4760f01b37
Model: shisa-ai/ablation-146-a128.dpo.armorm.rp.tl.1e6-shisa-v2-llama-3.1-8b Source: Original Platform
24485 lines
718 KiB
JSON
24485 lines
718 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 2.0,
|
|
"eval_steps": 500,
|
|
"global_step": 1630,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.001226993865030675,
|
|
"grad_norm": 151.34863688436073,
|
|
"learning_rate": 4e-08,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -474.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.6914,
|
|
"rewards/accuracies": 0.0,
|
|
"rewards/chosen": 0.0,
|
|
"rewards/margins": 0.0,
|
|
"rewards/rejected": 0.0,
|
|
"step": 1
|
|
},
|
|
{
|
|
"epoch": 0.00245398773006135,
|
|
"grad_norm": 162.78084419893554,
|
|
"learning_rate": 8e-08,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.6914,
|
|
"rewards/accuracies": 0.0,
|
|
"rewards/chosen": 0.0,
|
|
"rewards/margins": 0.0,
|
|
"rewards/rejected": 0.0,
|
|
"step": 2
|
|
},
|
|
{
|
|
"epoch": 0.0036809815950920245,
|
|
"grad_norm": 106.82291187813345,
|
|
"learning_rate": 1.2e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.6898,
|
|
"rewards/accuracies": 0.40625,
|
|
"rewards/chosen": 0.0250244140625,
|
|
"rewards/margins": 0.00946044921875,
|
|
"rewards/rejected": 0.01556396484375,
|
|
"step": 3
|
|
},
|
|
{
|
|
"epoch": 0.0049079754601227,
|
|
"grad_norm": 134.12056184559836,
|
|
"learning_rate": 1.6e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -242.0,
|
|
"loss": 0.7057,
|
|
"rewards/accuracies": 0.2578125,
|
|
"rewards/chosen": 0.00927734375,
|
|
"rewards/margins": -0.01495361328125,
|
|
"rewards/rejected": 0.0242919921875,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.006134969325153374,
|
|
"grad_norm": 303.3832578095922,
|
|
"learning_rate": 2e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.6884,
|
|
"rewards/accuracies": 0.34375,
|
|
"rewards/chosen": 0.03173828125,
|
|
"rewards/margins": 0.01458740234375,
|
|
"rewards/rejected": 0.01708984375,
|
|
"step": 5
|
|
},
|
|
{
|
|
"epoch": 0.007361963190184049,
|
|
"grad_norm": 323.77200794559565,
|
|
"learning_rate": 2.4e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.6888,
|
|
"rewards/accuracies": 0.4296875,
|
|
"rewards/chosen": 0.03076171875,
|
|
"rewards/margins": 0.0267333984375,
|
|
"rewards/rejected": 0.00408935546875,
|
|
"step": 6
|
|
},
|
|
{
|
|
"epoch": 0.008588957055214725,
|
|
"grad_norm": 242.90815448785312,
|
|
"learning_rate": 2.8e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.6953,
|
|
"rewards/accuracies": 0.3828125,
|
|
"rewards/chosen": 0.0037841796875,
|
|
"rewards/margins": 0.00604248046875,
|
|
"rewards/rejected": -0.0022430419921875,
|
|
"step": 7
|
|
},
|
|
{
|
|
"epoch": 0.0098159509202454,
|
|
"grad_norm": 127.34735766597795,
|
|
"learning_rate": 3.2e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.6919,
|
|
"rewards/accuracies": 0.3671875,
|
|
"rewards/chosen": 0.043701171875,
|
|
"rewards/margins": 0.01104736328125,
|
|
"rewards/rejected": 0.03271484375,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.011042944785276074,
|
|
"grad_norm": 115.54234285419575,
|
|
"learning_rate": 3.6e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.6835,
|
|
"rewards/accuracies": 0.3984375,
|
|
"rewards/chosen": 0.0537109375,
|
|
"rewards/margins": 0.0286865234375,
|
|
"rewards/rejected": 0.02490234375,
|
|
"step": 9
|
|
},
|
|
{
|
|
"epoch": 0.012269938650306749,
|
|
"grad_norm": 160.56175777535378,
|
|
"learning_rate": 4e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.6455,
|
|
"rewards/accuracies": 0.5390625,
|
|
"rewards/chosen": 0.0810546875,
|
|
"rewards/margins": 0.10888671875,
|
|
"rewards/rejected": -0.02783203125,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.013496932515337423,
|
|
"grad_norm": 184.2665155417569,
|
|
"learning_rate": 4.3999999999999997e-07,
|
|
"logits/chosen": -0.38671875,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.6427,
|
|
"rewards/accuracies": 0.5234375,
|
|
"rewards/chosen": 0.0732421875,
|
|
"rewards/margins": 0.1201171875,
|
|
"rewards/rejected": -0.04638671875,
|
|
"step": 11
|
|
},
|
|
{
|
|
"epoch": 0.014723926380368098,
|
|
"grad_norm": 263.0691476498303,
|
|
"learning_rate": 4.8e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.6103,
|
|
"rewards/accuracies": 0.6328125,
|
|
"rewards/chosen": 0.1474609375,
|
|
"rewards/margins": 0.21875,
|
|
"rewards/rejected": -0.0712890625,
|
|
"step": 12
|
|
},
|
|
{
|
|
"epoch": 0.015950920245398775,
|
|
"grad_norm": 102.73894172073577,
|
|
"learning_rate": 5.2e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -250.0,
|
|
"loss": 0.5678,
|
|
"rewards/accuracies": 0.6484375,
|
|
"rewards/chosen": 0.220703125,
|
|
"rewards/margins": 0.357421875,
|
|
"rewards/rejected": -0.1357421875,
|
|
"step": 13
|
|
},
|
|
{
|
|
"epoch": 0.01717791411042945,
|
|
"grad_norm": 91.63521794879738,
|
|
"learning_rate": 5.6e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -470.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.5302,
|
|
"rewards/accuracies": 0.71875,
|
|
"rewards/chosen": 0.396484375,
|
|
"rewards/margins": 0.52734375,
|
|
"rewards/rejected": -0.1298828125,
|
|
"step": 14
|
|
},
|
|
{
|
|
"epoch": 0.018404907975460124,
|
|
"grad_norm": 64.07563855366159,
|
|
"learning_rate": 6e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.1826171875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.5614,
|
|
"rewards/accuracies": 0.6171875,
|
|
"rewards/chosen": 0.310546875,
|
|
"rewards/margins": 0.435546875,
|
|
"rewards/rejected": -0.125,
|
|
"step": 15
|
|
},
|
|
{
|
|
"epoch": 0.0196319018404908,
|
|
"grad_norm": 66.54828155920123,
|
|
"learning_rate": 6.4e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.5093,
|
|
"rewards/accuracies": 0.71875,
|
|
"rewards/chosen": 0.412109375,
|
|
"rewards/margins": 0.62109375,
|
|
"rewards/rejected": -0.2080078125,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.020858895705521473,
|
|
"grad_norm": 69.61399588031946,
|
|
"learning_rate": 6.800000000000001e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -256.0,
|
|
"loss": 0.4793,
|
|
"rewards/accuracies": 0.7265625,
|
|
"rewards/chosen": 0.61328125,
|
|
"rewards/margins": 0.87890625,
|
|
"rewards/rejected": -0.267578125,
|
|
"step": 17
|
|
},
|
|
{
|
|
"epoch": 0.022085889570552148,
|
|
"grad_norm": 46.055018181732336,
|
|
"learning_rate": 7.2e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.4651,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 1.09375,
|
|
"rewards/rejected": -0.41796875,
|
|
"step": 18
|
|
},
|
|
{
|
|
"epoch": 0.023312883435582823,
|
|
"grad_norm": 77.14807287387903,
|
|
"learning_rate": 7.599999999999999e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.5136,
|
|
"rewards/accuracies": 0.7109375,
|
|
"rewards/chosen": 0.56640625,
|
|
"rewards/margins": 1.1171875,
|
|
"rewards/rejected": -0.5546875,
|
|
"step": 19
|
|
},
|
|
{
|
|
"epoch": 0.024539877300613498,
|
|
"grad_norm": 52.96981054915384,
|
|
"learning_rate": 8e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -460.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.4216,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 1.7890625,
|
|
"rewards/rejected": -0.83203125,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.025766871165644172,
|
|
"grad_norm": 49.55217029175373,
|
|
"learning_rate": 8.399999999999999e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -264.0,
|
|
"loss": 0.4123,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 1.390625,
|
|
"rewards/rejected": -0.6484375,
|
|
"step": 21
|
|
},
|
|
{
|
|
"epoch": 0.026993865030674847,
|
|
"grad_norm": 56.562310105836765,
|
|
"learning_rate": 8.799999999999999e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.4564,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 1.5,
|
|
"rewards/rejected": -0.81640625,
|
|
"step": 22
|
|
},
|
|
{
|
|
"epoch": 0.02822085889570552,
|
|
"grad_norm": 96.05342097645857,
|
|
"learning_rate": 9.2e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.4949,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 1.8359375,
|
|
"rewards/rejected": -0.86328125,
|
|
"step": 23
|
|
},
|
|
{
|
|
"epoch": 0.029447852760736196,
|
|
"grad_norm": 68.63010923117422,
|
|
"learning_rate": 9.6e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.4852,
|
|
"rewards/accuracies": 0.7265625,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 1.4375,
|
|
"rewards/rejected": -0.8671875,
|
|
"step": 24
|
|
},
|
|
{
|
|
"epoch": 0.03067484662576687,
|
|
"grad_norm": 40.52591666206953,
|
|
"learning_rate": 1e-06,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.375,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 2.46875,
|
|
"rewards/rejected": -1.3984375,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.03190184049079755,
|
|
"grad_norm": 47.31107870817392,
|
|
"learning_rate": 9.99586776859504e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.4293,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 2.59375,
|
|
"rewards/rejected": -1.203125,
|
|
"step": 26
|
|
},
|
|
{
|
|
"epoch": 0.033128834355828224,
|
|
"grad_norm": 42.03740863478121,
|
|
"learning_rate": 9.991735537190081e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3939,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 3.0,
|
|
"rewards/rejected": -1.4375,
|
|
"step": 27
|
|
},
|
|
{
|
|
"epoch": 0.0343558282208589,
|
|
"grad_norm": 38.22751255614571,
|
|
"learning_rate": 9.987603305785124e-07,
|
|
"logits/chosen": -0.0517578125,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.3875,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 2.578125,
|
|
"rewards/rejected": -0.98828125,
|
|
"step": 28
|
|
},
|
|
{
|
|
"epoch": 0.03558282208588957,
|
|
"grad_norm": 40.69846954289912,
|
|
"learning_rate": 9.983471074380165e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3733,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 3.125,
|
|
"rewards/rejected": -1.4140625,
|
|
"step": 29
|
|
},
|
|
{
|
|
"epoch": 0.03680981595092025,
|
|
"grad_norm": 46.55169582197367,
|
|
"learning_rate": 9.979338842975206e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.4096,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 2.484375,
|
|
"rewards/rejected": -1.234375,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.03803680981595092,
|
|
"grad_norm": 37.971298083652066,
|
|
"learning_rate": 9.975206611570247e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3889,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 3.25,
|
|
"rewards/rejected": -1.421875,
|
|
"step": 31
|
|
},
|
|
{
|
|
"epoch": 0.0392638036809816,
|
|
"grad_norm": 40.712562776553135,
|
|
"learning_rate": 9.971074380165288e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.384,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 3.125,
|
|
"rewards/rejected": -1.4765625,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.04049079754601227,
|
|
"grad_norm": 37.991087949351716,
|
|
"learning_rate": 9.96694214876033e-07,
|
|
"logits/chosen": -0.07958984375,
|
|
"logits/rejected": -0.2177734375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -238.0,
|
|
"loss": 0.3728,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 2.640625,
|
|
"rewards/rejected": -1.484375,
|
|
"step": 33
|
|
},
|
|
{
|
|
"epoch": 0.04171779141104295,
|
|
"grad_norm": 45.39723737748533,
|
|
"learning_rate": 9.962809917355372e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.4037,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 3.34375,
|
|
"rewards/rejected": -1.8984375,
|
|
"step": 34
|
|
},
|
|
{
|
|
"epoch": 0.04294478527607362,
|
|
"grad_norm": 46.41786525801273,
|
|
"learning_rate": 9.958677685950413e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.3792,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 3.15625,
|
|
"rewards/rejected": -2.109375,
|
|
"step": 35
|
|
},
|
|
{
|
|
"epoch": 0.044171779141104296,
|
|
"grad_norm": 42.51091401927911,
|
|
"learning_rate": 9.954545454545454e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3676,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 3.734375,
|
|
"rewards/rejected": -2.25,
|
|
"step": 36
|
|
},
|
|
{
|
|
"epoch": 0.04539877300613497,
|
|
"grad_norm": 40.33063661852525,
|
|
"learning_rate": 9.950413223140495e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.362,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 3.8125,
|
|
"rewards/rejected": -1.921875,
|
|
"step": 37
|
|
},
|
|
{
|
|
"epoch": 0.046625766871165646,
|
|
"grad_norm": 35.01835466546152,
|
|
"learning_rate": 9.946280991735538e-07,
|
|
"logits/chosen": -0.0947265625,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3457,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -2.09375,
|
|
"step": 38
|
|
},
|
|
{
|
|
"epoch": 0.04785276073619632,
|
|
"grad_norm": 46.14448752240649,
|
|
"learning_rate": 9.942148760330577e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3766,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.40625,
|
|
"step": 39
|
|
},
|
|
{
|
|
"epoch": 0.049079754601226995,
|
|
"grad_norm": 33.12245165013039,
|
|
"learning_rate": 9.93801652892562e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2986,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.46875,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 0.05030674846625767,
|
|
"grad_norm": 35.950653523061156,
|
|
"learning_rate": 9.93388429752066e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3625,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 3.984375,
|
|
"rewards/rejected": -2.328125,
|
|
"step": 41
|
|
},
|
|
{
|
|
"epoch": 0.051533742331288344,
|
|
"grad_norm": 33.25822119251682,
|
|
"learning_rate": 9.929752066115702e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3156,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 3.90625,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 42
|
|
},
|
|
{
|
|
"epoch": 0.05276073619631902,
|
|
"grad_norm": 35.99668142644002,
|
|
"learning_rate": 9.925619834710743e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2946,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.734375,
|
|
"step": 43
|
|
},
|
|
{
|
|
"epoch": 0.053987730061349694,
|
|
"grad_norm": 45.219634249357064,
|
|
"learning_rate": 9.921487603305784e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.4138,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 44
|
|
},
|
|
{
|
|
"epoch": 0.05521472392638037,
|
|
"grad_norm": 37.8165937134258,
|
|
"learning_rate": 9.917355371900827e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3291,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -2.875,
|
|
"step": 45
|
|
},
|
|
{
|
|
"epoch": 0.05644171779141104,
|
|
"grad_norm": 40.067988147487945,
|
|
"learning_rate": 9.913223140495868e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.302,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 46
|
|
},
|
|
{
|
|
"epoch": 0.05766871165644172,
|
|
"grad_norm": 62.331909497612145,
|
|
"learning_rate": 9.909090909090909e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.4649,
|
|
"rewards/accuracies": 0.7109375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 3.828125,
|
|
"rewards/rejected": -2.5,
|
|
"step": 47
|
|
},
|
|
{
|
|
"epoch": 0.05889570552147239,
|
|
"grad_norm": 35.63996127780391,
|
|
"learning_rate": 9.90495867768595e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3776,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 3.5,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 48
|
|
},
|
|
{
|
|
"epoch": 0.06012269938650307,
|
|
"grad_norm": 30.545142903578636,
|
|
"learning_rate": 9.90082644628099e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2534,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.0625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 49
|
|
},
|
|
{
|
|
"epoch": 0.06134969325153374,
|
|
"grad_norm": 46.15545512472155,
|
|
"learning_rate": 9.896694214876032e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3808,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 4.21875,
|
|
"rewards/rejected": -2.6875,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.06257668711656442,
|
|
"grad_norm": 42.158829592416716,
|
|
"learning_rate": 9.892561983471075e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3026,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -2.90625,
|
|
"step": 51
|
|
},
|
|
{
|
|
"epoch": 0.0638036809815951,
|
|
"grad_norm": 36.72373668526655,
|
|
"learning_rate": 9.888429752066116e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.32,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 52
|
|
},
|
|
{
|
|
"epoch": 0.06503067484662577,
|
|
"grad_norm": 35.57136422165816,
|
|
"learning_rate": 9.884297520661157e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2962,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.46875,
|
|
"step": 53
|
|
},
|
|
{
|
|
"epoch": 0.06625766871165645,
|
|
"grad_norm": 36.12409955593161,
|
|
"learning_rate": 9.880165289256197e-07,
|
|
"logits/chosen": -0.000728607177734375,
|
|
"logits/rejected": -0.17578125,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.2715,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -2.859375,
|
|
"step": 54
|
|
},
|
|
{
|
|
"epoch": 0.06748466257668712,
|
|
"grad_norm": 39.59210158362732,
|
|
"learning_rate": 9.87603305785124e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3207,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.75,
|
|
"step": 55
|
|
},
|
|
{
|
|
"epoch": 0.0687116564417178,
|
|
"grad_norm": 39.25115329003044,
|
|
"learning_rate": 9.871900826446281e-07,
|
|
"logits/chosen": -0.091796875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3727,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -2.4375,
|
|
"step": 56
|
|
},
|
|
{
|
|
"epoch": 0.06993865030674846,
|
|
"grad_norm": 46.72232187427657,
|
|
"learning_rate": 9.867768595041322e-07,
|
|
"logits/chosen": -0.09326171875,
|
|
"logits/rejected": -0.169921875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.357,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.59375,
|
|
"step": 57
|
|
},
|
|
{
|
|
"epoch": 0.07116564417177915,
|
|
"grad_norm": 43.616877681053715,
|
|
"learning_rate": 9.863636363636363e-07,
|
|
"logits/chosen": -0.1025390625,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3416,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.609375,
|
|
"step": 58
|
|
},
|
|
{
|
|
"epoch": 0.07239263803680981,
|
|
"grad_norm": 57.340851139568066,
|
|
"learning_rate": 9.859504132231404e-07,
|
|
"logits/chosen": -0.028564453125,
|
|
"logits/rejected": -0.11181640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.4351,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 3.578125,
|
|
"rewards/rejected": -2.21875,
|
|
"step": 59
|
|
},
|
|
{
|
|
"epoch": 0.0736196319018405,
|
|
"grad_norm": 40.41299854217699,
|
|
"learning_rate": 9.855371900826445e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3642,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 3.75,
|
|
"rewards/rejected": -2.5,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 0.07484662576687116,
|
|
"grad_norm": 38.38684468795477,
|
|
"learning_rate": 9.851239669421486e-07,
|
|
"logits/chosen": -0.1025390625,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3147,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -2.796875,
|
|
"step": 61
|
|
},
|
|
{
|
|
"epoch": 0.07607361963190185,
|
|
"grad_norm": 42.11986809638378,
|
|
"learning_rate": 9.84710743801653e-07,
|
|
"logits/chosen": -0.0869140625,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3517,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 2.609375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -2.796875,
|
|
"step": 62
|
|
},
|
|
{
|
|
"epoch": 0.07730061349693251,
|
|
"grad_norm": 48.46764704043021,
|
|
"learning_rate": 9.84297520661157e-07,
|
|
"logits/chosen": -0.005523681640625,
|
|
"logits/rejected": -0.1689453125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3154,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 4.21875,
|
|
"rewards/rejected": -2.5,
|
|
"step": 63
|
|
},
|
|
{
|
|
"epoch": 0.0785276073619632,
|
|
"grad_norm": 44.88518598351617,
|
|
"learning_rate": 9.838842975206611e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3722,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.25,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -2.546875,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.07975460122699386,
|
|
"grad_norm": 35.350496206715135,
|
|
"learning_rate": 9.834710743801652e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2996,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -2.515625,
|
|
"step": 65
|
|
},
|
|
{
|
|
"epoch": 0.08098159509202454,
|
|
"grad_norm": 34.14184871107505,
|
|
"learning_rate": 9.830578512396693e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.2818,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.328125,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 66
|
|
},
|
|
{
|
|
"epoch": 0.08220858895705521,
|
|
"grad_norm": 42.64233149605358,
|
|
"learning_rate": 9.826446280991734e-07,
|
|
"logits/chosen": -0.0869140625,
|
|
"logits/rejected": -0.205078125,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3617,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 67
|
|
},
|
|
{
|
|
"epoch": 0.0834355828220859,
|
|
"grad_norm": 41.39506967156979,
|
|
"learning_rate": 9.822314049586777e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3635,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.5,
|
|
"step": 68
|
|
},
|
|
{
|
|
"epoch": 0.08466257668711656,
|
|
"grad_norm": 32.22625490427397,
|
|
"learning_rate": 9.818181818181818e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.263,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.609375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 69
|
|
},
|
|
{
|
|
"epoch": 0.08588957055214724,
|
|
"grad_norm": 43.90551994332239,
|
|
"learning_rate": 9.814049586776859e-07,
|
|
"logits/chosen": -0.107421875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.315,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.59375,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 0.08711656441717791,
|
|
"grad_norm": 39.99506870265264,
|
|
"learning_rate": 9.8099173553719e-07,
|
|
"logits/chosen": -0.06201171875,
|
|
"logits/rejected": -0.1396484375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3872,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 3.9375,
|
|
"rewards/rejected": -2.34375,
|
|
"step": 71
|
|
},
|
|
{
|
|
"epoch": 0.08834355828220859,
|
|
"grad_norm": 41.349559255204774,
|
|
"learning_rate": 9.805785123966943e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.311,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -2.375,
|
|
"step": 72
|
|
},
|
|
{
|
|
"epoch": 0.08957055214723926,
|
|
"grad_norm": 45.978363733587855,
|
|
"learning_rate": 9.801652892561984e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.4089,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.671875,
|
|
"step": 73
|
|
},
|
|
{
|
|
"epoch": 0.09079754601226994,
|
|
"grad_norm": 38.839980321987454,
|
|
"learning_rate": 9.797520661157025e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2848,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 74
|
|
},
|
|
{
|
|
"epoch": 0.09202453987730061,
|
|
"grad_norm": 36.468204205869355,
|
|
"learning_rate": 9.793388429752066e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2732,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.09325153374233129,
|
|
"grad_norm": 51.75818303728974,
|
|
"learning_rate": 9.789256198347107e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3498,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 76
|
|
},
|
|
{
|
|
"epoch": 0.09447852760736196,
|
|
"grad_norm": 35.91522576533242,
|
|
"learning_rate": 9.785123966942148e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.284,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 77
|
|
},
|
|
{
|
|
"epoch": 0.09570552147239264,
|
|
"grad_norm": 39.01255683404935,
|
|
"learning_rate": 9.78099173553719e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2939,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.0,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.0,
|
|
"step": 78
|
|
},
|
|
{
|
|
"epoch": 0.09693251533742331,
|
|
"grad_norm": 41.678187193821934,
|
|
"learning_rate": 9.776859504132232e-07,
|
|
"logits/chosen": -0.1083984375,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3366,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 79
|
|
},
|
|
{
|
|
"epoch": 0.09815950920245399,
|
|
"grad_norm": 36.89274945940705,
|
|
"learning_rate": 9.772727272727273e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.285,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.671875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -3.265625,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 0.09938650306748466,
|
|
"grad_norm": 36.34213008191509,
|
|
"learning_rate": 9.768595041322313e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3048,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.28125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.28125,
|
|
"step": 81
|
|
},
|
|
{
|
|
"epoch": 0.10061349693251534,
|
|
"grad_norm": 51.83134303297906,
|
|
"learning_rate": 9.764462809917354e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3743,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 82
|
|
},
|
|
{
|
|
"epoch": 0.10184049079754601,
|
|
"grad_norm": 44.6306513701737,
|
|
"learning_rate": 9.760330578512395e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.326,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 83
|
|
},
|
|
{
|
|
"epoch": 0.10306748466257669,
|
|
"grad_norm": 43.368403842908165,
|
|
"learning_rate": 9.756198347107436e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2915,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 84
|
|
},
|
|
{
|
|
"epoch": 0.10429447852760736,
|
|
"grad_norm": 41.83545101757536,
|
|
"learning_rate": 9.75206611570248e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2993,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 85
|
|
},
|
|
{
|
|
"epoch": 0.10552147239263804,
|
|
"grad_norm": 42.79642482122795,
|
|
"learning_rate": 9.74793388429752e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3469,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 86
|
|
},
|
|
{
|
|
"epoch": 0.1067484662576687,
|
|
"grad_norm": 31.200859065220833,
|
|
"learning_rate": 9.743801652892561e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.2792,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 87
|
|
},
|
|
{
|
|
"epoch": 0.10797546012269939,
|
|
"grad_norm": 33.089204895794104,
|
|
"learning_rate": 9.739669421487602e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.2632,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 88
|
|
},
|
|
{
|
|
"epoch": 0.10920245398773006,
|
|
"grad_norm": 41.44285417719672,
|
|
"learning_rate": 9.735537190082645e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3377,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 89
|
|
},
|
|
{
|
|
"epoch": 0.11042944785276074,
|
|
"grad_norm": 33.95249948655705,
|
|
"learning_rate": 9.731404958677686e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3025,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 0.1116564417177914,
|
|
"grad_norm": 39.291583688315264,
|
|
"learning_rate": 9.727272727272727e-07,
|
|
"logits/chosen": -0.033935546875,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2872,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.203125,
|
|
"step": 91
|
|
},
|
|
{
|
|
"epoch": 0.11288343558282209,
|
|
"grad_norm": 34.11914530579394,
|
|
"learning_rate": 9.723140495867768e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2706,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 92
|
|
},
|
|
{
|
|
"epoch": 0.11411042944785275,
|
|
"grad_norm": 35.0129736595644,
|
|
"learning_rate": 9.71900826446281e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.271,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.359375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 93
|
|
},
|
|
{
|
|
"epoch": 0.11533742331288344,
|
|
"grad_norm": 39.56522119452336,
|
|
"learning_rate": 9.71487603305785e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2779,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.53125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 94
|
|
},
|
|
{
|
|
"epoch": 0.1165644171779141,
|
|
"grad_norm": 35.12017401888975,
|
|
"learning_rate": 9.710743801652893e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2912,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.0,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 95
|
|
},
|
|
{
|
|
"epoch": 0.11779141104294479,
|
|
"grad_norm": 35.951495412713705,
|
|
"learning_rate": 9.706611570247934e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2652,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.40625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 96
|
|
},
|
|
{
|
|
"epoch": 0.11901840490797547,
|
|
"grad_norm": 46.49379123938547,
|
|
"learning_rate": 9.702479338842975e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3607,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.21875,
|
|
"step": 97
|
|
},
|
|
{
|
|
"epoch": 0.12024539877300613,
|
|
"grad_norm": 38.4803512021978,
|
|
"learning_rate": 9.698347107438016e-07,
|
|
"logits/chosen": -0.04638671875,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.2977,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 98
|
|
},
|
|
{
|
|
"epoch": 0.12147239263803682,
|
|
"grad_norm": 39.97761004406295,
|
|
"learning_rate": 9.694214876033057e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -247.0,
|
|
"loss": 0.268,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 99
|
|
},
|
|
{
|
|
"epoch": 0.12269938650306748,
|
|
"grad_norm": 37.95641413102004,
|
|
"learning_rate": 9.6900826446281e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3193,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.5,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.12392638036809817,
|
|
"grad_norm": 46.76166817261012,
|
|
"learning_rate": 9.685950413223139e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.365,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 101
|
|
},
|
|
{
|
|
"epoch": 0.12515337423312883,
|
|
"grad_norm": 41.79765734110343,
|
|
"learning_rate": 9.681818181818182e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3324,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 102
|
|
},
|
|
{
|
|
"epoch": 0.1263803680981595,
|
|
"grad_norm": 36.794153558575765,
|
|
"learning_rate": 9.677685950413223e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2922,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 103
|
|
},
|
|
{
|
|
"epoch": 0.1276073619631902,
|
|
"grad_norm": 50.370859060453284,
|
|
"learning_rate": 9.673553719008264e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3734,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.078125,
|
|
"step": 104
|
|
},
|
|
{
|
|
"epoch": 0.12883435582822086,
|
|
"grad_norm": 34.529612759458495,
|
|
"learning_rate": 9.669421487603305e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2608,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 105
|
|
},
|
|
{
|
|
"epoch": 0.13006134969325153,
|
|
"grad_norm": 39.013908957263666,
|
|
"learning_rate": 9.665289256198348e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3065,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 106
|
|
},
|
|
{
|
|
"epoch": 0.1312883435582822,
|
|
"grad_norm": 42.150982723963395,
|
|
"learning_rate": 9.661157024793389e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2956,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 107
|
|
},
|
|
{
|
|
"epoch": 0.1325153374233129,
|
|
"grad_norm": 29.969554041041413,
|
|
"learning_rate": 9.65702479338843e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2406,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 108
|
|
},
|
|
{
|
|
"epoch": 0.13374233128834356,
|
|
"grad_norm": 36.72032157988445,
|
|
"learning_rate": 9.65289256198347e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.317,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.734375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 109
|
|
},
|
|
{
|
|
"epoch": 0.13496932515337423,
|
|
"grad_norm": 40.64788730782337,
|
|
"learning_rate": 9.648760330578511e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3312,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 0.1361963190184049,
|
|
"grad_norm": 34.71283913194937,
|
|
"learning_rate": 9.644628099173552e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2952,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 111
|
|
},
|
|
{
|
|
"epoch": 0.1374233128834356,
|
|
"grad_norm": 36.76218297956249,
|
|
"learning_rate": 9.640495867768595e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2795,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 112
|
|
},
|
|
{
|
|
"epoch": 0.13865030674846626,
|
|
"grad_norm": 39.011108940957456,
|
|
"learning_rate": 9.636363636363636e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.193359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3276,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 113
|
|
},
|
|
{
|
|
"epoch": 0.13987730061349693,
|
|
"grad_norm": 42.342001118592165,
|
|
"learning_rate": 9.632231404958677e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3342,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 114
|
|
},
|
|
{
|
|
"epoch": 0.1411042944785276,
|
|
"grad_norm": 48.02563322504298,
|
|
"learning_rate": 9.628099173553718e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3999,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.98046875,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 115
|
|
},
|
|
{
|
|
"epoch": 0.1423312883435583,
|
|
"grad_norm": 31.044871089885095,
|
|
"learning_rate": 9.62396694214876e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.262,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.53125,
|
|
"step": 116
|
|
},
|
|
{
|
|
"epoch": 0.14355828220858896,
|
|
"grad_norm": 37.17488322550573,
|
|
"learning_rate": 9.619834710743802e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.3079,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 117
|
|
},
|
|
{
|
|
"epoch": 0.14478527607361963,
|
|
"grad_norm": 38.57684603177769,
|
|
"learning_rate": 9.615702479338843e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3182,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.53125,
|
|
"step": 118
|
|
},
|
|
{
|
|
"epoch": 0.1460122699386503,
|
|
"grad_norm": 39.31837627767617,
|
|
"learning_rate": 9.611570247933884e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.308,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 119
|
|
},
|
|
{
|
|
"epoch": 0.147239263803681,
|
|
"grad_norm": 47.322798683163334,
|
|
"learning_rate": 9.607438016528925e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3156,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.53125,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 0.14846625766871166,
|
|
"grad_norm": 35.88276223707849,
|
|
"learning_rate": 9.603305785123966e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 2.34375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 121
|
|
},
|
|
{
|
|
"epoch": 0.14969325153374233,
|
|
"grad_norm": 43.922306327606435,
|
|
"learning_rate": 9.599173553719007e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3011,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.5,
|
|
"step": 122
|
|
},
|
|
{
|
|
"epoch": 0.150920245398773,
|
|
"grad_norm": 44.742388015169865,
|
|
"learning_rate": 9.59504132231405e-07,
|
|
"logits/chosen": -0.04931640625,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3574,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 123
|
|
},
|
|
{
|
|
"epoch": 0.1521472392638037,
|
|
"grad_norm": 34.160343795291915,
|
|
"learning_rate": 9.59090909090909e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.2674,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 124
|
|
},
|
|
{
|
|
"epoch": 0.15337423312883436,
|
|
"grad_norm": 39.518799854999045,
|
|
"learning_rate": 9.586776859504132e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2871,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.15460122699386503,
|
|
"grad_norm": 56.25861619658859,
|
|
"learning_rate": 9.582644628099173e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3637,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 126
|
|
},
|
|
{
|
|
"epoch": 0.1558282208588957,
|
|
"grad_norm": 42.50340403129208,
|
|
"learning_rate": 9.578512396694216e-07,
|
|
"logits/chosen": -0.08837890625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3851,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 127
|
|
},
|
|
{
|
|
"epoch": 0.1570552147239264,
|
|
"grad_norm": 44.339394645108435,
|
|
"learning_rate": 9.574380165289255e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3155,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.15828220858895706,
|
|
"grad_norm": 35.34227561050985,
|
|
"learning_rate": 9.570247933884298e-07,
|
|
"logits/chosen": -0.040283203125,
|
|
"logits/rejected": -0.169921875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3153,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 129
|
|
},
|
|
{
|
|
"epoch": 0.15950920245398773,
|
|
"grad_norm": 38.43919164614137,
|
|
"learning_rate": 9.566115702479339e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3349,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 0.1607361963190184,
|
|
"grad_norm": 34.097476299359684,
|
|
"learning_rate": 9.56198347107438e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2132,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 131
|
|
},
|
|
{
|
|
"epoch": 0.1619631901840491,
|
|
"grad_norm": 42.765865552519735,
|
|
"learning_rate": 9.55785123966942e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3402,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 132
|
|
},
|
|
{
|
|
"epoch": 0.16319018404907976,
|
|
"grad_norm": 40.18673593506736,
|
|
"learning_rate": 9.553719008264462e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2822,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 133
|
|
},
|
|
{
|
|
"epoch": 0.16441717791411042,
|
|
"grad_norm": 54.50121988482124,
|
|
"learning_rate": 9.549586776859505e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3145,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 134
|
|
},
|
|
{
|
|
"epoch": 0.1656441717791411,
|
|
"grad_norm": 44.50905430648629,
|
|
"learning_rate": 9.545454545454546e-07,
|
|
"logits/chosen": -0.05517578125,
|
|
"logits/rejected": -0.1669921875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3672,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 135
|
|
},
|
|
{
|
|
"epoch": 0.1668711656441718,
|
|
"grad_norm": 37.03632873479481,
|
|
"learning_rate": 9.541322314049586e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3149,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 136
|
|
},
|
|
{
|
|
"epoch": 0.16809815950920245,
|
|
"grad_norm": 46.53968885524061,
|
|
"learning_rate": 9.537190082644627e-07,
|
|
"logits/chosen": -0.0576171875,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3795,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 137
|
|
},
|
|
{
|
|
"epoch": 0.16932515337423312,
|
|
"grad_norm": 39.2528692217884,
|
|
"learning_rate": 9.533057851239668e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3313,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 138
|
|
},
|
|
{
|
|
"epoch": 0.1705521472392638,
|
|
"grad_norm": 41.85887635916101,
|
|
"learning_rate": 9.52892561983471e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3652,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 139
|
|
},
|
|
{
|
|
"epoch": 0.17177914110429449,
|
|
"grad_norm": 50.211692574184056,
|
|
"learning_rate": 9.524793388429751e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3576,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 0.17300613496932515,
|
|
"grad_norm": 34.61921033025654,
|
|
"learning_rate": 9.520661157024793e-07,
|
|
"logits/chosen": -0.08349609375,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.267,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 141
|
|
},
|
|
{
|
|
"epoch": 0.17423312883435582,
|
|
"grad_norm": 39.84749847418945,
|
|
"learning_rate": 9.516528925619834e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3431,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 142
|
|
},
|
|
{
|
|
"epoch": 0.1754601226993865,
|
|
"grad_norm": 41.19383820823238,
|
|
"learning_rate": 9.512396694214876e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2931,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 143
|
|
},
|
|
{
|
|
"epoch": 0.17668711656441718,
|
|
"grad_norm": 39.48774183772262,
|
|
"learning_rate": 9.508264462809917e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2937,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 144
|
|
},
|
|
{
|
|
"epoch": 0.17791411042944785,
|
|
"grad_norm": 39.31332603384939,
|
|
"learning_rate": 9.504132231404959e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2665,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 145
|
|
},
|
|
{
|
|
"epoch": 0.17914110429447852,
|
|
"grad_norm": 32.69853161252682,
|
|
"learning_rate": 9.499999999999999e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2369,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 146
|
|
},
|
|
{
|
|
"epoch": 0.18036809815950922,
|
|
"grad_norm": 33.92758965649418,
|
|
"learning_rate": 9.495867768595041e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2687,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 147
|
|
},
|
|
{
|
|
"epoch": 0.18159509202453988,
|
|
"grad_norm": 49.11649217825285,
|
|
"learning_rate": 9.491735537190082e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3872,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 148
|
|
},
|
|
{
|
|
"epoch": 0.18282208588957055,
|
|
"grad_norm": 42.0523811991046,
|
|
"learning_rate": 9.487603305785124e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.2569,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.921875,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 149
|
|
},
|
|
{
|
|
"epoch": 0.18404907975460122,
|
|
"grad_norm": 37.91815584989451,
|
|
"learning_rate": 9.483471074380165e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3146,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.18527607361963191,
|
|
"grad_norm": 39.01655480069067,
|
|
"learning_rate": 9.479338842975206e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3162,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 151
|
|
},
|
|
{
|
|
"epoch": 0.18650306748466258,
|
|
"grad_norm": 36.295557767109855,
|
|
"learning_rate": 9.475206611570248e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.284,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 152
|
|
},
|
|
{
|
|
"epoch": 0.18773006134969325,
|
|
"grad_norm": 35.43408456309788,
|
|
"learning_rate": 9.471074380165289e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.765625,
|
|
"step": 153
|
|
},
|
|
{
|
|
"epoch": 0.18895705521472392,
|
|
"grad_norm": 45.691475338279055,
|
|
"learning_rate": 9.466942148760331e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3274,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 154
|
|
},
|
|
{
|
|
"epoch": 0.1901840490797546,
|
|
"grad_norm": 28.365158188278386,
|
|
"learning_rate": 9.462809917355371e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.1938,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.625,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.0,
|
|
"step": 155
|
|
},
|
|
{
|
|
"epoch": 0.19141104294478528,
|
|
"grad_norm": 43.775694095180505,
|
|
"learning_rate": 9.458677685950413e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.22265625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.315,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 156
|
|
},
|
|
{
|
|
"epoch": 0.19263803680981595,
|
|
"grad_norm": 35.92254576769146,
|
|
"learning_rate": 9.454545454545454e-07,
|
|
"logits/chosen": -0.103515625,
|
|
"logits/rejected": -0.232421875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2914,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 157
|
|
},
|
|
{
|
|
"epoch": 0.19386503067484662,
|
|
"grad_norm": 32.68696317336219,
|
|
"learning_rate": 9.450413223140496e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2534,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 158
|
|
},
|
|
{
|
|
"epoch": 0.1950920245398773,
|
|
"grad_norm": 40.79749046301349,
|
|
"learning_rate": 9.446280991735537e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3199,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 159
|
|
},
|
|
{
|
|
"epoch": 0.19631901840490798,
|
|
"grad_norm": 28.669147296818444,
|
|
"learning_rate": 9.442148760330579e-07,
|
|
"logits/chosen": -0.0908203125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2185,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 0.19754601226993865,
|
|
"grad_norm": 35.436535721930035,
|
|
"learning_rate": 9.43801652892562e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2404,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 161
|
|
},
|
|
{
|
|
"epoch": 0.19877300613496932,
|
|
"grad_norm": 43.95617745827823,
|
|
"learning_rate": 9.433884297520662e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3681,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 162
|
|
},
|
|
{
|
|
"epoch": 0.2,
|
|
"grad_norm": 41.651239703295666,
|
|
"learning_rate": 9.429752066115701e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2972,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 163
|
|
},
|
|
{
|
|
"epoch": 0.20122699386503068,
|
|
"grad_norm": 30.82750182223144,
|
|
"learning_rate": 9.425619834710743e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2315,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 164
|
|
},
|
|
{
|
|
"epoch": 0.20245398773006135,
|
|
"grad_norm": 44.48878395259188,
|
|
"learning_rate": 9.421487603305784e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3043,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 165
|
|
},
|
|
{
|
|
"epoch": 0.20368098159509201,
|
|
"grad_norm": 41.12003400664833,
|
|
"learning_rate": 9.417355371900826e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3598,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 166
|
|
},
|
|
{
|
|
"epoch": 0.2049079754601227,
|
|
"grad_norm": 40.52632483767134,
|
|
"learning_rate": 9.413223140495867e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3636,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.3125,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -3.625,
|
|
"step": 167
|
|
},
|
|
{
|
|
"epoch": 0.20613496932515338,
|
|
"grad_norm": 38.13990823039671,
|
|
"learning_rate": 9.409090909090908e-07,
|
|
"logits/chosen": -0.0986328125,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.2792,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 168
|
|
},
|
|
{
|
|
"epoch": 0.20736196319018405,
|
|
"grad_norm": 37.3898766928124,
|
|
"learning_rate": 9.40495867768595e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2887,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 169
|
|
},
|
|
{
|
|
"epoch": 0.2085889570552147,
|
|
"grad_norm": 42.62803720646298,
|
|
"learning_rate": 9.400826446280991e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3073,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 0.2098159509202454,
|
|
"grad_norm": 33.8264512159167,
|
|
"learning_rate": 9.396694214876033e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2986,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 171
|
|
},
|
|
{
|
|
"epoch": 0.21104294478527608,
|
|
"grad_norm": 35.98076154174604,
|
|
"learning_rate": 9.392561983471073e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.27,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 172
|
|
},
|
|
{
|
|
"epoch": 0.21226993865030674,
|
|
"grad_norm": 27.651041582591024,
|
|
"learning_rate": 9.388429752066115e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.1967,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.453125,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 173
|
|
},
|
|
{
|
|
"epoch": 0.2134969325153374,
|
|
"grad_norm": 42.78349510928123,
|
|
"learning_rate": 9.384297520661156e-07,
|
|
"logits/chosen": -0.01483154296875,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3777,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 174
|
|
},
|
|
{
|
|
"epoch": 0.2147239263803681,
|
|
"grad_norm": 49.466708668886746,
|
|
"learning_rate": 9.380165289256198e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.361,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 0.21595092024539878,
|
|
"grad_norm": 37.891097032789595,
|
|
"learning_rate": 9.376033057851239e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3265,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 176
|
|
},
|
|
{
|
|
"epoch": 0.21717791411042944,
|
|
"grad_norm": 29.591204044267382,
|
|
"learning_rate": 9.371900826446281e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2365,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.578125,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 177
|
|
},
|
|
{
|
|
"epoch": 0.2184049079754601,
|
|
"grad_norm": 39.29643939569424,
|
|
"learning_rate": 9.367768595041322e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3276,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 178
|
|
},
|
|
{
|
|
"epoch": 0.2196319018404908,
|
|
"grad_norm": 32.63548187592793,
|
|
"learning_rate": 9.363636363636364e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2755,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.98046875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 179
|
|
},
|
|
{
|
|
"epoch": 0.22085889570552147,
|
|
"grad_norm": 41.35724310032657,
|
|
"learning_rate": 9.359504132231405e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3359,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 0.22208588957055214,
|
|
"grad_norm": 41.06602522381907,
|
|
"learning_rate": 9.355371900826446e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.3587,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 181
|
|
},
|
|
{
|
|
"epoch": 0.2233128834355828,
|
|
"grad_norm": 38.85418184494936,
|
|
"learning_rate": 9.351239669421487e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3357,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 182
|
|
},
|
|
{
|
|
"epoch": 0.2245398773006135,
|
|
"grad_norm": 42.84335154848466,
|
|
"learning_rate": 9.347107438016529e-07,
|
|
"logits/chosen": -0.07275390625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.342,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.83984375,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 183
|
|
},
|
|
{
|
|
"epoch": 0.22576687116564417,
|
|
"grad_norm": 46.28637855648245,
|
|
"learning_rate": 9.34297520661157e-07,
|
|
"logits/chosen": -0.10693359375,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3705,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 184
|
|
},
|
|
{
|
|
"epoch": 0.22699386503067484,
|
|
"grad_norm": 49.113931033261764,
|
|
"learning_rate": 9.338842975206612e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.2177734375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3514,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 185
|
|
},
|
|
{
|
|
"epoch": 0.2282208588957055,
|
|
"grad_norm": 37.05498191674873,
|
|
"learning_rate": 9.334710743801653e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3089,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 186
|
|
},
|
|
{
|
|
"epoch": 0.2294478527607362,
|
|
"grad_norm": 37.71647333523593,
|
|
"learning_rate": 9.330578512396694e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2659,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 187
|
|
},
|
|
{
|
|
"epoch": 0.23067484662576687,
|
|
"grad_norm": 33.868390403245435,
|
|
"learning_rate": 9.326446280991736e-07,
|
|
"logits/chosen": -0.056884765625,
|
|
"logits/rejected": -0.15234375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3049,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 188
|
|
},
|
|
{
|
|
"epoch": 0.23190184049079754,
|
|
"grad_norm": 32.6551143996549,
|
|
"learning_rate": 9.322314049586776e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2185,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8515625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 189
|
|
},
|
|
{
|
|
"epoch": 0.2331288343558282,
|
|
"grad_norm": 34.45171325611373,
|
|
"learning_rate": 9.318181818181817e-07,
|
|
"logits/chosen": -0.04296875,
|
|
"logits/rejected": -0.205078125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.292,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 0.2343558282208589,
|
|
"grad_norm": 50.99221887208661,
|
|
"learning_rate": 9.314049586776858e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3858,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 191
|
|
},
|
|
{
|
|
"epoch": 0.23558282208588957,
|
|
"grad_norm": 47.091436431958705,
|
|
"learning_rate": 9.3099173553719e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3259,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 192
|
|
},
|
|
{
|
|
"epoch": 0.23680981595092024,
|
|
"grad_norm": 36.36482495891595,
|
|
"learning_rate": 9.305785123966941e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2485,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 193
|
|
},
|
|
{
|
|
"epoch": 0.23803680981595093,
|
|
"grad_norm": 37.6633231814014,
|
|
"learning_rate": 9.301652892561983e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2939,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 194
|
|
},
|
|
{
|
|
"epoch": 0.2392638036809816,
|
|
"grad_norm": 34.69492122435969,
|
|
"learning_rate": 9.297520661157024e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2298,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.359375,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 195
|
|
},
|
|
{
|
|
"epoch": 0.24049079754601227,
|
|
"grad_norm": 41.61809014057929,
|
|
"learning_rate": 9.293388429752066e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3133,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 196
|
|
},
|
|
{
|
|
"epoch": 0.24171779141104294,
|
|
"grad_norm": 31.620797247972778,
|
|
"learning_rate": 9.289256198347107e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.282,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 197
|
|
},
|
|
{
|
|
"epoch": 0.24294478527607363,
|
|
"grad_norm": 39.95412220257755,
|
|
"learning_rate": 9.285123966942149e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.3762,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 198
|
|
},
|
|
{
|
|
"epoch": 0.2441717791411043,
|
|
"grad_norm": 27.11893031942543,
|
|
"learning_rate": 9.280991735537189e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.1959,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.46875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 199
|
|
},
|
|
{
|
|
"epoch": 0.24539877300613497,
|
|
"grad_norm": 42.12057014134938,
|
|
"learning_rate": 9.276859504132231e-07,
|
|
"logits/chosen": -0.11669921875,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.3142,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.234375,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.24662576687116564,
|
|
"grad_norm": 32.34510842493663,
|
|
"learning_rate": 9.272727272727272e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2316,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 201
|
|
},
|
|
{
|
|
"epoch": 0.24785276073619633,
|
|
"grad_norm": 34.14824944726285,
|
|
"learning_rate": 9.268595041322314e-07,
|
|
"logits/chosen": -0.10205078125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2685,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 202
|
|
},
|
|
{
|
|
"epoch": 0.249079754601227,
|
|
"grad_norm": 40.81109270790126,
|
|
"learning_rate": 9.264462809917355e-07,
|
|
"logits/chosen": -0.06787109375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.4324,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 203
|
|
},
|
|
{
|
|
"epoch": 0.25030674846625767,
|
|
"grad_norm": 39.08482579609809,
|
|
"learning_rate": 9.260330578512396e-07,
|
|
"logits/chosen": -0.05712890625,
|
|
"logits/rejected": -0.203125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.3833,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 204
|
|
},
|
|
{
|
|
"epoch": 0.25153374233128833,
|
|
"grad_norm": 36.559212995740964,
|
|
"learning_rate": 9.256198347107438e-07,
|
|
"logits/chosen": -0.00885009765625,
|
|
"logits/rejected": -0.1533203125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3464,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 205
|
|
},
|
|
{
|
|
"epoch": 0.252760736196319,
|
|
"grad_norm": 26.131004803554553,
|
|
"learning_rate": 9.252066115702479e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2208,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 206
|
|
},
|
|
{
|
|
"epoch": 0.25398773006134967,
|
|
"grad_norm": 42.01910176739678,
|
|
"learning_rate": 9.247933884297521e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3729,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 207
|
|
},
|
|
{
|
|
"epoch": 0.2552147239263804,
|
|
"grad_norm": 38.03191943662351,
|
|
"learning_rate": 9.243801652892561e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3343,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 208
|
|
},
|
|
{
|
|
"epoch": 0.25644171779141106,
|
|
"grad_norm": 36.80831687011056,
|
|
"learning_rate": 9.239669421487603e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3148,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.875,
|
|
"step": 209
|
|
},
|
|
{
|
|
"epoch": 0.25766871165644173,
|
|
"grad_norm": 44.11101399224612,
|
|
"learning_rate": 9.235537190082644e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.3628,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.63671875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 0.2588957055214724,
|
|
"grad_norm": 38.561232670471654,
|
|
"learning_rate": 9.231404958677686e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3366,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.5,
|
|
"step": 211
|
|
},
|
|
{
|
|
"epoch": 0.26012269938650306,
|
|
"grad_norm": 36.86026213129111,
|
|
"learning_rate": 9.227272727272727e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2974,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 212
|
|
},
|
|
{
|
|
"epoch": 0.26134969325153373,
|
|
"grad_norm": 33.76063146504063,
|
|
"learning_rate": 9.223140495867769e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.1962890625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2679,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 213
|
|
},
|
|
{
|
|
"epoch": 0.2625766871165644,
|
|
"grad_norm": 44.147098970181965,
|
|
"learning_rate": 9.21900826446281e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3562,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.0,
|
|
"step": 214
|
|
},
|
|
{
|
|
"epoch": 0.26380368098159507,
|
|
"grad_norm": 44.646862605268325,
|
|
"learning_rate": 9.214876033057852e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.1796875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3235,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.62890625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 215
|
|
},
|
|
{
|
|
"epoch": 0.2650306748466258,
|
|
"grad_norm": 40.25809763769465,
|
|
"learning_rate": 9.210743801652892e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3204,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.447265625,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 216
|
|
},
|
|
{
|
|
"epoch": 0.26625766871165646,
|
|
"grad_norm": 42.423324639307225,
|
|
"learning_rate": 9.206611570247933e-07,
|
|
"logits/chosen": -0.04248046875,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.4054,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 217
|
|
},
|
|
{
|
|
"epoch": 0.2674846625766871,
|
|
"grad_norm": 46.69438445358285,
|
|
"learning_rate": 9.202479338842974e-07,
|
|
"logits/chosen": -0.0849609375,
|
|
"logits/rejected": -0.2021484375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3932,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 218
|
|
},
|
|
{
|
|
"epoch": 0.2687116564417178,
|
|
"grad_norm": 35.2004707186147,
|
|
"learning_rate": 9.198347107438016e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3107,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 219
|
|
},
|
|
{
|
|
"epoch": 0.26993865030674846,
|
|
"grad_norm": 50.169633893127326,
|
|
"learning_rate": 9.194214876033057e-07,
|
|
"logits/chosen": -0.0322265625,
|
|
"logits/rejected": -0.1123046875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.4295,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.68359375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 0.27116564417177913,
|
|
"grad_norm": 42.093475734809076,
|
|
"learning_rate": 9.190082644628098e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3922,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 221
|
|
},
|
|
{
|
|
"epoch": 0.2723926380368098,
|
|
"grad_norm": 36.0657090040612,
|
|
"learning_rate": 9.18595041322314e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2895,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 222
|
|
},
|
|
{
|
|
"epoch": 0.27361963190184047,
|
|
"grad_norm": 37.91090520092486,
|
|
"learning_rate": 9.181818181818181e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2975,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 223
|
|
},
|
|
{
|
|
"epoch": 0.2748466257668712,
|
|
"grad_norm": 39.87455871283598,
|
|
"learning_rate": 9.177685950413223e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2872,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 224
|
|
},
|
|
{
|
|
"epoch": 0.27607361963190186,
|
|
"grad_norm": 35.385311266968216,
|
|
"learning_rate": 9.173553719008264e-07,
|
|
"logits/chosen": -0.09228515625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2718,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.75,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 0.2773006134969325,
|
|
"grad_norm": 37.620291584299586,
|
|
"learning_rate": 9.169421487603305e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.353,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 226
|
|
},
|
|
{
|
|
"epoch": 0.2785276073619632,
|
|
"grad_norm": 41.70527475343198,
|
|
"learning_rate": 9.165289256198346e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.22265625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.352,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 227
|
|
},
|
|
{
|
|
"epoch": 0.27975460122699386,
|
|
"grad_norm": 47.672730884256275,
|
|
"learning_rate": 9.161157024793388e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3342,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.375,
|
|
"step": 228
|
|
},
|
|
{
|
|
"epoch": 0.2809815950920245,
|
|
"grad_norm": 36.5310600090198,
|
|
"learning_rate": 9.157024793388429e-07,
|
|
"logits/chosen": -0.05908203125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2811,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 229
|
|
},
|
|
{
|
|
"epoch": 0.2822085889570552,
|
|
"grad_norm": 34.01874266449764,
|
|
"learning_rate": 9.152892561983471e-07,
|
|
"logits/chosen": -0.06005859375,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.253,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.625,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 0.28343558282208586,
|
|
"grad_norm": 43.76870205345571,
|
|
"learning_rate": 9.148760330578512e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.4191,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.875,
|
|
"step": 231
|
|
},
|
|
{
|
|
"epoch": 0.2846625766871166,
|
|
"grad_norm": 36.17795323546657,
|
|
"learning_rate": 9.144628099173554e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3191,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 232
|
|
},
|
|
{
|
|
"epoch": 0.28588957055214725,
|
|
"grad_norm": 35.1635442877376,
|
|
"learning_rate": 9.140495867768595e-07,
|
|
"logits/chosen": -0.10205078125,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3072,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.8828125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 233
|
|
},
|
|
{
|
|
"epoch": 0.2871165644171779,
|
|
"grad_norm": 37.827856451736956,
|
|
"learning_rate": 9.136363636363637e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3271,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.75,
|
|
"step": 234
|
|
},
|
|
{
|
|
"epoch": 0.2883435582822086,
|
|
"grad_norm": 36.88124628560585,
|
|
"learning_rate": 9.132231404958677e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3114,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.125,
|
|
"step": 235
|
|
},
|
|
{
|
|
"epoch": 0.28957055214723926,
|
|
"grad_norm": 48.08754663310893,
|
|
"learning_rate": 9.128099173553719e-07,
|
|
"logits/chosen": -0.0145263671875,
|
|
"logits/rejected": -0.1962890625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.4279,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.234375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.25,
|
|
"step": 236
|
|
},
|
|
{
|
|
"epoch": 0.2907975460122699,
|
|
"grad_norm": 33.733681477930965,
|
|
"learning_rate": 9.12396694214876e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2814,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 237
|
|
},
|
|
{
|
|
"epoch": 0.2920245398773006,
|
|
"grad_norm": 41.709782097463915,
|
|
"learning_rate": 9.119834710743802e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3809,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.765625,
|
|
"step": 238
|
|
},
|
|
{
|
|
"epoch": 0.29325153374233126,
|
|
"grad_norm": 41.613433419635136,
|
|
"learning_rate": 9.115702479338843e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3411,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.3125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.0,
|
|
"step": 239
|
|
},
|
|
{
|
|
"epoch": 0.294478527607362,
|
|
"grad_norm": 33.76902045077368,
|
|
"learning_rate": 9.111570247933884e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2174,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.625,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 0.29570552147239265,
|
|
"grad_norm": 35.54809707331976,
|
|
"learning_rate": 9.107438016528926e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2793,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.25,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 241
|
|
},
|
|
{
|
|
"epoch": 0.2969325153374233,
|
|
"grad_norm": 31.39332422823268,
|
|
"learning_rate": 9.103305785123967e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.238,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 242
|
|
},
|
|
{
|
|
"epoch": 0.298159509202454,
|
|
"grad_norm": 42.489213771046316,
|
|
"learning_rate": 9.099173553719007e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3768,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 243
|
|
},
|
|
{
|
|
"epoch": 0.29938650306748466,
|
|
"grad_norm": 42.520272970122726,
|
|
"learning_rate": 9.095041322314048e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.4879,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 244
|
|
},
|
|
{
|
|
"epoch": 0.3006134969325153,
|
|
"grad_norm": 36.80182572222518,
|
|
"learning_rate": 9.09090909090909e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2633,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 245
|
|
},
|
|
{
|
|
"epoch": 0.301840490797546,
|
|
"grad_norm": 36.10259081942398,
|
|
"learning_rate": 9.086776859504131e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3021,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.296875,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 246
|
|
},
|
|
{
|
|
"epoch": 0.3030674846625767,
|
|
"grad_norm": 37.745736662906765,
|
|
"learning_rate": 9.082644628099173e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.317,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 247
|
|
},
|
|
{
|
|
"epoch": 0.3042944785276074,
|
|
"grad_norm": 35.08194100895569,
|
|
"learning_rate": 9.078512396694214e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2352,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 248
|
|
},
|
|
{
|
|
"epoch": 0.30552147239263805,
|
|
"grad_norm": 39.937675917368395,
|
|
"learning_rate": 9.074380165289256e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -474.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.294,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 249
|
|
},
|
|
{
|
|
"epoch": 0.3067484662576687,
|
|
"grad_norm": 40.869769785212625,
|
|
"learning_rate": 9.070247933884297e-07,
|
|
"logits/chosen": -0.3359375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3549,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.3079754601226994,
|
|
"grad_norm": 40.52486042802848,
|
|
"learning_rate": 9.066115702479339e-07,
|
|
"logits/chosen": -0.02978515625,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3659,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 251
|
|
},
|
|
{
|
|
"epoch": 0.30920245398773005,
|
|
"grad_norm": 39.2129534588987,
|
|
"learning_rate": 9.061983471074379e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3449,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 252
|
|
},
|
|
{
|
|
"epoch": 0.3104294478527607,
|
|
"grad_norm": 32.24326164619934,
|
|
"learning_rate": 9.057851239669421e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2689,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.625,
|
|
"step": 253
|
|
},
|
|
{
|
|
"epoch": 0.3116564417177914,
|
|
"grad_norm": 42.0234467969428,
|
|
"learning_rate": 9.053719008264462e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3892,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 254
|
|
},
|
|
{
|
|
"epoch": 0.3128834355828221,
|
|
"grad_norm": 39.369143269027276,
|
|
"learning_rate": 9.049586776859504e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.375,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 255
|
|
},
|
|
{
|
|
"epoch": 0.3141104294478528,
|
|
"grad_norm": 33.202524065959715,
|
|
"learning_rate": 9.045454545454545e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2491,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.375,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 0.31533742331288345,
|
|
"grad_norm": 41.51571179035507,
|
|
"learning_rate": 9.041322314049586e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3173,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.265625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.0,
|
|
"step": 257
|
|
},
|
|
{
|
|
"epoch": 0.3165644171779141,
|
|
"grad_norm": 35.738504133116166,
|
|
"learning_rate": 9.037190082644628e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2726,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.921875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 258
|
|
},
|
|
{
|
|
"epoch": 0.3177914110429448,
|
|
"grad_norm": 34.59247141687721,
|
|
"learning_rate": 9.033057851239669e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2952,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 259
|
|
},
|
|
{
|
|
"epoch": 0.31901840490797545,
|
|
"grad_norm": 37.65821660368043,
|
|
"learning_rate": 9.028925619834711e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3336,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 0.3202453987730061,
|
|
"grad_norm": 40.17093324908733,
|
|
"learning_rate": 9.024793388429751e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3215,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 261
|
|
},
|
|
{
|
|
"epoch": 0.3214723926380368,
|
|
"grad_norm": 41.77028106523845,
|
|
"learning_rate": 9.020661157024793e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3204,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 262
|
|
},
|
|
{
|
|
"epoch": 0.3226993865030675,
|
|
"grad_norm": 45.33050032219676,
|
|
"learning_rate": 9.016528925619834e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3834,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 263
|
|
},
|
|
{
|
|
"epoch": 0.3239263803680982,
|
|
"grad_norm": 41.9279090830257,
|
|
"learning_rate": 9.012396694214876e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.34,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 264
|
|
},
|
|
{
|
|
"epoch": 0.32515337423312884,
|
|
"grad_norm": 47.31340400046535,
|
|
"learning_rate": 9.008264462809917e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2994,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 265
|
|
},
|
|
{
|
|
"epoch": 0.3263803680981595,
|
|
"grad_norm": 51.49399504173386,
|
|
"learning_rate": 9.004132231404959e-07,
|
|
"logits/chosen": -0.1220703125,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -320.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.464,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 0.423828125,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -4.125,
|
|
"step": 266
|
|
},
|
|
{
|
|
"epoch": 0.3276073619631902,
|
|
"grad_norm": 49.994468797582975,
|
|
"learning_rate": 9e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3743,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 267
|
|
},
|
|
{
|
|
"epoch": 0.32883435582822085,
|
|
"grad_norm": 31.9275570965626,
|
|
"learning_rate": 8.995867768595042e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2019,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 268
|
|
},
|
|
{
|
|
"epoch": 0.3300613496932515,
|
|
"grad_norm": 34.908633509533956,
|
|
"learning_rate": 8.991735537190083e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2272,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 269
|
|
},
|
|
{
|
|
"epoch": 0.3312883435582822,
|
|
"grad_norm": 42.11747091260519,
|
|
"learning_rate": 8.987603305785123e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3885,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 0.3325153374233129,
|
|
"grad_norm": 31.100938375184207,
|
|
"learning_rate": 8.983471074380164e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.262,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 271
|
|
},
|
|
{
|
|
"epoch": 0.3337423312883436,
|
|
"grad_norm": 36.40948876277434,
|
|
"learning_rate": 8.979338842975206e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.263,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.875,
|
|
"step": 272
|
|
},
|
|
{
|
|
"epoch": 0.33496932515337424,
|
|
"grad_norm": 43.97482452623596,
|
|
"learning_rate": 8.975206611570247e-07,
|
|
"logits/chosen": -0.025146484375,
|
|
"logits/rejected": -0.10498046875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.4597,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 273
|
|
},
|
|
{
|
|
"epoch": 0.3361963190184049,
|
|
"grad_norm": 38.58556750213493,
|
|
"learning_rate": 8.971074380165289e-07,
|
|
"logits/chosen": -0.04345703125,
|
|
"logits/rejected": -0.189453125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3517,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 274
|
|
},
|
|
{
|
|
"epoch": 0.3374233128834356,
|
|
"grad_norm": 36.32564369328471,
|
|
"learning_rate": 8.96694214876033e-07,
|
|
"logits/chosen": -0.1103515625,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3319,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 0.33865030674846625,
|
|
"grad_norm": 28.631593511584413,
|
|
"learning_rate": 8.962809917355371e-07,
|
|
"logits/chosen": -0.034912109375,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.295,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 276
|
|
},
|
|
{
|
|
"epoch": 0.3398773006134969,
|
|
"grad_norm": 30.563119880156428,
|
|
"learning_rate": 8.958677685950413e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2232,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8828125,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 277
|
|
},
|
|
{
|
|
"epoch": 0.3411042944785276,
|
|
"grad_norm": 43.29625763685435,
|
|
"learning_rate": 8.954545454545454e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3369,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.625,
|
|
"step": 278
|
|
},
|
|
{
|
|
"epoch": 0.3423312883435583,
|
|
"grad_norm": 23.231526168499595,
|
|
"learning_rate": 8.950413223140495e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.1765,
|
|
"rewards/accuracies": 0.9375,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 279
|
|
},
|
|
{
|
|
"epoch": 0.34355828220858897,
|
|
"grad_norm": 42.428969483206515,
|
|
"learning_rate": 8.946280991735536e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3827,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.875,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 0.34478527607361964,
|
|
"grad_norm": 37.34866118523232,
|
|
"learning_rate": 8.942148760330578e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3435,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 281
|
|
},
|
|
{
|
|
"epoch": 0.3460122699386503,
|
|
"grad_norm": 33.25917218474408,
|
|
"learning_rate": 8.938016528925619e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2298,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -5.0,
|
|
"step": 282
|
|
},
|
|
{
|
|
"epoch": 0.347239263803681,
|
|
"grad_norm": 34.34252527463677,
|
|
"learning_rate": 8.933884297520661e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2751,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.5,
|
|
"step": 283
|
|
},
|
|
{
|
|
"epoch": 0.34846625766871164,
|
|
"grad_norm": 38.92697687813909,
|
|
"learning_rate": 8.929752066115702e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2851,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 284
|
|
},
|
|
{
|
|
"epoch": 0.3496932515337423,
|
|
"grad_norm": 30.695326350223162,
|
|
"learning_rate": 8.925619834710744e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2414,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 285
|
|
},
|
|
{
|
|
"epoch": 0.350920245398773,
|
|
"grad_norm": 38.57430321053572,
|
|
"learning_rate": 8.921487603305785e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3505,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.0,
|
|
"step": 286
|
|
},
|
|
{
|
|
"epoch": 0.3521472392638037,
|
|
"grad_norm": 42.69834276963442,
|
|
"learning_rate": 8.917355371900827e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3256,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 287
|
|
},
|
|
{
|
|
"epoch": 0.35337423312883437,
|
|
"grad_norm": 36.1832160066718,
|
|
"learning_rate": 8.913223140495867e-07,
|
|
"logits/chosen": -0.036376953125,
|
|
"logits/rejected": -0.1865234375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3261,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 288
|
|
},
|
|
{
|
|
"epoch": 0.35460122699386504,
|
|
"grad_norm": 28.376305528750855,
|
|
"learning_rate": 8.909090909090909e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.1858,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 289
|
|
},
|
|
{
|
|
"epoch": 0.3558282208588957,
|
|
"grad_norm": 31.607934322430555,
|
|
"learning_rate": 8.90495867768595e-07,
|
|
"logits/chosen": -0.34375,
|
|
"logits/rejected": -0.48828125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.217,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 0.3570552147239264,
|
|
"grad_norm": 38.81396858549568,
|
|
"learning_rate": 8.900826446280992e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3515,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 291
|
|
},
|
|
{
|
|
"epoch": 0.35828220858895704,
|
|
"grad_norm": 37.738218399257526,
|
|
"learning_rate": 8.896694214876033e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2524,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 292
|
|
},
|
|
{
|
|
"epoch": 0.3595092024539877,
|
|
"grad_norm": 29.993709808838254,
|
|
"learning_rate": 8.892561983471074e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2178,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 293
|
|
},
|
|
{
|
|
"epoch": 0.36073619631901843,
|
|
"grad_norm": 26.82462113028186,
|
|
"learning_rate": 8.888429752066116e-07,
|
|
"logits/chosen": -0.4140625,
|
|
"logits/rejected": -0.5625,
|
|
"logps/chosen": -462.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2557,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 294
|
|
},
|
|
{
|
|
"epoch": 0.3619631901840491,
|
|
"grad_norm": 31.872360219352924,
|
|
"learning_rate": 8.884297520661157e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2396,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 295
|
|
},
|
|
{
|
|
"epoch": 0.36319018404907977,
|
|
"grad_norm": 40.252951579438076,
|
|
"learning_rate": 8.880165289256199e-07,
|
|
"logits/chosen": -0.0966796875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3687,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 296
|
|
},
|
|
{
|
|
"epoch": 0.36441717791411044,
|
|
"grad_norm": 31.43163856844169,
|
|
"learning_rate": 8.876033057851238e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2373,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 297
|
|
},
|
|
{
|
|
"epoch": 0.3656441717791411,
|
|
"grad_norm": 40.16169035445508,
|
|
"learning_rate": 8.87190082644628e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.2041015625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.322,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 298
|
|
},
|
|
{
|
|
"epoch": 0.36687116564417177,
|
|
"grad_norm": 44.752661502125484,
|
|
"learning_rate": 8.867768595041321e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3433,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 299
|
|
},
|
|
{
|
|
"epoch": 0.36809815950920244,
|
|
"grad_norm": 40.61630087470012,
|
|
"learning_rate": 8.863636363636363e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3683,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.3693251533742331,
|
|
"grad_norm": 34.184057413647636,
|
|
"learning_rate": 8.859504132231404e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3007,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 301
|
|
},
|
|
{
|
|
"epoch": 0.37055214723926383,
|
|
"grad_norm": 28.52262909170181,
|
|
"learning_rate": 8.855371900826446e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2335,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 302
|
|
},
|
|
{
|
|
"epoch": 0.3717791411042945,
|
|
"grad_norm": 45.58508428540745,
|
|
"learning_rate": 8.851239669421487e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3977,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 303
|
|
},
|
|
{
|
|
"epoch": 0.37300613496932516,
|
|
"grad_norm": 29.98788371317914,
|
|
"learning_rate": 8.847107438016529e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2267,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 304
|
|
},
|
|
{
|
|
"epoch": 0.37423312883435583,
|
|
"grad_norm": 35.17479410559687,
|
|
"learning_rate": 8.84297520661157e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3025,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 305
|
|
},
|
|
{
|
|
"epoch": 0.3754601226993865,
|
|
"grad_norm": 39.17340101734157,
|
|
"learning_rate": 8.838842975206611e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2699,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 306
|
|
},
|
|
{
|
|
"epoch": 0.37668711656441717,
|
|
"grad_norm": 37.77108585187058,
|
|
"learning_rate": 8.834710743801652e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3981,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 307
|
|
},
|
|
{
|
|
"epoch": 0.37791411042944784,
|
|
"grad_norm": 37.13471968173906,
|
|
"learning_rate": 8.830578512396694e-07,
|
|
"logits/chosen": -0.109375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3474,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.375,
|
|
"step": 308
|
|
},
|
|
{
|
|
"epoch": 0.3791411042944785,
|
|
"grad_norm": 32.80133096174931,
|
|
"learning_rate": 8.826446280991735e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2405,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 309
|
|
},
|
|
{
|
|
"epoch": 0.3803680981595092,
|
|
"grad_norm": 36.675395105919584,
|
|
"learning_rate": 8.822314049586776e-07,
|
|
"logits/chosen": -0.061767578125,
|
|
"logits/rejected": -0.158203125,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3773,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.15234375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.875,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 0.3815950920245399,
|
|
"grad_norm": 33.25381441708521,
|
|
"learning_rate": 8.818181818181818e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2785,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 311
|
|
},
|
|
{
|
|
"epoch": 0.38282208588957056,
|
|
"grad_norm": 42.64167876910758,
|
|
"learning_rate": 8.814049586776859e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.236328125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.4387,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 312
|
|
},
|
|
{
|
|
"epoch": 0.38404907975460123,
|
|
"grad_norm": 35.315697341783846,
|
|
"learning_rate": 8.809917355371901e-07,
|
|
"logits/chosen": -0.109375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3074,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 7.5625,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 313
|
|
},
|
|
{
|
|
"epoch": 0.3852760736196319,
|
|
"grad_norm": 28.577677082306792,
|
|
"learning_rate": 8.805785123966942e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.1843,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.625,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 314
|
|
},
|
|
{
|
|
"epoch": 0.38650306748466257,
|
|
"grad_norm": 34.67651692357887,
|
|
"learning_rate": 8.801652892561983e-07,
|
|
"logits/chosen": -0.03564453125,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2637,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.25,
|
|
"step": 315
|
|
},
|
|
{
|
|
"epoch": 0.38773006134969323,
|
|
"grad_norm": 31.070595317274808,
|
|
"learning_rate": 8.797520661157024e-07,
|
|
"logits/chosen": -0.04541015625,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2748,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 316
|
|
},
|
|
{
|
|
"epoch": 0.3889570552147239,
|
|
"grad_norm": 39.24846705876142,
|
|
"learning_rate": 8.793388429752066e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2469,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 317
|
|
},
|
|
{
|
|
"epoch": 0.3901840490797546,
|
|
"grad_norm": 41.29363946363224,
|
|
"learning_rate": 8.789256198347107e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3595,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.0,
|
|
"step": 318
|
|
},
|
|
{
|
|
"epoch": 0.3914110429447853,
|
|
"grad_norm": 21.922537381657932,
|
|
"learning_rate": 8.785123966942149e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.1571,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 7.9375,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 319
|
|
},
|
|
{
|
|
"epoch": 0.39263803680981596,
|
|
"grad_norm": 37.85813864992483,
|
|
"learning_rate": 8.78099173553719e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.3097,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 0.39386503067484663,
|
|
"grad_norm": 44.42516577854287,
|
|
"learning_rate": 8.776859504132232e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.3335,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 321
|
|
},
|
|
{
|
|
"epoch": 0.3950920245398773,
|
|
"grad_norm": 29.9172278345542,
|
|
"learning_rate": 8.772727272727273e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2506,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 322
|
|
},
|
|
{
|
|
"epoch": 0.39631901840490796,
|
|
"grad_norm": 38.156948353448634,
|
|
"learning_rate": 8.768595041322315e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.4208,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 323
|
|
},
|
|
{
|
|
"epoch": 0.39754601226993863,
|
|
"grad_norm": 32.767921456046125,
|
|
"learning_rate": 8.764462809917354e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3067,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.875,
|
|
"step": 324
|
|
},
|
|
{
|
|
"epoch": 0.3987730061349693,
|
|
"grad_norm": 40.13577337526863,
|
|
"learning_rate": 8.760330578512396e-07,
|
|
"logits/chosen": -0.119140625,
|
|
"logits/rejected": -0.1962890625,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.4481,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.6796875,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"grad_norm": 31.397049512750392,
|
|
"learning_rate": 8.756198347107437e-07,
|
|
"logits/chosen": -0.296875,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2169,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 326
|
|
},
|
|
{
|
|
"epoch": 0.4012269938650307,
|
|
"grad_norm": 43.09847438786523,
|
|
"learning_rate": 8.752066115702479e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.4159,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 327
|
|
},
|
|
{
|
|
"epoch": 0.40245398773006136,
|
|
"grad_norm": 38.59522801863003,
|
|
"learning_rate": 8.74793388429752e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.3241,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 328
|
|
},
|
|
{
|
|
"epoch": 0.403680981595092,
|
|
"grad_norm": 35.517969492099304,
|
|
"learning_rate": 8.743801652892561e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3229,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.0,
|
|
"rewards/margins": 7.625,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 329
|
|
},
|
|
{
|
|
"epoch": 0.4049079754601227,
|
|
"grad_norm": 32.81462048360016,
|
|
"learning_rate": 8.739669421487603e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2512,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 7.25,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 0.40613496932515336,
|
|
"grad_norm": 31.55365723251363,
|
|
"learning_rate": 8.735537190082644e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2528,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 331
|
|
},
|
|
{
|
|
"epoch": 0.40736196319018403,
|
|
"grad_norm": 25.058536630302516,
|
|
"learning_rate": 8.731404958677685e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2033,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.984375,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 332
|
|
},
|
|
{
|
|
"epoch": 0.4085889570552147,
|
|
"grad_norm": 37.27877790438504,
|
|
"learning_rate": 8.727272727272726e-07,
|
|
"logits/chosen": -0.087890625,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.332,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 333
|
|
},
|
|
{
|
|
"epoch": 0.4098159509202454,
|
|
"grad_norm": 32.893281250723994,
|
|
"learning_rate": 8.723140495867768e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2882,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 334
|
|
},
|
|
{
|
|
"epoch": 0.4110429447852761,
|
|
"grad_norm": 37.617815711991454,
|
|
"learning_rate": 8.719008264462809e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2311,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 335
|
|
},
|
|
{
|
|
"epoch": 0.41226993865030676,
|
|
"grad_norm": 34.178158756977034,
|
|
"learning_rate": 8.714876033057851e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3112,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 336
|
|
},
|
|
{
|
|
"epoch": 0.4134969325153374,
|
|
"grad_norm": 43.74423728378528,
|
|
"learning_rate": 8.710743801652892e-07,
|
|
"logits/chosen": -0.0673828125,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.4057,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 337
|
|
},
|
|
{
|
|
"epoch": 0.4147239263803681,
|
|
"grad_norm": 34.34678654972963,
|
|
"learning_rate": 8.706611570247934e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3277,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.875,
|
|
"step": 338
|
|
},
|
|
{
|
|
"epoch": 0.41595092024539876,
|
|
"grad_norm": 38.33964943892026,
|
|
"learning_rate": 8.702479338842975e-07,
|
|
"logits/chosen": -0.046630859375,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3247,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.875,
|
|
"step": 339
|
|
},
|
|
{
|
|
"epoch": 0.4171779141104294,
|
|
"grad_norm": 42.20123185254217,
|
|
"learning_rate": 8.698347107438017e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3163,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 0.41840490797546015,
|
|
"grad_norm": 30.617090577606344,
|
|
"learning_rate": 8.694214876033057e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3055,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 341
|
|
},
|
|
{
|
|
"epoch": 0.4196319018404908,
|
|
"grad_norm": 36.98686744550473,
|
|
"learning_rate": 8.690082644628099e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3403,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 342
|
|
},
|
|
{
|
|
"epoch": 0.4208588957055215,
|
|
"grad_norm": 38.00637794749286,
|
|
"learning_rate": 8.68595041322314e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3859,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 343
|
|
},
|
|
{
|
|
"epoch": 0.42208588957055215,
|
|
"grad_norm": 39.71440557055468,
|
|
"learning_rate": 8.681818181818182e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3756,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 344
|
|
},
|
|
{
|
|
"epoch": 0.4233128834355828,
|
|
"grad_norm": 35.94521943416743,
|
|
"learning_rate": 8.677685950413223e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3392,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.75,
|
|
"step": 345
|
|
},
|
|
{
|
|
"epoch": 0.4245398773006135,
|
|
"grad_norm": 38.26590405063427,
|
|
"learning_rate": 8.673553719008264e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.39,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 346
|
|
},
|
|
{
|
|
"epoch": 0.42576687116564416,
|
|
"grad_norm": 28.157756450395475,
|
|
"learning_rate": 8.669421487603306e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2541,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 347
|
|
},
|
|
{
|
|
"epoch": 0.4269938650306748,
|
|
"grad_norm": 30.247474673307313,
|
|
"learning_rate": 8.665289256198347e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2592,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.9140625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -6.0,
|
|
"step": 348
|
|
},
|
|
{
|
|
"epoch": 0.42822085889570555,
|
|
"grad_norm": 40.56583556395029,
|
|
"learning_rate": 8.661157024793389e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3461,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 349
|
|
},
|
|
{
|
|
"epoch": 0.4294478527607362,
|
|
"grad_norm": 37.32296255724777,
|
|
"learning_rate": 8.657024793388428e-07,
|
|
"logits/chosen": -0.0595703125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3201,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.625,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 0.4306748466257669,
|
|
"grad_norm": 38.61905817763815,
|
|
"learning_rate": 8.65289256198347e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3683,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -5.0,
|
|
"step": 351
|
|
},
|
|
{
|
|
"epoch": 0.43190184049079755,
|
|
"grad_norm": 32.77930853827487,
|
|
"learning_rate": 8.648760330578511e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3505,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 352
|
|
},
|
|
{
|
|
"epoch": 0.4331288343558282,
|
|
"grad_norm": 28.429883368705962,
|
|
"learning_rate": 8.644628099173553e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2385,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 353
|
|
},
|
|
{
|
|
"epoch": 0.4343558282208589,
|
|
"grad_norm": 32.10596291611156,
|
|
"learning_rate": 8.640495867768594e-07,
|
|
"logits/chosen": -0.1083984375,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.301,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.0,
|
|
"step": 354
|
|
},
|
|
{
|
|
"epoch": 0.43558282208588955,
|
|
"grad_norm": 35.70256329945714,
|
|
"learning_rate": 8.636363636363636e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3752,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 355
|
|
},
|
|
{
|
|
"epoch": 0.4368098159509202,
|
|
"grad_norm": 33.540082224759566,
|
|
"learning_rate": 8.632231404958677e-07,
|
|
"logits/chosen": -0.05908203125,
|
|
"logits/rejected": -0.2041015625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.351,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 356
|
|
},
|
|
{
|
|
"epoch": 0.43803680981595094,
|
|
"grad_norm": 32.54050040447977,
|
|
"learning_rate": 8.628099173553719e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.470703125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3107,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 357
|
|
},
|
|
{
|
|
"epoch": 0.4392638036809816,
|
|
"grad_norm": 34.06624107320208,
|
|
"learning_rate": 8.62396694214876e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2835,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 358
|
|
},
|
|
{
|
|
"epoch": 0.4404907975460123,
|
|
"grad_norm": 36.767095986046016,
|
|
"learning_rate": 8.619834710743801e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3176,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 359
|
|
},
|
|
{
|
|
"epoch": 0.44171779141104295,
|
|
"grad_norm": 34.37658092994157,
|
|
"learning_rate": 8.615702479338842e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.27,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 0.4429447852760736,
|
|
"grad_norm": 32.664077801869844,
|
|
"learning_rate": 8.611570247933884e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2781,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 361
|
|
},
|
|
{
|
|
"epoch": 0.4441717791411043,
|
|
"grad_norm": 31.9733261958367,
|
|
"learning_rate": 8.607438016528925e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2931,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 362
|
|
},
|
|
{
|
|
"epoch": 0.44539877300613495,
|
|
"grad_norm": 29.251958446930967,
|
|
"learning_rate": 8.603305785123967e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2371,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 363
|
|
},
|
|
{
|
|
"epoch": 0.4466257668711656,
|
|
"grad_norm": 37.57398787735362,
|
|
"learning_rate": 8.599173553719008e-07,
|
|
"logits/chosen": -0.042724609375,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3235,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 364
|
|
},
|
|
{
|
|
"epoch": 0.44785276073619634,
|
|
"grad_norm": 29.870035318324586,
|
|
"learning_rate": 8.595041322314049e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2644,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 365
|
|
},
|
|
{
|
|
"epoch": 0.449079754601227,
|
|
"grad_norm": 34.90800717205546,
|
|
"learning_rate": 8.590909090909091e-07,
|
|
"logits/chosen": -0.08056640625,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3873,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 366
|
|
},
|
|
{
|
|
"epoch": 0.4503067484662577,
|
|
"grad_norm": 29.571048038433712,
|
|
"learning_rate": 8.586776859504132e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2315,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.5,
|
|
"step": 367
|
|
},
|
|
{
|
|
"epoch": 0.45153374233128835,
|
|
"grad_norm": 42.056035274022705,
|
|
"learning_rate": 8.582644628099173e-07,
|
|
"logits/chosen": -0.034423828125,
|
|
"logits/rejected": -0.0859375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3686,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 368
|
|
},
|
|
{
|
|
"epoch": 0.452760736196319,
|
|
"grad_norm": 38.8917843400358,
|
|
"learning_rate": 8.578512396694214e-07,
|
|
"logits/chosen": -0.05712890625,
|
|
"logits/rejected": -0.1689453125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3427,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.62109375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 369
|
|
},
|
|
{
|
|
"epoch": 0.4539877300613497,
|
|
"grad_norm": 35.30535390014327,
|
|
"learning_rate": 8.574380165289256e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3331,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.375,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 0.45521472392638035,
|
|
"grad_norm": 31.840615729545142,
|
|
"learning_rate": 8.570247933884297e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2692,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 371
|
|
},
|
|
{
|
|
"epoch": 0.456441717791411,
|
|
"grad_norm": 29.955978159088186,
|
|
"learning_rate": 8.566115702479339e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2245,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 372
|
|
},
|
|
{
|
|
"epoch": 0.45766871165644174,
|
|
"grad_norm": 36.82441545636138,
|
|
"learning_rate": 8.56198347107438e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.37,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.828125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 373
|
|
},
|
|
{
|
|
"epoch": 0.4588957055214724,
|
|
"grad_norm": 29.01028759814529,
|
|
"learning_rate": 8.557851239669422e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2316,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.0625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 374
|
|
},
|
|
{
|
|
"epoch": 0.4601226993865031,
|
|
"grad_norm": 34.54866122563966,
|
|
"learning_rate": 8.553719008264463e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2607,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 0.46134969325153374,
|
|
"grad_norm": 33.10026205924125,
|
|
"learning_rate": 8.549586776859505e-07,
|
|
"logits/chosen": -0.1044921875,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2614,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.25,
|
|
"step": 376
|
|
},
|
|
{
|
|
"epoch": 0.4625766871165644,
|
|
"grad_norm": 35.75040093231158,
|
|
"learning_rate": 8.545454545454544e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.24609375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3217,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 377
|
|
},
|
|
{
|
|
"epoch": 0.4638036809815951,
|
|
"grad_norm": 37.937967119942094,
|
|
"learning_rate": 8.541322314049586e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.332,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.0,
|
|
"step": 378
|
|
},
|
|
{
|
|
"epoch": 0.46503067484662575,
|
|
"grad_norm": 32.72768732817253,
|
|
"learning_rate": 8.537190082644627e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2597,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 379
|
|
},
|
|
{
|
|
"epoch": 0.4662576687116564,
|
|
"grad_norm": 33.93477462013243,
|
|
"learning_rate": 8.533057851239669e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2747,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 0.46748466257668714,
|
|
"grad_norm": 33.523029253094535,
|
|
"learning_rate": 8.52892561983471e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3434,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.125,
|
|
"step": 381
|
|
},
|
|
{
|
|
"epoch": 0.4687116564417178,
|
|
"grad_norm": 30.950990880904673,
|
|
"learning_rate": 8.524793388429751e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.291,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 382
|
|
},
|
|
{
|
|
"epoch": 0.4699386503067485,
|
|
"grad_norm": 37.335141994131185,
|
|
"learning_rate": 8.520661157024793e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.3653,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 383
|
|
},
|
|
{
|
|
"epoch": 0.47116564417177914,
|
|
"grad_norm": 35.708837972051896,
|
|
"learning_rate": 8.516528925619834e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2715,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.546875,
|
|
"rewards/margins": 7.5,
|
|
"rewards/rejected": -6.9375,
|
|
"step": 384
|
|
},
|
|
{
|
|
"epoch": 0.4723926380368098,
|
|
"grad_norm": 38.095621644948935,
|
|
"learning_rate": 8.512396694214876e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3249,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.279296875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 385
|
|
},
|
|
{
|
|
"epoch": 0.4736196319018405,
|
|
"grad_norm": 42.56085647887984,
|
|
"learning_rate": 8.508264462809916e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3519,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": -0.337890625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 386
|
|
},
|
|
{
|
|
"epoch": 0.47484662576687114,
|
|
"grad_norm": 63.07829836210694,
|
|
"learning_rate": 8.504132231404958e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3377,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.87890625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 387
|
|
},
|
|
{
|
|
"epoch": 0.47607361963190187,
|
|
"grad_norm": 44.508149553651,
|
|
"learning_rate": 8.499999999999999e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.4576,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.318359375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 388
|
|
},
|
|
{
|
|
"epoch": 0.47730061349693254,
|
|
"grad_norm": 43.15219698694189,
|
|
"learning_rate": 8.495867768595041e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4252,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 389
|
|
},
|
|
{
|
|
"epoch": 0.4785276073619632,
|
|
"grad_norm": 43.3947576259124,
|
|
"learning_rate": 8.491735537190082e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.4287,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.10595703125,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 0.47975460122699387,
|
|
"grad_norm": 42.71401974468104,
|
|
"learning_rate": 8.487603305785124e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4278,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 391
|
|
},
|
|
{
|
|
"epoch": 0.48098159509202454,
|
|
"grad_norm": 31.928442662001206,
|
|
"learning_rate": 8.483471074380165e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2866,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 392
|
|
},
|
|
{
|
|
"epoch": 0.4822085889570552,
|
|
"grad_norm": 34.01837933004929,
|
|
"learning_rate": 8.479338842975207e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2862,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 393
|
|
},
|
|
{
|
|
"epoch": 0.4834355828220859,
|
|
"grad_norm": 34.069017331837635,
|
|
"learning_rate": 8.475206611570248e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.255,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 394
|
|
},
|
|
{
|
|
"epoch": 0.48466257668711654,
|
|
"grad_norm": 36.10911763794192,
|
|
"learning_rate": 8.471074380165289e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3367,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 395
|
|
},
|
|
{
|
|
"epoch": 0.48588957055214727,
|
|
"grad_norm": 24.87975181789429,
|
|
"learning_rate": 8.46694214876033e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2201,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.439453125,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -6.0,
|
|
"step": 396
|
|
},
|
|
{
|
|
"epoch": 0.48711656441717793,
|
|
"grad_norm": 37.31242821688829,
|
|
"learning_rate": 8.462809917355372e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3358,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -6.0,
|
|
"step": 397
|
|
},
|
|
{
|
|
"epoch": 0.4883435582822086,
|
|
"grad_norm": 38.435432269740204,
|
|
"learning_rate": 8.458677685950413e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3625,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 398
|
|
},
|
|
{
|
|
"epoch": 0.48957055214723927,
|
|
"grad_norm": 31.226281146954268,
|
|
"learning_rate": 8.454545454545454e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2934,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 399
|
|
},
|
|
{
|
|
"epoch": 0.49079754601226994,
|
|
"grad_norm": 45.39512753724527,
|
|
"learning_rate": 8.450413223140496e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.4427,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.4920245398773006,
|
|
"grad_norm": 30.28556425067376,
|
|
"learning_rate": 8.446280991735537e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.2465,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 401
|
|
},
|
|
{
|
|
"epoch": 0.49325153374233127,
|
|
"grad_norm": 34.97541893034036,
|
|
"learning_rate": 8.442148760330579e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3267,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 402
|
|
},
|
|
{
|
|
"epoch": 0.49447852760736194,
|
|
"grad_norm": 35.34876316760721,
|
|
"learning_rate": 8.43801652892562e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.31,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.71875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 403
|
|
},
|
|
{
|
|
"epoch": 0.49570552147239266,
|
|
"grad_norm": 30.378373289318976,
|
|
"learning_rate": 8.43388429752066e-07,
|
|
"logits/chosen": -0.11865234375,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2139,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 404
|
|
},
|
|
{
|
|
"epoch": 0.49693251533742333,
|
|
"grad_norm": 38.10870942161704,
|
|
"learning_rate": 8.429752066115701e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3558,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.375,
|
|
"step": 405
|
|
},
|
|
{
|
|
"epoch": 0.498159509202454,
|
|
"grad_norm": 41.99213740680795,
|
|
"learning_rate": 8.425619834710743e-07,
|
|
"logits/chosen": -0.09521484375,
|
|
"logits/rejected": -0.1875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.44,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 406
|
|
},
|
|
{
|
|
"epoch": 0.49938650306748467,
|
|
"grad_norm": 41.75199039490215,
|
|
"learning_rate": 8.421487603305784e-07,
|
|
"logits/chosen": -0.08544921875,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3833,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.125,
|
|
"step": 407
|
|
},
|
|
{
|
|
"epoch": 0.5006134969325153,
|
|
"grad_norm": 46.46791360512976,
|
|
"learning_rate": 8.417355371900826e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.4566,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 408
|
|
},
|
|
{
|
|
"epoch": 0.501840490797546,
|
|
"grad_norm": 44.673203673536904,
|
|
"learning_rate": 8.413223140495867e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3461,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 409
|
|
},
|
|
{
|
|
"epoch": 0.5030674846625767,
|
|
"grad_norm": 38.196100384677806,
|
|
"learning_rate": 8.409090909090909e-07,
|
|
"logits/chosen": -0.1025390625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3237,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.9296875,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 0.5042944785276073,
|
|
"grad_norm": 36.65382709836008,
|
|
"learning_rate": 8.40495867768595e-07,
|
|
"logits/chosen": -0.09326171875,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2836,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.6796875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 411
|
|
},
|
|
{
|
|
"epoch": 0.505521472392638,
|
|
"grad_norm": 46.13136333228446,
|
|
"learning_rate": 8.400826446280992e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3856,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 412
|
|
},
|
|
{
|
|
"epoch": 0.5067484662576687,
|
|
"grad_norm": 40.318779960108294,
|
|
"learning_rate": 8.396694214876032e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3462,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 413
|
|
},
|
|
{
|
|
"epoch": 0.5079754601226993,
|
|
"grad_norm": 34.87712228181845,
|
|
"learning_rate": 8.392561983471074e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2741,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 414
|
|
},
|
|
{
|
|
"epoch": 0.50920245398773,
|
|
"grad_norm": 34.161638850062126,
|
|
"learning_rate": 8.388429752066115e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2949,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 415
|
|
},
|
|
{
|
|
"epoch": 0.5104294478527608,
|
|
"grad_norm": 28.62101759801655,
|
|
"learning_rate": 8.384297520661157e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.1788,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.125,
|
|
"step": 416
|
|
},
|
|
{
|
|
"epoch": 0.5116564417177915,
|
|
"grad_norm": 32.11517958415963,
|
|
"learning_rate": 8.380165289256198e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.302,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -5.75,
|
|
"step": 417
|
|
},
|
|
{
|
|
"epoch": 0.5128834355828221,
|
|
"grad_norm": 36.91670370496033,
|
|
"learning_rate": 8.376033057851239e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3162,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 418
|
|
},
|
|
{
|
|
"epoch": 0.5141104294478528,
|
|
"grad_norm": 29.74220467370693,
|
|
"learning_rate": 8.371900826446281e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2755,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 419
|
|
},
|
|
{
|
|
"epoch": 0.5153374233128835,
|
|
"grad_norm": 35.53103540854566,
|
|
"learning_rate": 8.367768595041322e-07,
|
|
"logits/chosen": -0.08837890625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.335,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 0.5165644171779141,
|
|
"grad_norm": 32.97503154682443,
|
|
"learning_rate": 8.363636363636363e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3545,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 421
|
|
},
|
|
{
|
|
"epoch": 0.5177914110429448,
|
|
"grad_norm": 38.11996274683729,
|
|
"learning_rate": 8.359504132231404e-07,
|
|
"logits/chosen": -0.0546875,
|
|
"logits/rejected": -0.2158203125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3392,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 422
|
|
},
|
|
{
|
|
"epoch": 0.5190184049079755,
|
|
"grad_norm": 39.06478111720593,
|
|
"learning_rate": 8.355371900826446e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3845,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 423
|
|
},
|
|
{
|
|
"epoch": 0.5202453987730061,
|
|
"grad_norm": 35.69529736649316,
|
|
"learning_rate": 8.351239669421487e-07,
|
|
"logits/chosen": -0.09619140625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3743,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 424
|
|
},
|
|
{
|
|
"epoch": 0.5214723926380368,
|
|
"grad_norm": 41.60888435807761,
|
|
"learning_rate": 8.347107438016529e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.4486,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 0.5226993865030675,
|
|
"grad_norm": 42.36087992828029,
|
|
"learning_rate": 8.34297520661157e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3332,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 426
|
|
},
|
|
{
|
|
"epoch": 0.5239263803680981,
|
|
"grad_norm": 77.1683428566809,
|
|
"learning_rate": 8.338842975206612e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3382,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 427
|
|
},
|
|
{
|
|
"epoch": 0.5251533742331288,
|
|
"grad_norm": 35.0971325671411,
|
|
"learning_rate": 8.334710743801653e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3745,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.546875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 428
|
|
},
|
|
{
|
|
"epoch": 0.5263803680981595,
|
|
"grad_norm": 30.4960218875002,
|
|
"learning_rate": 8.330578512396695e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.2397,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.275390625,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -7.125,
|
|
"step": 429
|
|
},
|
|
{
|
|
"epoch": 0.5276073619631901,
|
|
"grad_norm": 27.007644146694663,
|
|
"learning_rate": 8.326446280991735e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2181,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.80859375,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 0.5288343558282208,
|
|
"grad_norm": 46.10911626487466,
|
|
"learning_rate": 8.322314049586777e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.4699,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.337890625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 431
|
|
},
|
|
{
|
|
"epoch": 0.5300613496932516,
|
|
"grad_norm": 39.706928116253934,
|
|
"learning_rate": 8.318181818181817e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.1083984375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3755,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 432
|
|
},
|
|
{
|
|
"epoch": 0.5312883435582823,
|
|
"grad_norm": 44.60896484929849,
|
|
"learning_rate": 8.314049586776859e-07,
|
|
"logits/chosen": -0.04736328125,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3756,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.2353515625,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 433
|
|
},
|
|
{
|
|
"epoch": 0.5325153374233129,
|
|
"grad_norm": 39.167901096522954,
|
|
"learning_rate": 8.3099173553719e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2737,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.34765625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 434
|
|
},
|
|
{
|
|
"epoch": 0.5337423312883436,
|
|
"grad_norm": 41.42002661257632,
|
|
"learning_rate": 8.305785123966941e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.349,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.93359375,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 435
|
|
},
|
|
{
|
|
"epoch": 0.5349693251533743,
|
|
"grad_norm": 36.28330999120649,
|
|
"learning_rate": 8.301652892561983e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.231,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 436
|
|
},
|
|
{
|
|
"epoch": 0.5361963190184049,
|
|
"grad_norm": 29.931742040479904,
|
|
"learning_rate": 8.297520661157024e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2471,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 437
|
|
},
|
|
{
|
|
"epoch": 0.5374233128834356,
|
|
"grad_norm": 40.51322990584841,
|
|
"learning_rate": 8.293388429752066e-07,
|
|
"logits/chosen": -0.10546875,
|
|
"logits/rejected": -0.2060546875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.4088,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 438
|
|
},
|
|
{
|
|
"epoch": 0.5386503067484663,
|
|
"grad_norm": 36.14747773969047,
|
|
"learning_rate": 8.289256198347106e-07,
|
|
"logits/chosen": -0.057373046875,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2942,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.59765625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 439
|
|
},
|
|
{
|
|
"epoch": 0.5398773006134969,
|
|
"grad_norm": 47.70699134185717,
|
|
"learning_rate": 8.285123966942148e-07,
|
|
"logits/chosen": -0.099609375,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.4004,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.169921875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 0.5411042944785276,
|
|
"grad_norm": 37.12029582406457,
|
|
"learning_rate": 8.280991735537189e-07,
|
|
"logits/chosen": -0.03271484375,
|
|
"logits/rejected": -0.14453125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3617,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.4609375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 441
|
|
},
|
|
{
|
|
"epoch": 0.5423312883435583,
|
|
"grad_norm": 36.833773823132546,
|
|
"learning_rate": 8.276859504132231e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3691,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 442
|
|
},
|
|
{
|
|
"epoch": 0.5435582822085889,
|
|
"grad_norm": 38.764256019654724,
|
|
"learning_rate": 8.272727272727272e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3263,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 443
|
|
},
|
|
{
|
|
"epoch": 0.5447852760736196,
|
|
"grad_norm": 36.57479438910318,
|
|
"learning_rate": 8.268595041322314e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2696,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 444
|
|
},
|
|
{
|
|
"epoch": 0.5460122699386503,
|
|
"grad_norm": 37.423432013581284,
|
|
"learning_rate": 8.264462809917355e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.4224,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.5,
|
|
"step": 445
|
|
},
|
|
{
|
|
"epoch": 0.5472392638036809,
|
|
"grad_norm": 32.85951175366752,
|
|
"learning_rate": 8.260330578512397e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.261,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 446
|
|
},
|
|
{
|
|
"epoch": 0.5484662576687117,
|
|
"grad_norm": 30.593381836628694,
|
|
"learning_rate": 8.256198347107438e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2775,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 7.5625,
|
|
"rewards/rejected": -5.75,
|
|
"step": 447
|
|
},
|
|
{
|
|
"epoch": 0.5496932515337424,
|
|
"grad_norm": 37.96919240385222,
|
|
"learning_rate": 8.252066115702479e-07,
|
|
"logits/chosen": -0.076171875,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3504,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.75390625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 448
|
|
},
|
|
{
|
|
"epoch": 0.550920245398773,
|
|
"grad_norm": 32.94147744887095,
|
|
"learning_rate": 8.24793388429752e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8359375,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 449
|
|
},
|
|
{
|
|
"epoch": 0.5521472392638037,
|
|
"grad_norm": 32.316931210250736,
|
|
"learning_rate": 8.243801652892562e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2695,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 0.5533742331288344,
|
|
"grad_norm": 32.65287065671748,
|
|
"learning_rate": 8.239669421487603e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3194,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 451
|
|
},
|
|
{
|
|
"epoch": 0.554601226993865,
|
|
"grad_norm": 34.48235820988141,
|
|
"learning_rate": 8.235537190082645e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.259,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.34375,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 452
|
|
},
|
|
{
|
|
"epoch": 0.5558282208588957,
|
|
"grad_norm": 28.400538678800142,
|
|
"learning_rate": 8.231404958677686e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2402,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.515625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 453
|
|
},
|
|
{
|
|
"epoch": 0.5570552147239264,
|
|
"grad_norm": 33.9480760163604,
|
|
"learning_rate": 8.227272727272727e-07,
|
|
"logits/chosen": -0.078125,
|
|
"logits/rejected": -0.154296875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.285,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 454
|
|
},
|
|
{
|
|
"epoch": 0.558282208588957,
|
|
"grad_norm": 30.854904595369923,
|
|
"learning_rate": 8.223140495867769e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2265,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.921875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 455
|
|
},
|
|
{
|
|
"epoch": 0.5595092024539877,
|
|
"grad_norm": 39.11103336894796,
|
|
"learning_rate": 8.21900826446281e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.1845703125,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.34,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 456
|
|
},
|
|
{
|
|
"epoch": 0.5607361963190184,
|
|
"grad_norm": 34.29324527378237,
|
|
"learning_rate": 8.214876033057851e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3053,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 457
|
|
},
|
|
{
|
|
"epoch": 0.561963190184049,
|
|
"grad_norm": 32.239574863581325,
|
|
"learning_rate": 8.210743801652891e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2319,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 458
|
|
},
|
|
{
|
|
"epoch": 0.5631901840490797,
|
|
"grad_norm": 33.179929411101874,
|
|
"learning_rate": 8.206611570247933e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2898,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.5,
|
|
"step": 459
|
|
},
|
|
{
|
|
"epoch": 0.5644171779141104,
|
|
"grad_norm": 33.55136315479753,
|
|
"learning_rate": 8.202479338842974e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.286,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 0.5656441717791411,
|
|
"grad_norm": 36.94174797842278,
|
|
"learning_rate": 8.198347107438016e-07,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 461
|
|
},
|
|
{
|
|
"epoch": 0.5668711656441717,
|
|
"grad_norm": 39.294473057532144,
|
|
"learning_rate": 8.194214876033057e-07,
|
|
"logits/chosen": -0.0751953125,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3504,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.38671875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 462
|
|
},
|
|
{
|
|
"epoch": 0.5680981595092025,
|
|
"grad_norm": 36.351270220398256,
|
|
"learning_rate": 8.190082644628099e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2884,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.875,
|
|
"step": 463
|
|
},
|
|
{
|
|
"epoch": 0.5693251533742332,
|
|
"grad_norm": 39.910546519558004,
|
|
"learning_rate": 8.18595041322314e-07,
|
|
"logits/chosen": -0.0634765625,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.312,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 464
|
|
},
|
|
{
|
|
"epoch": 0.5705521472392638,
|
|
"grad_norm": 25.392205352978074,
|
|
"learning_rate": 8.181818181818182e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.1799,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 7.5625,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 465
|
|
},
|
|
{
|
|
"epoch": 0.5717791411042945,
|
|
"grad_norm": 38.85817198852564,
|
|
"learning_rate": 8.177685950413222e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3154,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.44921875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 466
|
|
},
|
|
{
|
|
"epoch": 0.5730061349693252,
|
|
"grad_norm": 37.87083062435175,
|
|
"learning_rate": 8.173553719008264e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3243,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 467
|
|
},
|
|
{
|
|
"epoch": 0.5742331288343558,
|
|
"grad_norm": 39.325632460733274,
|
|
"learning_rate": 8.169421487603305e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3072,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -6.25,
|
|
"step": 468
|
|
},
|
|
{
|
|
"epoch": 0.5754601226993865,
|
|
"grad_norm": 28.14882814674194,
|
|
"learning_rate": 8.165289256198347e-07,
|
|
"logits/chosen": -0.08740234375,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2011,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 469
|
|
},
|
|
{
|
|
"epoch": 0.5766871165644172,
|
|
"grad_norm": 33.58329571055999,
|
|
"learning_rate": 8.161157024793388e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.1943359375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3149,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.875,
|
|
"step": 470
|
|
},
|
|
{
|
|
"epoch": 0.5779141104294478,
|
|
"grad_norm": 33.52640840566966,
|
|
"learning_rate": 8.157024793388429e-07,
|
|
"logits/chosen": 0.002197265625,
|
|
"logits/rejected": -0.1650390625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3098,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 471
|
|
},
|
|
{
|
|
"epoch": 0.5791411042944785,
|
|
"grad_norm": 37.764625171132664,
|
|
"learning_rate": 8.152892561983471e-07,
|
|
"logits/chosen": -0.046142578125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3377,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.87109375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 472
|
|
},
|
|
{
|
|
"epoch": 0.5803680981595092,
|
|
"grad_norm": 33.16690966573419,
|
|
"learning_rate": 8.148760330578512e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.3486,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.5,
|
|
"step": 473
|
|
},
|
|
{
|
|
"epoch": 0.5815950920245399,
|
|
"grad_norm": 37.71975865074261,
|
|
"learning_rate": 8.144628099173554e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.296,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 474
|
|
},
|
|
{
|
|
"epoch": 0.5828220858895705,
|
|
"grad_norm": 31.866582554566868,
|
|
"learning_rate": 8.140495867768594e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2639,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 7.78125,
|
|
"rewards/rejected": -6.1875,
|
|
"step": 475
|
|
},
|
|
{
|
|
"epoch": 0.5840490797546012,
|
|
"grad_norm": 32.60067460423666,
|
|
"learning_rate": 8.136363636363636e-07,
|
|
"logits/chosen": -0.09619140625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2944,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": -0.1923828125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 476
|
|
},
|
|
{
|
|
"epoch": 0.5852760736196319,
|
|
"grad_norm": 34.300412867411,
|
|
"learning_rate": 8.132231404958677e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2808,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 477
|
|
},
|
|
{
|
|
"epoch": 0.5865030674846625,
|
|
"grad_norm": 33.50508466908412,
|
|
"learning_rate": 8.128099173553719e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3184,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.87890625,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 478
|
|
},
|
|
{
|
|
"epoch": 0.5877300613496933,
|
|
"grad_norm": 32.74590499905971,
|
|
"learning_rate": 8.12396694214876e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.04638671875,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 479
|
|
},
|
|
{
|
|
"epoch": 0.588957055214724,
|
|
"grad_norm": 31.25419724588422,
|
|
"learning_rate": 8.119834710743802e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2338,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.296875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 480
|
|
},
|
|
{
|
|
"epoch": 0.5901840490797546,
|
|
"grad_norm": 40.0079772160971,
|
|
"learning_rate": 8.115702479338843e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.4119,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.7578125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 481
|
|
},
|
|
{
|
|
"epoch": 0.5914110429447853,
|
|
"grad_norm": 35.87779767449304,
|
|
"learning_rate": 8.111570247933885e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2926,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 482
|
|
},
|
|
{
|
|
"epoch": 0.592638036809816,
|
|
"grad_norm": 34.477936142548806,
|
|
"learning_rate": 8.107438016528926e-07,
|
|
"logits/chosen": -0.0306396484375,
|
|
"logits/rejected": -0.1708984375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3076,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.7578125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.75,
|
|
"step": 483
|
|
},
|
|
{
|
|
"epoch": 0.5938650306748466,
|
|
"grad_norm": 35.06510512407324,
|
|
"learning_rate": 8.103305785123967e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3428,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 484
|
|
},
|
|
{
|
|
"epoch": 0.5950920245398773,
|
|
"grad_norm": 29.58257600279775,
|
|
"learning_rate": 8.099173553719008e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.276,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 485
|
|
},
|
|
{
|
|
"epoch": 0.596319018404908,
|
|
"grad_norm": 35.63248619960468,
|
|
"learning_rate": 8.09504132231405e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3049,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 486
|
|
},
|
|
{
|
|
"epoch": 0.5975460122699386,
|
|
"grad_norm": 27.478788009307355,
|
|
"learning_rate": 8.09090909090909e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2195,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 487
|
|
},
|
|
{
|
|
"epoch": 0.5987730061349693,
|
|
"grad_norm": 35.65492860883581,
|
|
"learning_rate": 8.086776859504131e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.25,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 488
|
|
},
|
|
{
|
|
"epoch": 0.6,
|
|
"grad_norm": 29.439269568854748,
|
|
"learning_rate": 8.082644628099173e-07,
|
|
"logits/chosen": -0.09619140625,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2831,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 489
|
|
},
|
|
{
|
|
"epoch": 0.6012269938650306,
|
|
"grad_norm": 40.34135361403921,
|
|
"learning_rate": 8.078512396694214e-07,
|
|
"logits/chosen": -0.06103515625,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3637,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 490
|
|
},
|
|
{
|
|
"epoch": 0.6024539877300613,
|
|
"grad_norm": 33.6425847997183,
|
|
"learning_rate": 8.074380165289256e-07,
|
|
"logits/chosen": -0.04443359375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2811,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 491
|
|
},
|
|
{
|
|
"epoch": 0.603680981595092,
|
|
"grad_norm": 37.77385653884553,
|
|
"learning_rate": 8.070247933884297e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.306,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 492
|
|
},
|
|
{
|
|
"epoch": 0.6049079754601226,
|
|
"grad_norm": 43.24053528343554,
|
|
"learning_rate": 8.066115702479338e-07,
|
|
"logits/chosen": 0.0159912109375,
|
|
"logits/rejected": -0.09228515625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.4301,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 493
|
|
},
|
|
{
|
|
"epoch": 0.6061349693251534,
|
|
"grad_norm": 32.3711711931091,
|
|
"learning_rate": 8.061983471074379e-07,
|
|
"logits/chosen": -0.076171875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2809,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 494
|
|
},
|
|
{
|
|
"epoch": 0.6073619631901841,
|
|
"grad_norm": 35.66825520473123,
|
|
"learning_rate": 8.057851239669421e-07,
|
|
"logits/chosen": -0.0859375,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3078,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 495
|
|
},
|
|
{
|
|
"epoch": 0.6085889570552148,
|
|
"grad_norm": 38.66039815304769,
|
|
"learning_rate": 8.053719008264462e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3333,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 496
|
|
},
|
|
{
|
|
"epoch": 0.6098159509202454,
|
|
"grad_norm": 39.75597363433562,
|
|
"learning_rate": 8.049586776859504e-07,
|
|
"logits/chosen": -0.08251953125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.4202,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 497
|
|
},
|
|
{
|
|
"epoch": 0.6110429447852761,
|
|
"grad_norm": 33.630068628465295,
|
|
"learning_rate": 8.045454545454545e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3002,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 498
|
|
},
|
|
{
|
|
"epoch": 0.6122699386503068,
|
|
"grad_norm": 35.64431198341231,
|
|
"learning_rate": 8.041322314049587e-07,
|
|
"logits/chosen": -0.07666015625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.291,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 499
|
|
},
|
|
{
|
|
"epoch": 0.6134969325153374,
|
|
"grad_norm": 33.74414745074971,
|
|
"learning_rate": 8.037190082644628e-07,
|
|
"logits/chosen": -0.09716796875,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2982,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.375,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.6147239263803681,
|
|
"grad_norm": 33.58342944370235,
|
|
"learning_rate": 8.033057851239669e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2988,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 501
|
|
},
|
|
{
|
|
"epoch": 0.6159509202453988,
|
|
"grad_norm": 32.39121968024134,
|
|
"learning_rate": 8.02892561983471e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2782,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.125,
|
|
"step": 502
|
|
},
|
|
{
|
|
"epoch": 0.6171779141104294,
|
|
"grad_norm": 33.84549068331807,
|
|
"learning_rate": 8.024793388429752e-07,
|
|
"logits/chosen": 0.055908203125,
|
|
"logits/rejected": -0.1455078125,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2825,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 503
|
|
},
|
|
{
|
|
"epoch": 0.6184049079754601,
|
|
"grad_norm": 27.907537064187128,
|
|
"learning_rate": 8.020661157024793e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2727,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.5,
|
|
"step": 504
|
|
},
|
|
{
|
|
"epoch": 0.6196319018404908,
|
|
"grad_norm": 42.97391892879846,
|
|
"learning_rate": 8.016528925619835e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.3986,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.30078125,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -6.375,
|
|
"step": 505
|
|
},
|
|
{
|
|
"epoch": 0.6208588957055214,
|
|
"grad_norm": 38.43298718810276,
|
|
"learning_rate": 8.012396694214876e-07,
|
|
"logits/chosen": -0.045654296875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.268,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 506
|
|
},
|
|
{
|
|
"epoch": 0.6220858895705521,
|
|
"grad_norm": 27.936273038832482,
|
|
"learning_rate": 8.008264462809917e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.1868,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 507
|
|
},
|
|
{
|
|
"epoch": 0.6233128834355828,
|
|
"grad_norm": 32.8519316556872,
|
|
"learning_rate": 8.004132231404959e-07,
|
|
"logits/chosen": 0.0201416015625,
|
|
"logits/rejected": -0.16015625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2689,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.6171875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 508
|
|
},
|
|
{
|
|
"epoch": 0.6245398773006134,
|
|
"grad_norm": 41.63795799495562,
|
|
"learning_rate": 8e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -466.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.4381,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 8.25,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 509
|
|
},
|
|
{
|
|
"epoch": 0.6257668711656442,
|
|
"grad_norm": 28.89555459383984,
|
|
"learning_rate": 7.995867768595041e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2237,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.341796875,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 510
|
|
},
|
|
{
|
|
"epoch": 0.6269938650306749,
|
|
"grad_norm": 38.19688149477771,
|
|
"learning_rate": 7.991735537190082e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3636,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 511
|
|
},
|
|
{
|
|
"epoch": 0.6282208588957056,
|
|
"grad_norm": 41.54065794607386,
|
|
"learning_rate": 7.987603305785124e-07,
|
|
"logits/chosen": -0.051025390625,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.3803,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.265625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 512
|
|
},
|
|
{
|
|
"epoch": 0.6294478527607362,
|
|
"grad_norm": 35.913329772336745,
|
|
"learning_rate": 7.983471074380164e-07,
|
|
"logits/chosen": -0.052490234375,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3437,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 513
|
|
},
|
|
{
|
|
"epoch": 0.6306748466257669,
|
|
"grad_norm": 29.78232755098569,
|
|
"learning_rate": 7.979338842975206e-07,
|
|
"logits/chosen": -0.038818359375,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2618,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": -0.2451171875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 514
|
|
},
|
|
{
|
|
"epoch": 0.6319018404907976,
|
|
"grad_norm": 36.15654336969748,
|
|
"learning_rate": 7.975206611570247e-07,
|
|
"logits/chosen": -0.056396484375,
|
|
"logits/rejected": -0.2109375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3156,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": -0.28125,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 515
|
|
},
|
|
{
|
|
"epoch": 0.6331288343558282,
|
|
"grad_norm": 31.44788161564659,
|
|
"learning_rate": 7.971074380165289e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2639,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": -0.064453125,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 516
|
|
},
|
|
{
|
|
"epoch": 0.6343558282208589,
|
|
"grad_norm": 42.37255733997886,
|
|
"learning_rate": 7.96694214876033e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.4386,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.3359375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 517
|
|
},
|
|
{
|
|
"epoch": 0.6355828220858896,
|
|
"grad_norm": 33.14166173125557,
|
|
"learning_rate": 7.962809917355372e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2235,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 518
|
|
},
|
|
{
|
|
"epoch": 0.6368098159509202,
|
|
"grad_norm": 40.145641610013044,
|
|
"learning_rate": 7.958677685950412e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3229,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.4765625,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 519
|
|
},
|
|
{
|
|
"epoch": 0.6380368098159509,
|
|
"grad_norm": 38.367748711689615,
|
|
"learning_rate": 7.954545454545454e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3632,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 520
|
|
},
|
|
{
|
|
"epoch": 0.6392638036809816,
|
|
"grad_norm": 31.706634028120178,
|
|
"learning_rate": 7.950413223140495e-07,
|
|
"logits/chosen": 0.01434326171875,
|
|
"logits/rejected": -0.125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2611,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 521
|
|
},
|
|
{
|
|
"epoch": 0.6404907975460122,
|
|
"grad_norm": 36.111031138767444,
|
|
"learning_rate": 7.946280991735537e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3182,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 7.5,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 522
|
|
},
|
|
{
|
|
"epoch": 0.6417177914110429,
|
|
"grad_norm": 43.661807402482516,
|
|
"learning_rate": 7.942148760330578e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.4124,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.2392578125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 523
|
|
},
|
|
{
|
|
"epoch": 0.6429447852760736,
|
|
"grad_norm": 44.04386819066262,
|
|
"learning_rate": 7.938016528925619e-07,
|
|
"logits/chosen": -0.07958984375,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3794,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.251953125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.625,
|
|
"step": 524
|
|
},
|
|
{
|
|
"epoch": 0.6441717791411042,
|
|
"grad_norm": 48.408238800052914,
|
|
"learning_rate": 7.933884297520661e-07,
|
|
"logits/chosen": -0.0751953125,
|
|
"logits/rejected": -0.1337890625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.4399,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.0625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 525
|
|
},
|
|
{
|
|
"epoch": 0.645398773006135,
|
|
"grad_norm": 33.71138762956367,
|
|
"learning_rate": 7.929752066115702e-07,
|
|
"logits/chosen": -0.09814453125,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2684,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.734375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 526
|
|
},
|
|
{
|
|
"epoch": 0.6466257668711657,
|
|
"grad_norm": 35.72880037638548,
|
|
"learning_rate": 7.925619834710744e-07,
|
|
"logits/chosen": -0.031005859375,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 527
|
|
},
|
|
{
|
|
"epoch": 0.6478527607361964,
|
|
"grad_norm": 30.922550042147858,
|
|
"learning_rate": 7.921487603305784e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3056,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.404296875,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 528
|
|
},
|
|
{
|
|
"epoch": 0.649079754601227,
|
|
"grad_norm": 33.98515479129212,
|
|
"learning_rate": 7.917355371900826e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3684,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 529
|
|
},
|
|
{
|
|
"epoch": 0.6503067484662577,
|
|
"grad_norm": 37.186485518201295,
|
|
"learning_rate": 7.913223140495867e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.201171875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3376,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 530
|
|
},
|
|
{
|
|
"epoch": 0.6515337423312884,
|
|
"grad_norm": 27.176074854888466,
|
|
"learning_rate": 7.909090909090909e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2583,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.75,
|
|
"step": 531
|
|
},
|
|
{
|
|
"epoch": 0.652760736196319,
|
|
"grad_norm": 31.28326806837223,
|
|
"learning_rate": 7.90495867768595e-07,
|
|
"logits/chosen": -0.0263671875,
|
|
"logits/rejected": -0.16796875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2852,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 532
|
|
},
|
|
{
|
|
"epoch": 0.6539877300613497,
|
|
"grad_norm": 36.08815904584555,
|
|
"learning_rate": 7.900826446280992e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2781,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 533
|
|
},
|
|
{
|
|
"epoch": 0.6552147239263804,
|
|
"grad_norm": 31.873251516918746,
|
|
"learning_rate": 7.896694214876033e-07,
|
|
"logits/chosen": -0.051513671875,
|
|
"logits/rejected": -0.173828125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2876,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 534
|
|
},
|
|
{
|
|
"epoch": 0.656441717791411,
|
|
"grad_norm": 36.530095888975836,
|
|
"learning_rate": 7.892561983471075e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.312,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.06103515625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -6.0,
|
|
"step": 535
|
|
},
|
|
{
|
|
"epoch": 0.6576687116564417,
|
|
"grad_norm": 32.74817222483817,
|
|
"learning_rate": 7.888429752066116e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2461,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 536
|
|
},
|
|
{
|
|
"epoch": 0.6588957055214724,
|
|
"grad_norm": 44.707496162945326,
|
|
"learning_rate": 7.884297520661157e-07,
|
|
"logits/chosen": -0.087890625,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.4648,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": -0.0277099609375,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 537
|
|
},
|
|
{
|
|
"epoch": 0.660122699386503,
|
|
"grad_norm": 39.672945580644054,
|
|
"learning_rate": 7.880165289256198e-07,
|
|
"logits/chosen": -0.06689453125,
|
|
"logits/rejected": -0.2421875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.387,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 538
|
|
},
|
|
{
|
|
"epoch": 0.6613496932515337,
|
|
"grad_norm": 35.46415326892615,
|
|
"learning_rate": 7.87603305785124e-07,
|
|
"logits/chosen": -0.06494140625,
|
|
"logits/rejected": -0.201171875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3394,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 539
|
|
},
|
|
{
|
|
"epoch": 0.6625766871165644,
|
|
"grad_norm": 37.23193708193184,
|
|
"learning_rate": 7.87190082644628e-07,
|
|
"logits/chosen": -0.04248046875,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3496,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.5,
|
|
"step": 540
|
|
},
|
|
{
|
|
"epoch": 0.6638036809815951,
|
|
"grad_norm": 35.550266538453776,
|
|
"learning_rate": 7.867768595041322e-07,
|
|
"logits/chosen": -0.059814453125,
|
|
"logits/rejected": -0.1552734375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.313,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 541
|
|
},
|
|
{
|
|
"epoch": 0.6650306748466258,
|
|
"grad_norm": 32.113839687383255,
|
|
"learning_rate": 7.863636363636363e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2916,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 542
|
|
},
|
|
{
|
|
"epoch": 0.6662576687116565,
|
|
"grad_norm": 35.51517611628603,
|
|
"learning_rate": 7.859504132231404e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.293,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 543
|
|
},
|
|
{
|
|
"epoch": 0.6674846625766871,
|
|
"grad_norm": 44.19375136908175,
|
|
"learning_rate": 7.855371900826446e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.4179,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.5,
|
|
"step": 544
|
|
},
|
|
{
|
|
"epoch": 0.6687116564417178,
|
|
"grad_norm": 40.698950248901056,
|
|
"learning_rate": 7.851239669421487e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3983,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.75,
|
|
"step": 545
|
|
},
|
|
{
|
|
"epoch": 0.6699386503067485,
|
|
"grad_norm": 37.546979812882604,
|
|
"learning_rate": 7.847107438016528e-07,
|
|
"logits/chosen": -0.0537109375,
|
|
"logits/rejected": -0.1572265625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3606,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 546
|
|
},
|
|
{
|
|
"epoch": 0.6711656441717792,
|
|
"grad_norm": 31.44946831622725,
|
|
"learning_rate": 7.842975206611569e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2976,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 547
|
|
},
|
|
{
|
|
"epoch": 0.6723926380368098,
|
|
"grad_norm": 40.34732101502891,
|
|
"learning_rate": 7.838842975206611e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2922,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 548
|
|
},
|
|
{
|
|
"epoch": 0.6736196319018405,
|
|
"grad_norm": 36.61410881664837,
|
|
"learning_rate": 7.834710743801652e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3224,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 549
|
|
},
|
|
{
|
|
"epoch": 0.6748466257668712,
|
|
"grad_norm": 36.09415599886978,
|
|
"learning_rate": 7.830578512396694e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3663,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 0.6760736196319018,
|
|
"grad_norm": 32.33592929698809,
|
|
"learning_rate": 7.826446280991735e-07,
|
|
"logits/chosen": -0.050048828125,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.307,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 551
|
|
},
|
|
{
|
|
"epoch": 0.6773006134969325,
|
|
"grad_norm": 34.998070605055936,
|
|
"learning_rate": 7.822314049586777e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 552
|
|
},
|
|
{
|
|
"epoch": 0.6785276073619632,
|
|
"grad_norm": 36.85987881722463,
|
|
"learning_rate": 7.818181818181818e-07,
|
|
"logits/chosen": -0.08935546875,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3871,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.267578125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 553
|
|
},
|
|
{
|
|
"epoch": 0.6797546012269938,
|
|
"grad_norm": 33.055846271688495,
|
|
"learning_rate": 7.81404958677686e-07,
|
|
"logits/chosen": -0.03515625,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2646,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.125,
|
|
"step": 554
|
|
},
|
|
{
|
|
"epoch": 0.6809815950920245,
|
|
"grad_norm": 35.60901600819588,
|
|
"learning_rate": 7.8099173553719e-07,
|
|
"logits/chosen": 0.07568359375,
|
|
"logits/rejected": -0.047119140625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3521,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 555
|
|
},
|
|
{
|
|
"epoch": 0.6822085889570552,
|
|
"grad_norm": 25.690739155417347,
|
|
"learning_rate": 7.805785123966942e-07,
|
|
"logits/chosen": -0.0576171875,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2129,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 556
|
|
},
|
|
{
|
|
"epoch": 0.6834355828220859,
|
|
"grad_norm": 30.514742939928254,
|
|
"learning_rate": 7.801652892561983e-07,
|
|
"logits/chosen": -0.0654296875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2947,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 557
|
|
},
|
|
{
|
|
"epoch": 0.6846625766871166,
|
|
"grad_norm": 29.65623720712839,
|
|
"learning_rate": 7.797520661157025e-07,
|
|
"logits/chosen": -0.0272216796875,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2704,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.443359375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -6.25,
|
|
"step": 558
|
|
},
|
|
{
|
|
"epoch": 0.6858895705521473,
|
|
"grad_norm": 37.74863410035385,
|
|
"learning_rate": 7.793388429752066e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3346,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 559
|
|
},
|
|
{
|
|
"epoch": 0.6871165644171779,
|
|
"grad_norm": 22.447578429969834,
|
|
"learning_rate": 7.789256198347107e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2058,
|
|
"rewards/accuracies": 0.9375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 560
|
|
},
|
|
{
|
|
"epoch": 0.6883435582822086,
|
|
"grad_norm": 28.553930750757928,
|
|
"learning_rate": 7.785123966942149e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2121,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 561
|
|
},
|
|
{
|
|
"epoch": 0.6895705521472393,
|
|
"grad_norm": 28.88112185443711,
|
|
"learning_rate": 7.78099173553719e-07,
|
|
"logits/chosen": 0.0203857421875,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2434,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.458984375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 562
|
|
},
|
|
{
|
|
"epoch": 0.69079754601227,
|
|
"grad_norm": 30.44618082475287,
|
|
"learning_rate": 7.776859504132232e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3087,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 563
|
|
},
|
|
{
|
|
"epoch": 0.6920245398773006,
|
|
"grad_norm": 23.53292610789856,
|
|
"learning_rate": 7.772727272727272e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2313,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 564
|
|
},
|
|
{
|
|
"epoch": 0.6932515337423313,
|
|
"grad_norm": 32.54292501432025,
|
|
"learning_rate": 7.768595041322314e-07,
|
|
"logits/chosen": -0.099609375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3924,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 565
|
|
},
|
|
{
|
|
"epoch": 0.694478527607362,
|
|
"grad_norm": 32.727602830013666,
|
|
"learning_rate": 7.764462809917354e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2518,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 566
|
|
},
|
|
{
|
|
"epoch": 0.6957055214723926,
|
|
"grad_norm": 23.022890741215004,
|
|
"learning_rate": 7.760330578512397e-07,
|
|
"logits/chosen": -0.004852294921875,
|
|
"logits/rejected": -0.1357421875,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.193,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 567
|
|
},
|
|
{
|
|
"epoch": 0.6969325153374233,
|
|
"grad_norm": 36.23377796534644,
|
|
"learning_rate": 7.756198347107437e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.2041015625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.326,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 568
|
|
},
|
|
{
|
|
"epoch": 0.698159509202454,
|
|
"grad_norm": 29.465085248933423,
|
|
"learning_rate": 7.752066115702479e-07,
|
|
"logits/chosen": -0.060791015625,
|
|
"logits/rejected": -0.2060546875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2385,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 569
|
|
},
|
|
{
|
|
"epoch": 0.6993865030674846,
|
|
"grad_norm": 24.94400322793083,
|
|
"learning_rate": 7.74793388429752e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.1676,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.25,
|
|
"step": 570
|
|
},
|
|
{
|
|
"epoch": 0.7006134969325153,
|
|
"grad_norm": 31.9046154994296,
|
|
"learning_rate": 7.743801652892562e-07,
|
|
"logits/chosen": -0.0927734375,
|
|
"logits/rejected": -0.21875,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2539,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.7578125,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 571
|
|
},
|
|
{
|
|
"epoch": 0.701840490797546,
|
|
"grad_norm": 31.18991534975819,
|
|
"learning_rate": 7.739669421487603e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2406,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.96875,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 572
|
|
},
|
|
{
|
|
"epoch": 0.7030674846625767,
|
|
"grad_norm": 40.389538917168906,
|
|
"learning_rate": 7.735537190082644e-07,
|
|
"logits/chosen": -0.0277099609375,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.4632,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 573
|
|
},
|
|
{
|
|
"epoch": 0.7042944785276074,
|
|
"grad_norm": 39.70707746238728,
|
|
"learning_rate": 7.731404958677685e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3206,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 574
|
|
},
|
|
{
|
|
"epoch": 0.7055214723926381,
|
|
"grad_norm": 31.61799879254746,
|
|
"learning_rate": 7.727272727272727e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2326,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.8515625,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 575
|
|
},
|
|
{
|
|
"epoch": 0.7067484662576687,
|
|
"grad_norm": 24.65806387193105,
|
|
"learning_rate": 7.723140495867768e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.1948,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 7.65625,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 576
|
|
},
|
|
{
|
|
"epoch": 0.7079754601226994,
|
|
"grad_norm": 36.42980548819235,
|
|
"learning_rate": 7.719008264462809e-07,
|
|
"logits/chosen": -0.00396728515625,
|
|
"logits/rejected": -0.1328125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3269,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 577
|
|
},
|
|
{
|
|
"epoch": 0.7092024539877301,
|
|
"grad_norm": 33.09293756143091,
|
|
"learning_rate": 7.714876033057851e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3248,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 7.65625,
|
|
"rewards/rejected": -6.0,
|
|
"step": 578
|
|
},
|
|
{
|
|
"epoch": 0.7104294478527607,
|
|
"grad_norm": 41.891361592921854,
|
|
"learning_rate": 7.710743801652892e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.4104,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.2392578125,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 579
|
|
},
|
|
{
|
|
"epoch": 0.7116564417177914,
|
|
"grad_norm": 34.636238626793094,
|
|
"learning_rate": 7.706611570247934e-07,
|
|
"logits/chosen": -0.033447265625,
|
|
"logits/rejected": -0.2041015625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2367,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 580
|
|
},
|
|
{
|
|
"epoch": 0.7128834355828221,
|
|
"grad_norm": 28.382108013456723,
|
|
"learning_rate": 7.702479338842975e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2203,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 581
|
|
},
|
|
{
|
|
"epoch": 0.7141104294478527,
|
|
"grad_norm": 37.59655611342421,
|
|
"learning_rate": 7.698347107438016e-07,
|
|
"logits/chosen": 0.06982421875,
|
|
"logits/rejected": -0.1484375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.309,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.6796875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 582
|
|
},
|
|
{
|
|
"epoch": 0.7153374233128834,
|
|
"grad_norm": 36.220508133309536,
|
|
"learning_rate": 7.694214876033057e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3079,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 583
|
|
},
|
|
{
|
|
"epoch": 0.7165644171779141,
|
|
"grad_norm": 40.69950074705535,
|
|
"learning_rate": 7.690082644628099e-07,
|
|
"logits/chosen": -0.027099609375,
|
|
"logits/rejected": -0.09521484375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.4144,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.09521484375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -5.125,
|
|
"step": 584
|
|
},
|
|
{
|
|
"epoch": 0.7177914110429447,
|
|
"grad_norm": 43.10567674720949,
|
|
"learning_rate": 7.68595041322314e-07,
|
|
"logits/chosen": 0.06591796875,
|
|
"logits/rejected": -0.0537109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.4515,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 585
|
|
},
|
|
{
|
|
"epoch": 0.7190184049079754,
|
|
"grad_norm": 36.145469481313405,
|
|
"learning_rate": 7.681818181818182e-07,
|
|
"logits/chosen": -0.0673828125,
|
|
"logits/rejected": -0.17578125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3494,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.470703125,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 586
|
|
},
|
|
{
|
|
"epoch": 0.7202453987730061,
|
|
"grad_norm": 32.7614093132039,
|
|
"learning_rate": 7.677685950413223e-07,
|
|
"logits/chosen": -0.0966796875,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.333,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 587
|
|
},
|
|
{
|
|
"epoch": 0.7214723926380369,
|
|
"grad_norm": 25.043925429714594,
|
|
"learning_rate": 7.673553719008265e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2467,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 588
|
|
},
|
|
{
|
|
"epoch": 0.7226993865030675,
|
|
"grad_norm": 34.35378043572271,
|
|
"learning_rate": 7.669421487603306e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3603,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.05224609375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 589
|
|
},
|
|
{
|
|
"epoch": 0.7239263803680982,
|
|
"grad_norm": 28.467941642279985,
|
|
"learning_rate": 7.665289256198347e-07,
|
|
"logits/chosen": -0.08154296875,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2867,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.59375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 590
|
|
},
|
|
{
|
|
"epoch": 0.7251533742331289,
|
|
"grad_norm": 37.68214390316246,
|
|
"learning_rate": 7.661157024793388e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2616,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 7.5625,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 591
|
|
},
|
|
{
|
|
"epoch": 0.7263803680981595,
|
|
"grad_norm": 29.27377189590371,
|
|
"learning_rate": 7.65702479338843e-07,
|
|
"logits/chosen": -0.10693359375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2369,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 7.78125,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 592
|
|
},
|
|
{
|
|
"epoch": 0.7276073619631902,
|
|
"grad_norm": 40.332707371918964,
|
|
"learning_rate": 7.65289256198347e-07,
|
|
"logits/chosen": -0.0478515625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.3628,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 593
|
|
},
|
|
{
|
|
"epoch": 0.7288343558282209,
|
|
"grad_norm": 33.289607036623224,
|
|
"learning_rate": 7.648760330578513e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.03125,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 594
|
|
},
|
|
{
|
|
"epoch": 0.7300613496932515,
|
|
"grad_norm": 34.24935390172643,
|
|
"learning_rate": 7.644628099173553e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2221,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.80859375,
|
|
"rewards/margins": 7.625,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 595
|
|
},
|
|
{
|
|
"epoch": 0.7312883435582822,
|
|
"grad_norm": 34.78409025027798,
|
|
"learning_rate": 7.640495867768594e-07,
|
|
"logits/chosen": -0.09521484375,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3039,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.419921875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 596
|
|
},
|
|
{
|
|
"epoch": 0.7325153374233129,
|
|
"grad_norm": 35.11934595828285,
|
|
"learning_rate": 7.636363636363636e-07,
|
|
"logits/chosen": 0.0028533935546875,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3534,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.474609375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -6.0,
|
|
"step": 597
|
|
},
|
|
{
|
|
"epoch": 0.7337423312883435,
|
|
"grad_norm": 28.132698718957986,
|
|
"learning_rate": 7.632231404958677e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2167,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 7.6875,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 598
|
|
},
|
|
{
|
|
"epoch": 0.7349693251533742,
|
|
"grad_norm": 30.286056210482066,
|
|
"learning_rate": 7.628099173553718e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.2975,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 7.6875,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 599
|
|
},
|
|
{
|
|
"epoch": 0.7361963190184049,
|
|
"grad_norm": 36.213324901908116,
|
|
"learning_rate": 7.623966942148759e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.4124,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.2890625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.7374233128834355,
|
|
"grad_norm": 32.06779553662014,
|
|
"learning_rate": 7.619834710743801e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3017,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.79296875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -7.03125,
|
|
"step": 601
|
|
},
|
|
{
|
|
"epoch": 0.7386503067484662,
|
|
"grad_norm": 26.190653496247283,
|
|
"learning_rate": 7.615702479338842e-07,
|
|
"logits/chosen": 0.0299072265625,
|
|
"logits/rejected": -0.12890625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2026,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.333984375,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -7.125,
|
|
"step": 602
|
|
},
|
|
{
|
|
"epoch": 0.7398773006134969,
|
|
"grad_norm": 27.143434858530537,
|
|
"learning_rate": 7.611570247933884e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2978,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.578125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 603
|
|
},
|
|
{
|
|
"epoch": 0.7411042944785277,
|
|
"grad_norm": 34.631139500617245,
|
|
"learning_rate": 7.607438016528925e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2591,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.0869140625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -7.125,
|
|
"step": 604
|
|
},
|
|
{
|
|
"epoch": 0.7423312883435583,
|
|
"grad_norm": 31.74919177883954,
|
|
"learning_rate": 7.603305785123967e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.261,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": -0.0184326171875,
|
|
"rewards/margins": 7.375,
|
|
"rewards/rejected": -7.375,
|
|
"step": 605
|
|
},
|
|
{
|
|
"epoch": 0.743558282208589,
|
|
"grad_norm": 36.161235720593524,
|
|
"learning_rate": 7.599173553719008e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3577,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.109375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 606
|
|
},
|
|
{
|
|
"epoch": 0.7447852760736197,
|
|
"grad_norm": 32.105367044031574,
|
|
"learning_rate": 7.59504132231405e-07,
|
|
"logits/chosen": -0.09814453125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.2672,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": -0.6796875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -7.84375,
|
|
"step": 607
|
|
},
|
|
{
|
|
"epoch": 0.7460122699386503,
|
|
"grad_norm": 33.485252705726495,
|
|
"learning_rate": 7.59090909090909e-07,
|
|
"logits/chosen": 0.01544189453125,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3304,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.94921875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 608
|
|
},
|
|
{
|
|
"epoch": 0.747239263803681,
|
|
"grad_norm": 36.86942720098226,
|
|
"learning_rate": 7.586776859504132e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3437,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": -0.57421875,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -7.53125,
|
|
"step": 609
|
|
},
|
|
{
|
|
"epoch": 0.7484662576687117,
|
|
"grad_norm": 30.686568178530944,
|
|
"learning_rate": 7.582644628099173e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.2626,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.049072265625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -7.6875,
|
|
"step": 610
|
|
},
|
|
{
|
|
"epoch": 0.7496932515337423,
|
|
"grad_norm": 49.02806362281854,
|
|
"learning_rate": 7.578512396694215e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.4296,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": -1.2109375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -7.40625,
|
|
"step": 611
|
|
},
|
|
{
|
|
"epoch": 0.750920245398773,
|
|
"grad_norm": 37.48523436202966,
|
|
"learning_rate": 7.574380165289256e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.4003,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": -0.1787109375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 612
|
|
},
|
|
{
|
|
"epoch": 0.7521472392638037,
|
|
"grad_norm": 36.236160655246614,
|
|
"learning_rate": 7.570247933884297e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.185546875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3752,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": -0.68359375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -7.125,
|
|
"step": 613
|
|
},
|
|
{
|
|
"epoch": 0.7533742331288343,
|
|
"grad_norm": 36.75268374092951,
|
|
"learning_rate": 7.566115702479339e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.1337890625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -7.0,
|
|
"step": 614
|
|
},
|
|
{
|
|
"epoch": 0.754601226993865,
|
|
"grad_norm": 33.698627786550716,
|
|
"learning_rate": 7.56198347107438e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3104,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.37109375,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -6.9375,
|
|
"step": 615
|
|
},
|
|
{
|
|
"epoch": 0.7558282208588957,
|
|
"grad_norm": 30.559780985185174,
|
|
"learning_rate": 7.557851239669422e-07,
|
|
"logits/chosen": -0.109375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2703,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": -0.1494140625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 616
|
|
},
|
|
{
|
|
"epoch": 0.7570552147239263,
|
|
"grad_norm": 39.386718246924126,
|
|
"learning_rate": 7.553719008264462e-07,
|
|
"logits/chosen": -0.08544921875,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3872,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.1982421875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.875,
|
|
"step": 617
|
|
},
|
|
{
|
|
"epoch": 0.758282208588957,
|
|
"grad_norm": 46.324705589027026,
|
|
"learning_rate": 7.549586776859504e-07,
|
|
"logits/chosen": -0.011474609375,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.4326,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.25,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 618
|
|
},
|
|
{
|
|
"epoch": 0.7595092024539877,
|
|
"grad_norm": 30.627967348839967,
|
|
"learning_rate": 7.545454545454545e-07,
|
|
"logits/chosen": -0.037109375,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2452,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 619
|
|
},
|
|
{
|
|
"epoch": 0.7607361963190185,
|
|
"grad_norm": 44.10697642876468,
|
|
"learning_rate": 7.541322314049587e-07,
|
|
"logits/chosen": -0.041259765625,
|
|
"logits/rejected": -0.2421875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3218,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.392578125,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 620
|
|
},
|
|
{
|
|
"epoch": 0.7619631901840491,
|
|
"grad_norm": 32.581029083272384,
|
|
"learning_rate": 7.537190082644627e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -478.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2759,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 621
|
|
},
|
|
{
|
|
"epoch": 0.7631901840490798,
|
|
"grad_norm": 31.98384527719372,
|
|
"learning_rate": 7.53305785123967e-07,
|
|
"logits/chosen": -0.046875,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2609,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.66796875,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 622
|
|
},
|
|
{
|
|
"epoch": 0.7644171779141105,
|
|
"grad_norm": 36.238195387056656,
|
|
"learning_rate": 7.52892561983471e-07,
|
|
"logits/chosen": -0.076171875,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.3195,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 623
|
|
},
|
|
{
|
|
"epoch": 0.7656441717791411,
|
|
"grad_norm": 30.409808531447826,
|
|
"learning_rate": 7.524793388429752e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2453,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 7.875,
|
|
"rewards/rejected": -6.625,
|
|
"step": 624
|
|
},
|
|
{
|
|
"epoch": 0.7668711656441718,
|
|
"grad_norm": 30.338663314610898,
|
|
"learning_rate": 7.520661157024793e-07,
|
|
"logits/chosen": -0.10546875,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2278,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.5,
|
|
"step": 625
|
|
},
|
|
{
|
|
"epoch": 0.7680981595092025,
|
|
"grad_norm": 37.82309263616244,
|
|
"learning_rate": 7.516528925619834e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3596,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 626
|
|
},
|
|
{
|
|
"epoch": 0.7693251533742331,
|
|
"grad_norm": 32.50530579632077,
|
|
"learning_rate": 7.512396694214875e-07,
|
|
"logits/chosen": -0.0218505859375,
|
|
"logits/rejected": -0.10498046875,
|
|
"logps/chosen": -310.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3278,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": -0.1591796875,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 627
|
|
},
|
|
{
|
|
"epoch": 0.7705521472392638,
|
|
"grad_norm": 37.587412820262834,
|
|
"learning_rate": 7.508264462809917e-07,
|
|
"logits/chosen": -0.047607421875,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3091,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 628
|
|
},
|
|
{
|
|
"epoch": 0.7717791411042945,
|
|
"grad_norm": 21.71383824575051,
|
|
"learning_rate": 7.504132231404958e-07,
|
|
"logits/chosen": -0.07958984375,
|
|
"logits/rejected": -0.201171875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.1735,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 629
|
|
},
|
|
{
|
|
"epoch": 0.7730061349693251,
|
|
"grad_norm": 30.12378099002339,
|
|
"learning_rate": 7.5e-07,
|
|
"logits/chosen": -0.016845703125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2675,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 630
|
|
},
|
|
{
|
|
"epoch": 0.7742331288343558,
|
|
"grad_norm": 33.98419118837188,
|
|
"learning_rate": 7.495867768595041e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3587,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 631
|
|
},
|
|
{
|
|
"epoch": 0.7754601226993865,
|
|
"grad_norm": 35.183542839414216,
|
|
"learning_rate": 7.491735537190082e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3281,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 632
|
|
},
|
|
{
|
|
"epoch": 0.7766871165644171,
|
|
"grad_norm": 40.51431962905515,
|
|
"learning_rate": 7.487603305785124e-07,
|
|
"logits/chosen": -0.04248046875,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3706,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 633
|
|
},
|
|
{
|
|
"epoch": 0.7779141104294478,
|
|
"grad_norm": 30.98589205588177,
|
|
"learning_rate": 7.483471074380165e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2213,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 2.875,
|
|
"rewards/margins": 7.90625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 634
|
|
},
|
|
{
|
|
"epoch": 0.7791411042944786,
|
|
"grad_norm": 26.611336041491974,
|
|
"learning_rate": 7.479338842975206e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.208984375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2088,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 635
|
|
},
|
|
{
|
|
"epoch": 0.7803680981595092,
|
|
"grad_norm": 38.13714899235208,
|
|
"learning_rate": 7.475206611570247e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3472,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 636
|
|
},
|
|
{
|
|
"epoch": 0.7815950920245399,
|
|
"grad_norm": 43.92799656490318,
|
|
"learning_rate": 7.471074380165289e-07,
|
|
"logits/chosen": 0.01190185546875,
|
|
"logits/rejected": -0.16796875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.4102,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 637
|
|
},
|
|
{
|
|
"epoch": 0.7828220858895706,
|
|
"grad_norm": 31.881872275499614,
|
|
"learning_rate": 7.46694214876033e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2583,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 638
|
|
},
|
|
{
|
|
"epoch": 0.7840490797546013,
|
|
"grad_norm": 38.49434546180358,
|
|
"learning_rate": 7.462809917355372e-07,
|
|
"logits/chosen": -0.01019287109375,
|
|
"logits/rejected": -0.1845703125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3851,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 639
|
|
},
|
|
{
|
|
"epoch": 0.7852760736196319,
|
|
"grad_norm": 36.173889678321956,
|
|
"learning_rate": 7.458677685950413e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.3122,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 640
|
|
},
|
|
{
|
|
"epoch": 0.7865030674846626,
|
|
"grad_norm": 42.67341870666886,
|
|
"learning_rate": 7.454545454545455e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.4884,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 641
|
|
},
|
|
{
|
|
"epoch": 0.7877300613496933,
|
|
"grad_norm": 27.896406400263544,
|
|
"learning_rate": 7.450413223140496e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.268,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 642
|
|
},
|
|
{
|
|
"epoch": 0.7889570552147239,
|
|
"grad_norm": 36.50632780938656,
|
|
"learning_rate": 7.446280991735538e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2863,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 643
|
|
},
|
|
{
|
|
"epoch": 0.7901840490797546,
|
|
"grad_norm": 36.90254530403193,
|
|
"learning_rate": 7.442148760330578e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3362,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 644
|
|
},
|
|
{
|
|
"epoch": 0.7914110429447853,
|
|
"grad_norm": 31.726545254851587,
|
|
"learning_rate": 7.43801652892562e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2737,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 645
|
|
},
|
|
{
|
|
"epoch": 0.7926380368098159,
|
|
"grad_norm": 32.26303661920976,
|
|
"learning_rate": 7.433884297520661e-07,
|
|
"logits/chosen": -0.051025390625,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2876,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 646
|
|
},
|
|
{
|
|
"epoch": 0.7938650306748466,
|
|
"grad_norm": 34.65899753363137,
|
|
"learning_rate": 7.429752066115703e-07,
|
|
"logits/chosen": -0.1201171875,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3098,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.984375,
|
|
"rewards/margins": 7.375,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 647
|
|
},
|
|
{
|
|
"epoch": 0.7950920245398773,
|
|
"grad_norm": 39.232231660651884,
|
|
"learning_rate": 7.425619834710743e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.374,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.8359375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 648
|
|
},
|
|
{
|
|
"epoch": 0.7963190184049079,
|
|
"grad_norm": 41.02647010732188,
|
|
"learning_rate": 7.421487603305784e-07,
|
|
"logits/chosen": -0.033447265625,
|
|
"logits/rejected": -0.12890625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.5179,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 649
|
|
},
|
|
{
|
|
"epoch": 0.7975460122699386,
|
|
"grad_norm": 43.926958102848964,
|
|
"learning_rate": 7.417355371900826e-07,
|
|
"logits/chosen": -0.05810546875,
|
|
"logits/rejected": -0.2197265625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.4029,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.5,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 0.7987730061349694,
|
|
"grad_norm": 37.4846322870058,
|
|
"learning_rate": 7.413223140495867e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3213,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 7.625,
|
|
"rewards/rejected": -6.0,
|
|
"step": 651
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"grad_norm": 39.82102192858506,
|
|
"learning_rate": 7.409090909090909e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3244,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 652
|
|
},
|
|
{
|
|
"epoch": 0.8012269938650307,
|
|
"grad_norm": 39.721500208654376,
|
|
"learning_rate": 7.404958677685949e-07,
|
|
"logits/chosen": -0.0712890625,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2804,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 653
|
|
},
|
|
{
|
|
"epoch": 0.8024539877300614,
|
|
"grad_norm": 29.27062388899099,
|
|
"learning_rate": 7.400826446280991e-07,
|
|
"logits/chosen": -0.05126953125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2265,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.875,
|
|
"step": 654
|
|
},
|
|
{
|
|
"epoch": 0.803680981595092,
|
|
"grad_norm": 35.588927933308156,
|
|
"learning_rate": 7.396694214876032e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2825,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 655
|
|
},
|
|
{
|
|
"epoch": 0.8049079754601227,
|
|
"grad_norm": 38.2147503401273,
|
|
"learning_rate": 7.392561983471074e-07,
|
|
"logits/chosen": -0.0458984375,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3879,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 656
|
|
},
|
|
{
|
|
"epoch": 0.8061349693251534,
|
|
"grad_norm": 31.082086756562887,
|
|
"learning_rate": 7.388429752066115e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3235,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 657
|
|
},
|
|
{
|
|
"epoch": 0.807361963190184,
|
|
"grad_norm": 34.397342278852506,
|
|
"learning_rate": 7.384297520661157e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2744,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.72265625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.75,
|
|
"step": 658
|
|
},
|
|
{
|
|
"epoch": 0.8085889570552147,
|
|
"grad_norm": 39.01877683461593,
|
|
"learning_rate": 7.380165289256198e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3594,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": -0.004241943359375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 659
|
|
},
|
|
{
|
|
"epoch": 0.8098159509202454,
|
|
"grad_norm": 29.796080293736573,
|
|
"learning_rate": 7.37603305785124e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2772,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 660
|
|
},
|
|
{
|
|
"epoch": 0.811042944785276,
|
|
"grad_norm": 43.970542180811556,
|
|
"learning_rate": 7.371900826446281e-07,
|
|
"logits/chosen": -0.06884765625,
|
|
"logits/rejected": -0.177734375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.4211,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.1943359375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 661
|
|
},
|
|
{
|
|
"epoch": 0.8122699386503067,
|
|
"grad_norm": 27.972472797325533,
|
|
"learning_rate": 7.367768595041322e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.2060546875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2248,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 662
|
|
},
|
|
{
|
|
"epoch": 0.8134969325153374,
|
|
"grad_norm": 31.48669247587522,
|
|
"learning_rate": 7.363636363636363e-07,
|
|
"logits/chosen": -0.0186767578125,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3033,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 663
|
|
},
|
|
{
|
|
"epoch": 0.8147239263803681,
|
|
"grad_norm": 26.36831097405094,
|
|
"learning_rate": 7.359504132231405e-07,
|
|
"logits/chosen": -0.31640625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.1771,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 7.84375,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 664
|
|
},
|
|
{
|
|
"epoch": 0.8159509202453987,
|
|
"grad_norm": 41.87820109300682,
|
|
"learning_rate": 7.355371900826446e-07,
|
|
"logits/chosen": 0.000423431396484375,
|
|
"logits/rejected": -0.166015625,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.342,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 665
|
|
},
|
|
{
|
|
"epoch": 0.8171779141104294,
|
|
"grad_norm": 38.91570112915555,
|
|
"learning_rate": 7.351239669421487e-07,
|
|
"logits/chosen": -0.039794921875,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3886,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 666
|
|
},
|
|
{
|
|
"epoch": 0.8184049079754602,
|
|
"grad_norm": 30.797773261094097,
|
|
"learning_rate": 7.347107438016529e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2568,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.5,
|
|
"step": 667
|
|
},
|
|
{
|
|
"epoch": 0.8196319018404908,
|
|
"grad_norm": 30.324274734242014,
|
|
"learning_rate": 7.34297520661157e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2997,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 668
|
|
},
|
|
{
|
|
"epoch": 0.8208588957055215,
|
|
"grad_norm": 39.407727222455065,
|
|
"learning_rate": 7.338842975206612e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3349,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 669
|
|
},
|
|
{
|
|
"epoch": 0.8220858895705522,
|
|
"grad_norm": 29.030268011802306,
|
|
"learning_rate": 7.334710743801653e-07,
|
|
"logits/chosen": 0.038818359375,
|
|
"logits/rejected": -0.10595703125,
|
|
"logps/chosen": -320.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2506,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 670
|
|
},
|
|
{
|
|
"epoch": 0.8233128834355828,
|
|
"grad_norm": 32.39302896464466,
|
|
"learning_rate": 7.330578512396694e-07,
|
|
"logits/chosen": -0.08056640625,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2618,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 671
|
|
},
|
|
{
|
|
"epoch": 0.8245398773006135,
|
|
"grad_norm": 32.730627166939506,
|
|
"learning_rate": 7.326446280991735e-07,
|
|
"logits/chosen": -0.06396484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2935,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 672
|
|
},
|
|
{
|
|
"epoch": 0.8257668711656442,
|
|
"grad_norm": 35.46284206569561,
|
|
"learning_rate": 7.322314049586777e-07,
|
|
"logits/chosen": 0.00579833984375,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2854,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 673
|
|
},
|
|
{
|
|
"epoch": 0.8269938650306748,
|
|
"grad_norm": 35.48501784668173,
|
|
"learning_rate": 7.318181818181818e-07,
|
|
"logits/chosen": 0.037841796875,
|
|
"logits/rejected": -0.12255859375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3332,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.5,
|
|
"step": 674
|
|
},
|
|
{
|
|
"epoch": 0.8282208588957055,
|
|
"grad_norm": 34.07602348490195,
|
|
"learning_rate": 7.31404958677686e-07,
|
|
"logits/chosen": -0.037353515625,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3553,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 675
|
|
},
|
|
{
|
|
"epoch": 0.8294478527607362,
|
|
"grad_norm": 34.87811781040948,
|
|
"learning_rate": 7.3099173553719e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2859,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.8359375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 676
|
|
},
|
|
{
|
|
"epoch": 0.8306748466257668,
|
|
"grad_norm": 32.93168351329577,
|
|
"learning_rate": 7.305785123966942e-07,
|
|
"logits/chosen": -0.0242919921875,
|
|
"logits/rejected": -0.1953125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.546875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 677
|
|
},
|
|
{
|
|
"epoch": 0.8319018404907975,
|
|
"grad_norm": 34.12760581259703,
|
|
"learning_rate": 7.301652892561983e-07,
|
|
"logits/chosen": -0.052001953125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3125,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.125,
|
|
"step": 678
|
|
},
|
|
{
|
|
"epoch": 0.8331288343558282,
|
|
"grad_norm": 36.89463698994786,
|
|
"learning_rate": 7.297520661157024e-07,
|
|
"logits/chosen": 0.037841796875,
|
|
"logits/rejected": -0.0712890625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3702,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 679
|
|
},
|
|
{
|
|
"epoch": 0.8343558282208589,
|
|
"grad_norm": 29.256142841767886,
|
|
"learning_rate": 7.293388429752065e-07,
|
|
"logits/chosen": -0.09814453125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2672,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 680
|
|
},
|
|
{
|
|
"epoch": 0.8355828220858895,
|
|
"grad_norm": 26.757154748897623,
|
|
"learning_rate": 7.289256198347107e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2032,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -6.25,
|
|
"step": 681
|
|
},
|
|
{
|
|
"epoch": 0.8368098159509203,
|
|
"grad_norm": 34.48662870784652,
|
|
"learning_rate": 7.285123966942148e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.36,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.75,
|
|
"step": 682
|
|
},
|
|
{
|
|
"epoch": 0.838036809815951,
|
|
"grad_norm": 31.6242482051113,
|
|
"learning_rate": 7.28099173553719e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2923,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 683
|
|
},
|
|
{
|
|
"epoch": 0.8392638036809816,
|
|
"grad_norm": 33.064102799803514,
|
|
"learning_rate": 7.276859504132231e-07,
|
|
"logits/chosen": 0.047119140625,
|
|
"logits/rejected": -0.0693359375,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3284,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": -0.388671875,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 684
|
|
},
|
|
{
|
|
"epoch": 0.8404907975460123,
|
|
"grad_norm": 32.22229237207361,
|
|
"learning_rate": 7.272727272727272e-07,
|
|
"logits/chosen": -0.08544921875,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.298,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.498046875,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -7.03125,
|
|
"step": 685
|
|
},
|
|
{
|
|
"epoch": 0.841717791411043,
|
|
"grad_norm": 27.326122854839305,
|
|
"learning_rate": 7.268595041322314e-07,
|
|
"logits/chosen": -0.09814453125,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.266,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.40625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.5,
|
|
"step": 686
|
|
},
|
|
{
|
|
"epoch": 0.8429447852760736,
|
|
"grad_norm": 34.77177398420566,
|
|
"learning_rate": 7.264462809917355e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2982,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.47265625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 687
|
|
},
|
|
{
|
|
"epoch": 0.8441717791411043,
|
|
"grad_norm": 32.90103137234285,
|
|
"learning_rate": 7.260330578512396e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.286,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": -0.53515625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 688
|
|
},
|
|
{
|
|
"epoch": 0.845398773006135,
|
|
"grad_norm": 35.4677538273116,
|
|
"learning_rate": 7.256198347107437e-07,
|
|
"logits/chosen": -0.095703125,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3169,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.408203125,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 689
|
|
},
|
|
{
|
|
"epoch": 0.8466257668711656,
|
|
"grad_norm": 47.82010746411854,
|
|
"learning_rate": 7.252066115702479e-07,
|
|
"logits/chosen": 0.03857421875,
|
|
"logits/rejected": -0.16015625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.5312,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.5390625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.375,
|
|
"step": 690
|
|
},
|
|
{
|
|
"epoch": 0.8478527607361963,
|
|
"grad_norm": 28.75628225752802,
|
|
"learning_rate": 7.24793388429752e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2626,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.40625,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.875,
|
|
"step": 691
|
|
},
|
|
{
|
|
"epoch": 0.849079754601227,
|
|
"grad_norm": 42.35779725913325,
|
|
"learning_rate": 7.243801652892562e-07,
|
|
"logits/chosen": -0.1044921875,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2985,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.87890625,
|
|
"rewards/margins": 7.375,
|
|
"rewards/rejected": -6.5,
|
|
"step": 692
|
|
},
|
|
{
|
|
"epoch": 0.8503067484662576,
|
|
"grad_norm": 33.77138589351699,
|
|
"learning_rate": 7.239669421487603e-07,
|
|
"logits/chosen": -0.12109375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.333,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 693
|
|
},
|
|
{
|
|
"epoch": 0.8515337423312883,
|
|
"grad_norm": 45.20119760328906,
|
|
"learning_rate": 7.235537190082645e-07,
|
|
"logits/chosen": -0.10498046875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.4492,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": -0.2080078125,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 694
|
|
},
|
|
{
|
|
"epoch": 0.852760736196319,
|
|
"grad_norm": 45.718560051692215,
|
|
"learning_rate": 7.231404958677686e-07,
|
|
"logits/chosen": -0.07666015625,
|
|
"logits/rejected": -0.16796875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.4966,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": -0.06884765625,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 695
|
|
},
|
|
{
|
|
"epoch": 0.8539877300613496,
|
|
"grad_norm": 35.81247616707383,
|
|
"learning_rate": 7.227272727272728e-07,
|
|
"logits/chosen": -0.039306640625,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.309,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 696
|
|
},
|
|
{
|
|
"epoch": 0.8552147239263803,
|
|
"grad_norm": 43.825897731212095,
|
|
"learning_rate": 7.223140495867768e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.3419,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 697
|
|
},
|
|
{
|
|
"epoch": 0.8564417177914111,
|
|
"grad_norm": 35.78736283192075,
|
|
"learning_rate": 7.21900826446281e-07,
|
|
"logits/chosen": -0.040771484375,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3256,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 698
|
|
},
|
|
{
|
|
"epoch": 0.8576687116564418,
|
|
"grad_norm": 36.453198015080005,
|
|
"learning_rate": 7.214876033057851e-07,
|
|
"logits/chosen": -0.10693359375,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2859,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.357421875,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 699
|
|
},
|
|
{
|
|
"epoch": 0.8588957055214724,
|
|
"grad_norm": 37.8768222815939,
|
|
"learning_rate": 7.210743801652893e-07,
|
|
"logits/chosen": -0.00787353515625,
|
|
"logits/rejected": -0.1123046875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3363,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.8601226993865031,
|
|
"grad_norm": 39.25723778290927,
|
|
"learning_rate": 7.206611570247934e-07,
|
|
"logits/chosen": 0.0167236328125,
|
|
"logits/rejected": -0.09521484375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3918,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.75,
|
|
"step": 701
|
|
},
|
|
{
|
|
"epoch": 0.8613496932515338,
|
|
"grad_norm": 29.45196545763872,
|
|
"learning_rate": 7.202479338842974e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2043,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 7.8125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 702
|
|
},
|
|
{
|
|
"epoch": 0.8625766871165644,
|
|
"grad_norm": 30.738694822628716,
|
|
"learning_rate": 7.198347107438016e-07,
|
|
"logits/chosen": -0.004669189453125,
|
|
"logits/rejected": -0.1435546875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2985,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 703
|
|
},
|
|
{
|
|
"epoch": 0.8638036809815951,
|
|
"grad_norm": 32.62903012160973,
|
|
"learning_rate": 7.194214876033057e-07,
|
|
"logits/chosen": -0.08154296875,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3484,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 704
|
|
},
|
|
{
|
|
"epoch": 0.8650306748466258,
|
|
"grad_norm": 37.268433179228865,
|
|
"learning_rate": 7.190082644628099e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2941,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 705
|
|
},
|
|
{
|
|
"epoch": 0.8662576687116564,
|
|
"grad_norm": 35.93317835979168,
|
|
"learning_rate": 7.185950413223139e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3444,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 706
|
|
},
|
|
{
|
|
"epoch": 0.8674846625766871,
|
|
"grad_norm": 32.83261430682924,
|
|
"learning_rate": 7.181818181818181e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.27,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 707
|
|
},
|
|
{
|
|
"epoch": 0.8687116564417178,
|
|
"grad_norm": 31.02174274459476,
|
|
"learning_rate": 7.177685950413222e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2662,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 708
|
|
},
|
|
{
|
|
"epoch": 0.8699386503067484,
|
|
"grad_norm": 24.67697083946156,
|
|
"learning_rate": 7.173553719008264e-07,
|
|
"logits/chosen": -0.025146484375,
|
|
"logits/rejected": -0.1865234375,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2054,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 8.25,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 709
|
|
},
|
|
{
|
|
"epoch": 0.8711656441717791,
|
|
"grad_norm": 31.37694718110565,
|
|
"learning_rate": 7.169421487603305e-07,
|
|
"logits/chosen": 0.0296630859375,
|
|
"logits/rejected": -0.1044921875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3072,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.6484375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 710
|
|
},
|
|
{
|
|
"epoch": 0.8723926380368098,
|
|
"grad_norm": 27.905237479227235,
|
|
"learning_rate": 7.165289256198347e-07,
|
|
"logits/chosen": 0.003997802734375,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2746,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 711
|
|
},
|
|
{
|
|
"epoch": 0.8736196319018404,
|
|
"grad_norm": 38.712158048146684,
|
|
"learning_rate": 7.161157024793388e-07,
|
|
"logits/chosen": 0.005615234375,
|
|
"logits/rejected": -0.14453125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3845,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 712
|
|
},
|
|
{
|
|
"epoch": 0.8748466257668711,
|
|
"grad_norm": 28.765796451166402,
|
|
"learning_rate": 7.15702479338843e-07,
|
|
"logits/chosen": -0.08203125,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2628,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 713
|
|
},
|
|
{
|
|
"epoch": 0.8760736196319019,
|
|
"grad_norm": 33.59381716607132,
|
|
"learning_rate": 7.152892561983471e-07,
|
|
"logits/chosen": 0.03515625,
|
|
"logits/rejected": -0.1826171875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3023,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.91796875,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -5.75,
|
|
"step": 714
|
|
},
|
|
{
|
|
"epoch": 0.8773006134969326,
|
|
"grad_norm": 28.144956963469994,
|
|
"learning_rate": 7.148760330578512e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2619,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 715
|
|
},
|
|
{
|
|
"epoch": 0.8785276073619632,
|
|
"grad_norm": 31.37035640952665,
|
|
"learning_rate": 7.144628099173553e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3094,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 716
|
|
},
|
|
{
|
|
"epoch": 0.8797546012269939,
|
|
"grad_norm": 31.006168306577173,
|
|
"learning_rate": 7.140495867768595e-07,
|
|
"logits/chosen": -0.0498046875,
|
|
"logits/rejected": -0.208984375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2815,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 717
|
|
},
|
|
{
|
|
"epoch": 0.8809815950920246,
|
|
"grad_norm": 30.73153346857215,
|
|
"learning_rate": 7.136363636363636e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2255,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 7.6875,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 718
|
|
},
|
|
{
|
|
"epoch": 0.8822085889570552,
|
|
"grad_norm": 38.41588372897521,
|
|
"learning_rate": 7.132231404958677e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3139,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 719
|
|
},
|
|
{
|
|
"epoch": 0.8834355828220859,
|
|
"grad_norm": 26.827242556670615,
|
|
"learning_rate": 7.128099173553719e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2223,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 7.65625,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 720
|
|
},
|
|
{
|
|
"epoch": 0.8846625766871166,
|
|
"grad_norm": 25.00213343436406,
|
|
"learning_rate": 7.12396694214876e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2234,
|
|
"rewards/accuracies": 0.9375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 721
|
|
},
|
|
{
|
|
"epoch": 0.8858895705521472,
|
|
"grad_norm": 28.769798448354592,
|
|
"learning_rate": 7.119834710743802e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2256,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 722
|
|
},
|
|
{
|
|
"epoch": 0.8871165644171779,
|
|
"grad_norm": 33.86575550751347,
|
|
"learning_rate": 7.115702479338843e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 723
|
|
},
|
|
{
|
|
"epoch": 0.8883435582822086,
|
|
"grad_norm": 31.590261922698286,
|
|
"learning_rate": 7.111570247933884e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2706,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 724
|
|
},
|
|
{
|
|
"epoch": 0.8895705521472392,
|
|
"grad_norm": 33.780056784071235,
|
|
"learning_rate": 7.107438016528925e-07,
|
|
"logits/chosen": -0.002960205078125,
|
|
"logits/rejected": -0.1767578125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3451,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 725
|
|
},
|
|
{
|
|
"epoch": 0.8907975460122699,
|
|
"grad_norm": 30.012891436757428,
|
|
"learning_rate": 7.103305785123967e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2249,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 726
|
|
},
|
|
{
|
|
"epoch": 0.8920245398773006,
|
|
"grad_norm": 35.1770485043286,
|
|
"learning_rate": 7.099173553719008e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2507,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.375,
|
|
"step": 727
|
|
},
|
|
{
|
|
"epoch": 0.8932515337423312,
|
|
"grad_norm": 36.51114365671209,
|
|
"learning_rate": 7.09504132231405e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3237,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.96875,
|
|
"rewards/margins": 7.84375,
|
|
"rewards/rejected": -5.875,
|
|
"step": 728
|
|
},
|
|
{
|
|
"epoch": 0.894478527607362,
|
|
"grad_norm": 28.430454437088613,
|
|
"learning_rate": 7.09090909090909e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.221,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 7.9375,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 729
|
|
},
|
|
{
|
|
"epoch": 0.8957055214723927,
|
|
"grad_norm": 29.81908334892623,
|
|
"learning_rate": 7.086776859504132e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2212,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 730
|
|
},
|
|
{
|
|
"epoch": 0.8969325153374234,
|
|
"grad_norm": 32.79564398674248,
|
|
"learning_rate": 7.082644628099173e-07,
|
|
"logits/chosen": -0.0034332275390625,
|
|
"logits/rejected": -0.125,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2765,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -6.125,
|
|
"step": 731
|
|
},
|
|
{
|
|
"epoch": 0.898159509202454,
|
|
"grad_norm": 38.63951488407778,
|
|
"learning_rate": 7.078512396694215e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3878,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.625,
|
|
"step": 732
|
|
},
|
|
{
|
|
"epoch": 0.8993865030674847,
|
|
"grad_norm": 29.56794348682216,
|
|
"learning_rate": 7.074380165289255e-07,
|
|
"logits/chosen": -0.05712890625,
|
|
"logits/rejected": -0.2080078125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.236,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 733
|
|
},
|
|
{
|
|
"epoch": 0.9006134969325154,
|
|
"grad_norm": 34.52211164354205,
|
|
"learning_rate": 7.070247933884297e-07,
|
|
"logits/chosen": -0.042236328125,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.295,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 734
|
|
},
|
|
{
|
|
"epoch": 0.901840490797546,
|
|
"grad_norm": 30.150478068869884,
|
|
"learning_rate": 7.066115702479338e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.1939,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 8.25,
|
|
"rewards/rejected": -7.09375,
|
|
"step": 735
|
|
},
|
|
{
|
|
"epoch": 0.9030674846625767,
|
|
"grad_norm": 36.53381753968603,
|
|
"learning_rate": 7.06198347107438e-07,
|
|
"logits/chosen": 0.026123046875,
|
|
"logits/rejected": -0.09912109375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3826,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.1220703125,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 736
|
|
},
|
|
{
|
|
"epoch": 0.9042944785276074,
|
|
"grad_norm": 45.25449310101674,
|
|
"learning_rate": 7.057851239669421e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.1708984375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.4576,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.203125,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 737
|
|
},
|
|
{
|
|
"epoch": 0.905521472392638,
|
|
"grad_norm": 29.955058399708445,
|
|
"learning_rate": 7.053719008264462e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2561,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 7.71875,
|
|
"rewards/rejected": -6.375,
|
|
"step": 738
|
|
},
|
|
{
|
|
"epoch": 0.9067484662576687,
|
|
"grad_norm": 33.57627676756485,
|
|
"learning_rate": 7.049586776859504e-07,
|
|
"logits/chosen": -0.06103515625,
|
|
"logits/rejected": -0.1943359375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3007,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.40234375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 739
|
|
},
|
|
{
|
|
"epoch": 0.9079754601226994,
|
|
"grad_norm": 33.78022485190333,
|
|
"learning_rate": 7.045454545454545e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3257,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.4609375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 740
|
|
},
|
|
{
|
|
"epoch": 0.90920245398773,
|
|
"grad_norm": 33.1650640297202,
|
|
"learning_rate": 7.041322314049587e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3045,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 8.125,
|
|
"rewards/rejected": -6.84375,
|
|
"step": 741
|
|
},
|
|
{
|
|
"epoch": 0.9104294478527607,
|
|
"grad_norm": 30.90480025389736,
|
|
"learning_rate": 7.037190082644627e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2353,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 8.0,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 742
|
|
},
|
|
{
|
|
"epoch": 0.9116564417177914,
|
|
"grad_norm": 33.61329805299957,
|
|
"learning_rate": 7.033057851239669e-07,
|
|
"logits/chosen": -0.09716796875,
|
|
"logits/rejected": -0.1572265625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.3054,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.875,
|
|
"step": 743
|
|
},
|
|
{
|
|
"epoch": 0.912883435582822,
|
|
"grad_norm": 25.590518259699632,
|
|
"learning_rate": 7.02892561983471e-07,
|
|
"logits/chosen": -0.07666015625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2006,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 744
|
|
},
|
|
{
|
|
"epoch": 0.9141104294478528,
|
|
"grad_norm": 34.43136654833863,
|
|
"learning_rate": 7.024793388429752e-07,
|
|
"logits/chosen": -0.0284423828125,
|
|
"logits/rejected": -0.125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3174,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 745
|
|
},
|
|
{
|
|
"epoch": 0.9153374233128835,
|
|
"grad_norm": 40.51516468048696,
|
|
"learning_rate": 7.020661157024793e-07,
|
|
"logits/chosen": -0.0322265625,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3759,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 746
|
|
},
|
|
{
|
|
"epoch": 0.9165644171779141,
|
|
"grad_norm": 31.597769899241445,
|
|
"learning_rate": 7.016528925619835e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2728,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.453125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 747
|
|
},
|
|
{
|
|
"epoch": 0.9177914110429448,
|
|
"grad_norm": 39.04762321513073,
|
|
"learning_rate": 7.012396694214876e-07,
|
|
"logits/chosen": 0.00433349609375,
|
|
"logits/rejected": -0.10888671875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.416,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.01544189453125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 748
|
|
},
|
|
{
|
|
"epoch": 0.9190184049079755,
|
|
"grad_norm": 38.417255638109644,
|
|
"learning_rate": 7.008264462809918e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3824,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -6.21875,
|
|
"step": 749
|
|
},
|
|
{
|
|
"epoch": 0.9202453987730062,
|
|
"grad_norm": 41.17828471732116,
|
|
"learning_rate": 7.004132231404959e-07,
|
|
"logits/chosen": -0.1044921875,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.381,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -6.1875,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 0.9214723926380368,
|
|
"grad_norm": 40.041781400868416,
|
|
"learning_rate": 7e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3298,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 7.75,
|
|
"rewards/rejected": -6.75,
|
|
"step": 751
|
|
},
|
|
{
|
|
"epoch": 0.9226993865030675,
|
|
"grad_norm": 31.287358924979277,
|
|
"learning_rate": 6.995867768595041e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2493,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 8.0,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 752
|
|
},
|
|
{
|
|
"epoch": 0.9239263803680982,
|
|
"grad_norm": 35.190849602144475,
|
|
"learning_rate": 6.991735537190083e-07,
|
|
"logits/chosen": 0.01312255859375,
|
|
"logits/rejected": -0.12451171875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3014,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 753
|
|
},
|
|
{
|
|
"epoch": 0.9251533742331288,
|
|
"grad_norm": 37.30128361063688,
|
|
"learning_rate": 6.987603305785124e-07,
|
|
"logits/chosen": -0.05908203125,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3875,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.470703125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 754
|
|
},
|
|
{
|
|
"epoch": 0.9263803680981595,
|
|
"grad_norm": 31.219996632117976,
|
|
"learning_rate": 6.983471074380165e-07,
|
|
"logits/chosen": -0.10546875,
|
|
"logits/rejected": -0.2275390625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2524,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 755
|
|
},
|
|
{
|
|
"epoch": 0.9276073619631902,
|
|
"grad_norm": 30.607420425359958,
|
|
"learning_rate": 6.979338842975207e-07,
|
|
"logits/chosen": -0.01458740234375,
|
|
"logits/rejected": -0.10546875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2667,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.08984375,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 756
|
|
},
|
|
{
|
|
"epoch": 0.9288343558282208,
|
|
"grad_norm": 25.673659823784536,
|
|
"learning_rate": 6.975206611570247e-07,
|
|
"logits/chosen": 0.0308837890625,
|
|
"logits/rejected": -0.1904296875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2237,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 7.5,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 757
|
|
},
|
|
{
|
|
"epoch": 0.9300613496932515,
|
|
"grad_norm": 38.08378193724403,
|
|
"learning_rate": 6.971074380165289e-07,
|
|
"logits/chosen": -0.034423828125,
|
|
"logits/rejected": -0.203125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.357,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.283203125,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 758
|
|
},
|
|
{
|
|
"epoch": 0.9312883435582822,
|
|
"grad_norm": 33.845509877999,
|
|
"learning_rate": 6.96694214876033e-07,
|
|
"logits/chosen": -0.00537109375,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2702,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 759
|
|
},
|
|
{
|
|
"epoch": 0.9325153374233128,
|
|
"grad_norm": 28.0088409746571,
|
|
"learning_rate": 6.962809917355371e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2775,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 8.3125,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 760
|
|
},
|
|
{
|
|
"epoch": 0.9337423312883436,
|
|
"grad_norm": 34.61271932757203,
|
|
"learning_rate": 6.958677685950412e-07,
|
|
"logits/chosen": 0.0277099609375,
|
|
"logits/rejected": -0.10546875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2985,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 761
|
|
},
|
|
{
|
|
"epoch": 0.9349693251533743,
|
|
"grad_norm": 36.436482750324764,
|
|
"learning_rate": 6.954545454545454e-07,
|
|
"logits/chosen": 0.04052734375,
|
|
"logits/rejected": -0.142578125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2972,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 762
|
|
},
|
|
{
|
|
"epoch": 0.9361963190184049,
|
|
"grad_norm": 31.885728071569915,
|
|
"learning_rate": 6.950413223140495e-07,
|
|
"logits/chosen": -0.0198974609375,
|
|
"logits/rejected": -0.11181640625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.306,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 7.375,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 763
|
|
},
|
|
{
|
|
"epoch": 0.9374233128834356,
|
|
"grad_norm": 32.14943706006715,
|
|
"learning_rate": 6.946280991735537e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.27,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.1904296875,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -6.625,
|
|
"step": 764
|
|
},
|
|
{
|
|
"epoch": 0.9386503067484663,
|
|
"grad_norm": 40.62569491353796,
|
|
"learning_rate": 6.942148760330578e-07,
|
|
"logits/chosen": 0.000492095947265625,
|
|
"logits/rejected": -0.12890625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3557,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.6484375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 765
|
|
},
|
|
{
|
|
"epoch": 0.939877300613497,
|
|
"grad_norm": 31.76476132663633,
|
|
"learning_rate": 6.93801652892562e-07,
|
|
"logits/chosen": -0.0654296875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3003,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.0,
|
|
"step": 766
|
|
},
|
|
{
|
|
"epoch": 0.9411042944785276,
|
|
"grad_norm": 35.07353126279611,
|
|
"learning_rate": 6.933884297520661e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3239,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 767
|
|
},
|
|
{
|
|
"epoch": 0.9423312883435583,
|
|
"grad_norm": 34.24541590745614,
|
|
"learning_rate": 6.929752066115702e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.3416,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 768
|
|
},
|
|
{
|
|
"epoch": 0.943558282208589,
|
|
"grad_norm": 29.496588883092915,
|
|
"learning_rate": 6.925619834710743e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.264,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.75,
|
|
"step": 769
|
|
},
|
|
{
|
|
"epoch": 0.9447852760736196,
|
|
"grad_norm": 37.888017377085696,
|
|
"learning_rate": 6.921487603305785e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3036,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 770
|
|
},
|
|
{
|
|
"epoch": 0.9460122699386503,
|
|
"grad_norm": 41.74591781088978,
|
|
"learning_rate": 6.917355371900826e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.1416015625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.4706,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 0.1103515625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 771
|
|
},
|
|
{
|
|
"epoch": 0.947239263803681,
|
|
"grad_norm": 31.969953781435493,
|
|
"learning_rate": 6.913223140495868e-07,
|
|
"logits/chosen": 0.0159912109375,
|
|
"logits/rejected": -0.2041015625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3091,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 772
|
|
},
|
|
{
|
|
"epoch": 0.9484662576687116,
|
|
"grad_norm": 33.70149897966481,
|
|
"learning_rate": 6.909090909090909e-07,
|
|
"logits/chosen": 0.021240234375,
|
|
"logits/rejected": -0.1689453125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3331,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.1796875,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 773
|
|
},
|
|
{
|
|
"epoch": 0.9496932515337423,
|
|
"grad_norm": 36.75815167674088,
|
|
"learning_rate": 6.90495867768595e-07,
|
|
"logits/chosen": 0.002044677734375,
|
|
"logits/rejected": -0.1396484375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3729,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": -0.287109375,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -5.25,
|
|
"step": 774
|
|
},
|
|
{
|
|
"epoch": 0.950920245398773,
|
|
"grad_norm": 36.27807417125134,
|
|
"learning_rate": 6.900826446280992e-07,
|
|
"logits/chosen": 0.087890625,
|
|
"logits/rejected": -0.09814453125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3326,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 775
|
|
},
|
|
{
|
|
"epoch": 0.9521472392638037,
|
|
"grad_norm": 31.589189156684085,
|
|
"learning_rate": 6.896694214876033e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3345,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.5,
|
|
"step": 776
|
|
},
|
|
{
|
|
"epoch": 0.9533742331288344,
|
|
"grad_norm": 34.426563462025776,
|
|
"learning_rate": 6.892561983471074e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.2275390625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3353,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.435546875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 777
|
|
},
|
|
{
|
|
"epoch": 0.9546012269938651,
|
|
"grad_norm": 32.93001270220387,
|
|
"learning_rate": 6.888429752066115e-07,
|
|
"logits/chosen": -0.054443359375,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2456,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.357421875,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 778
|
|
},
|
|
{
|
|
"epoch": 0.9558282208588957,
|
|
"grad_norm": 40.1711499351323,
|
|
"learning_rate": 6.884297520661157e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3138,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.93359375,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -6.25,
|
|
"step": 779
|
|
},
|
|
{
|
|
"epoch": 0.9570552147239264,
|
|
"grad_norm": 35.21476896811508,
|
|
"learning_rate": 6.880165289256198e-07,
|
|
"logits/chosen": -0.00063323974609375,
|
|
"logits/rejected": -0.173828125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3715,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.06103515625,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 780
|
|
},
|
|
{
|
|
"epoch": 0.9582822085889571,
|
|
"grad_norm": 25.471392662636358,
|
|
"learning_rate": 6.87603305785124e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2024,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 781
|
|
},
|
|
{
|
|
"epoch": 0.9595092024539877,
|
|
"grad_norm": 27.786411223699663,
|
|
"learning_rate": 6.87190082644628e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.1904296875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2199,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 782
|
|
},
|
|
{
|
|
"epoch": 0.9607361963190184,
|
|
"grad_norm": 35.92383214242273,
|
|
"learning_rate": 6.867768595041323e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2667,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 7.6875,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 783
|
|
},
|
|
{
|
|
"epoch": 0.9619631901840491,
|
|
"grad_norm": 38.77277265712556,
|
|
"learning_rate": 6.863636363636363e-07,
|
|
"logits/chosen": -0.04931640625,
|
|
"logits/rejected": -0.0966796875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3682,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": -0.0244140625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -6.1875,
|
|
"step": 784
|
|
},
|
|
{
|
|
"epoch": 0.9631901840490797,
|
|
"grad_norm": 31.90266851164007,
|
|
"learning_rate": 6.859504132231405e-07,
|
|
"logits/chosen": -0.0179443359375,
|
|
"logits/rejected": -0.185546875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2512,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 785
|
|
},
|
|
{
|
|
"epoch": 0.9644171779141104,
|
|
"grad_norm": 27.113863566340694,
|
|
"learning_rate": 6.855371900826445e-07,
|
|
"logits/chosen": -0.0164794921875,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2269,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.2314453125,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -6.96875,
|
|
"step": 786
|
|
},
|
|
{
|
|
"epoch": 0.9656441717791411,
|
|
"grad_norm": 38.48315867389771,
|
|
"learning_rate": 6.851239669421487e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3133,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.318359375,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -7.15625,
|
|
"step": 787
|
|
},
|
|
{
|
|
"epoch": 0.9668711656441717,
|
|
"grad_norm": 25.75799559419808,
|
|
"learning_rate": 6.847107438016528e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.1881,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 7.90625,
|
|
"rewards/rejected": -7.03125,
|
|
"step": 788
|
|
},
|
|
{
|
|
"epoch": 0.9680981595092024,
|
|
"grad_norm": 32.90135828255642,
|
|
"learning_rate": 6.84297520661157e-07,
|
|
"logits/chosen": -0.0203857421875,
|
|
"logits/rejected": -0.2060546875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3296,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.009033203125,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 789
|
|
},
|
|
{
|
|
"epoch": 0.9693251533742331,
|
|
"grad_norm": 29.130488464702395,
|
|
"learning_rate": 6.838842975206611e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.1817,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 0.1689453125,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -6.875,
|
|
"step": 790
|
|
},
|
|
{
|
|
"epoch": 0.9705521472392638,
|
|
"grad_norm": 36.24126812710581,
|
|
"learning_rate": 6.834710743801652e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3177,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 7.90625,
|
|
"rewards/rejected": -7.21875,
|
|
"step": 791
|
|
},
|
|
{
|
|
"epoch": 0.9717791411042945,
|
|
"grad_norm": 34.75459181078778,
|
|
"learning_rate": 6.830578512396694e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3603,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.8359375,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 792
|
|
},
|
|
{
|
|
"epoch": 0.9730061349693252,
|
|
"grad_norm": 36.36771763644911,
|
|
"learning_rate": 6.826446280991735e-07,
|
|
"logits/chosen": -0.099609375,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3444,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.3984375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 793
|
|
},
|
|
{
|
|
"epoch": 0.9742331288343559,
|
|
"grad_norm": 37.02795141772375,
|
|
"learning_rate": 6.822314049586777e-07,
|
|
"logits/chosen": -0.046875,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.4307,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": -0.1904296875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -6.375,
|
|
"step": 794
|
|
},
|
|
{
|
|
"epoch": 0.9754601226993865,
|
|
"grad_norm": 26.039807081549554,
|
|
"learning_rate": 6.818181818181817e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2507,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -7.59375,
|
|
"step": 795
|
|
},
|
|
{
|
|
"epoch": 0.9766871165644172,
|
|
"grad_norm": 40.332433215846514,
|
|
"learning_rate": 6.814049586776859e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3141,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": -0.421875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -6.875,
|
|
"step": 796
|
|
},
|
|
{
|
|
"epoch": 0.9779141104294479,
|
|
"grad_norm": 26.54735552093693,
|
|
"learning_rate": 6.8099173553719e-07,
|
|
"logits/chosen": -0.06298828125,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2381,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 797
|
|
},
|
|
{
|
|
"epoch": 0.9791411042944785,
|
|
"grad_norm": 35.069027268070755,
|
|
"learning_rate": 6.805785123966942e-07,
|
|
"logits/chosen": -0.10498046875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3053,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 798
|
|
},
|
|
{
|
|
"epoch": 0.9803680981595092,
|
|
"grad_norm": 37.27286246589222,
|
|
"learning_rate": 6.801652892561983e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3271,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 799
|
|
},
|
|
{
|
|
"epoch": 0.9815950920245399,
|
|
"grad_norm": 36.01628957518016,
|
|
"learning_rate": 6.797520661157025e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2853,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.48828125,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.9828220858895705,
|
|
"grad_norm": 39.65535558250296,
|
|
"learning_rate": 6.793388429752066e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.4193,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 801
|
|
},
|
|
{
|
|
"epoch": 0.9840490797546012,
|
|
"grad_norm": 41.69927659280073,
|
|
"learning_rate": 6.789256198347108e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.426,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.400390625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 802
|
|
},
|
|
{
|
|
"epoch": 0.9852760736196319,
|
|
"grad_norm": 31.096547435657403,
|
|
"learning_rate": 6.785123966942149e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3024,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 803
|
|
},
|
|
{
|
|
"epoch": 0.9865030674846625,
|
|
"grad_norm": 32.05019392351643,
|
|
"learning_rate": 6.78099173553719e-07,
|
|
"logits/chosen": -0.06884765625,
|
|
"logits/rejected": -0.1826171875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3229,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.26953125,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 804
|
|
},
|
|
{
|
|
"epoch": 0.9877300613496932,
|
|
"grad_norm": 38.743266159886026,
|
|
"learning_rate": 6.776859504132231e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.439,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 805
|
|
},
|
|
{
|
|
"epoch": 0.9889570552147239,
|
|
"grad_norm": 39.2535896844492,
|
|
"learning_rate": 6.772727272727273e-07,
|
|
"logits/chosen": 0.04345703125,
|
|
"logits/rejected": -0.08740234375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3836,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 806
|
|
},
|
|
{
|
|
"epoch": 0.9901840490797545,
|
|
"grad_norm": 34.46845540454637,
|
|
"learning_rate": 6.768595041322314e-07,
|
|
"logits/chosen": -0.048583984375,
|
|
"logits/rejected": -0.2080078125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2423,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.125,
|
|
"step": 807
|
|
},
|
|
{
|
|
"epoch": 0.9914110429447853,
|
|
"grad_norm": 30.33905056557921,
|
|
"learning_rate": 6.764462809917355e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3216,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.255859375,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 808
|
|
},
|
|
{
|
|
"epoch": 0.992638036809816,
|
|
"grad_norm": 33.590212816681365,
|
|
"learning_rate": 6.760330578512397e-07,
|
|
"logits/chosen": -0.109375,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2833,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 7.78125,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 809
|
|
},
|
|
{
|
|
"epoch": 0.9938650306748467,
|
|
"grad_norm": 33.059294007862654,
|
|
"learning_rate": 6.756198347107437e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.239,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 7.25,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 810
|
|
},
|
|
{
|
|
"epoch": 0.9950920245398773,
|
|
"grad_norm": 30.59632262177732,
|
|
"learning_rate": 6.75206611570248e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2783,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 811
|
|
},
|
|
{
|
|
"epoch": 0.996319018404908,
|
|
"grad_norm": 38.69919040079445,
|
|
"learning_rate": 6.74793388429752e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2784,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.4609375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.5625,
|
|
"step": 812
|
|
},
|
|
{
|
|
"epoch": 0.9975460122699387,
|
|
"grad_norm": 27.207499335475898,
|
|
"learning_rate": 6.743801652892561e-07,
|
|
"logits/chosen": -0.08935546875,
|
|
"logits/rejected": -0.2275390625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.1876,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 813
|
|
},
|
|
{
|
|
"epoch": 0.9987730061349693,
|
|
"grad_norm": 35.79902409901904,
|
|
"learning_rate": 6.739669421487602e-07,
|
|
"logits/chosen": -0.0634765625,
|
|
"logits/rejected": -0.142578125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3407,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": -0.166015625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 814
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"grad_norm": 42.42708489033231,
|
|
"learning_rate": 6.735537190082644e-07,
|
|
"logits/chosen": 0.006988525390625,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2248,
|
|
"rewards/accuracies": 0.8823530077934265,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 815
|
|
},
|
|
{
|
|
"epoch": 1.0012269938650307,
|
|
"grad_norm": 8.198778956705105,
|
|
"learning_rate": 6.731404958677685e-07,
|
|
"logits/chosen": -0.054931640625,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.0418,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -7.0,
|
|
"step": 816
|
|
},
|
|
{
|
|
"epoch": 1.0024539877300613,
|
|
"grad_norm": 10.034582862263894,
|
|
"learning_rate": 6.727272727272727e-07,
|
|
"logits/chosen": -0.0771484375,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.0562,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 7.84375,
|
|
"rewards/rejected": -6.84375,
|
|
"step": 817
|
|
},
|
|
{
|
|
"epoch": 1.003680981595092,
|
|
"grad_norm": 4.884188251750306,
|
|
"learning_rate": 6.723140495867768e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 818
|
|
},
|
|
{
|
|
"epoch": 1.0049079754601227,
|
|
"grad_norm": 11.241423829810111,
|
|
"learning_rate": 6.71900826446281e-07,
|
|
"logits/chosen": -0.05908203125,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0466,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.25,
|
|
"step": 819
|
|
},
|
|
{
|
|
"epoch": 1.0061349693251533,
|
|
"grad_norm": 7.660217134045902,
|
|
"learning_rate": 6.714876033057851e-07,
|
|
"logits/chosen": -0.025390625,
|
|
"logits/rejected": -0.158203125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0417,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 820
|
|
},
|
|
{
|
|
"epoch": 1.007361963190184,
|
|
"grad_norm": 15.209753822479508,
|
|
"learning_rate": 6.710743801652893e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.093,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 821
|
|
},
|
|
{
|
|
"epoch": 1.0085889570552147,
|
|
"grad_norm": 8.899845665402786,
|
|
"learning_rate": 6.706611570247933e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0405,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 8.4375,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 822
|
|
},
|
|
{
|
|
"epoch": 1.0098159509202453,
|
|
"grad_norm": 9.280453529135258,
|
|
"learning_rate": 6.702479338842975e-07,
|
|
"logits/chosen": -0.07763671875,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0447,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.28125,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 823
|
|
},
|
|
{
|
|
"epoch": 1.011042944785276,
|
|
"grad_norm": 9.526356346868496,
|
|
"learning_rate": 6.698347107438016e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0254,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.3125,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 824
|
|
},
|
|
{
|
|
"epoch": 1.0122699386503067,
|
|
"grad_norm": 7.714029962461168,
|
|
"learning_rate": 6.694214876033058e-07,
|
|
"logits/chosen": 0.0263671875,
|
|
"logits/rejected": -0.2421875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0344,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -6.71875,
|
|
"step": 825
|
|
},
|
|
{
|
|
"epoch": 1.0134969325153373,
|
|
"grad_norm": 14.929023390632208,
|
|
"learning_rate": 6.690082644628099e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0788,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 7.8125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 826
|
|
},
|
|
{
|
|
"epoch": 1.014723926380368,
|
|
"grad_norm": 7.6760614756143335,
|
|
"learning_rate": 6.68595041322314e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.0463,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 827
|
|
},
|
|
{
|
|
"epoch": 1.0159509202453987,
|
|
"grad_norm": 11.290687439548616,
|
|
"learning_rate": 6.681818181818182e-07,
|
|
"logits/chosen": 0.02783203125,
|
|
"logits/rejected": -0.1650390625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.0672,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 8.0,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 828
|
|
},
|
|
{
|
|
"epoch": 1.0171779141104293,
|
|
"grad_norm": 14.625653305060517,
|
|
"learning_rate": 6.677685950413223e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0818,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -6.75,
|
|
"step": 829
|
|
},
|
|
{
|
|
"epoch": 1.01840490797546,
|
|
"grad_norm": 9.498059779706894,
|
|
"learning_rate": 6.673553719008265e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0476,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 2.515625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -7.40625,
|
|
"step": 830
|
|
},
|
|
{
|
|
"epoch": 1.019631901840491,
|
|
"grad_norm": 5.956361360927761,
|
|
"learning_rate": 6.669421487603305e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.205078125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.296875,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -7.75,
|
|
"step": 831
|
|
},
|
|
{
|
|
"epoch": 1.0208588957055216,
|
|
"grad_norm": 18.522639372015494,
|
|
"learning_rate": 6.665289256198347e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0719,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -7.125,
|
|
"step": 832
|
|
},
|
|
{
|
|
"epoch": 1.0220858895705522,
|
|
"grad_norm": 21.19037653129941,
|
|
"learning_rate": 6.661157024793388e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0952,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -7.5625,
|
|
"step": 833
|
|
},
|
|
{
|
|
"epoch": 1.023312883435583,
|
|
"grad_norm": 8.63167531108866,
|
|
"learning_rate": 6.65702479338843e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.12255859375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.037,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 7.78125,
|
|
"rewards/rejected": -7.1875,
|
|
"step": 834
|
|
},
|
|
{
|
|
"epoch": 1.0245398773006136,
|
|
"grad_norm": 9.923697620512455,
|
|
"learning_rate": 6.652892561983471e-07,
|
|
"logits/chosen": -0.045654296875,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0542,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.5,
|
|
"step": 835
|
|
},
|
|
{
|
|
"epoch": 1.0257668711656442,
|
|
"grad_norm": 16.19774857132627,
|
|
"learning_rate": 6.648760330578513e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0526,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 836
|
|
},
|
|
{
|
|
"epoch": 1.026993865030675,
|
|
"grad_norm": 8.012099562239197,
|
|
"learning_rate": 6.644628099173554e-07,
|
|
"logits/chosen": 0.00225830078125,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0454,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -7.8125,
|
|
"step": 837
|
|
},
|
|
{
|
|
"epoch": 1.0282208588957056,
|
|
"grad_norm": 7.474951818436832,
|
|
"learning_rate": 6.640495867768596e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0341,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 838
|
|
},
|
|
{
|
|
"epoch": 1.0294478527607362,
|
|
"grad_norm": 21.892352741432404,
|
|
"learning_rate": 6.636363636363636e-07,
|
|
"logits/chosen": -0.09423828125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0642,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -7.90625,
|
|
"step": 839
|
|
},
|
|
{
|
|
"epoch": 1.030674846625767,
|
|
"grad_norm": 10.435422644764143,
|
|
"learning_rate": 6.632231404958677e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0431,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 840
|
|
},
|
|
{
|
|
"epoch": 1.0319018404907976,
|
|
"grad_norm": 8.864374542310735,
|
|
"learning_rate": 6.628099173553718e-07,
|
|
"logits/chosen": -0.051513671875,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0487,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.47265625,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.625,
|
|
"step": 841
|
|
},
|
|
{
|
|
"epoch": 1.0331288343558283,
|
|
"grad_norm": 7.613434573367824,
|
|
"learning_rate": 6.62396694214876e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.046,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.87109375,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.0,
|
|
"step": 842
|
|
},
|
|
{
|
|
"epoch": 1.034355828220859,
|
|
"grad_norm": 6.2462766771315925,
|
|
"learning_rate": 6.619834710743801e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0352,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.25,
|
|
"step": 843
|
|
},
|
|
{
|
|
"epoch": 1.0355828220858896,
|
|
"grad_norm": 9.404365859872662,
|
|
"learning_rate": 6.615702479338842e-07,
|
|
"logits/chosen": -0.03955078125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.125,
|
|
"step": 844
|
|
},
|
|
{
|
|
"epoch": 1.0368098159509203,
|
|
"grad_norm": 10.78850901488838,
|
|
"learning_rate": 6.611570247933884e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0364,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 8.1875,
|
|
"rewards/rejected": -7.1875,
|
|
"step": 845
|
|
},
|
|
{
|
|
"epoch": 1.038036809815951,
|
|
"grad_norm": 10.165561087370767,
|
|
"learning_rate": 6.607438016528925e-07,
|
|
"logits/chosen": -0.013671875,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0424,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.125,
|
|
"step": 846
|
|
},
|
|
{
|
|
"epoch": 1.0392638036809816,
|
|
"grad_norm": 7.940010314439831,
|
|
"learning_rate": 6.603305785123967e-07,
|
|
"logits/chosen": -0.038818359375,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0412,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -7.40625,
|
|
"step": 847
|
|
},
|
|
{
|
|
"epoch": 1.0404907975460123,
|
|
"grad_norm": 6.116154548492314,
|
|
"learning_rate": 6.599173553719007e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0369,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.375,
|
|
"step": 848
|
|
},
|
|
{
|
|
"epoch": 1.041717791411043,
|
|
"grad_norm": 8.890378518155226,
|
|
"learning_rate": 6.595041322314049e-07,
|
|
"logits/chosen": -0.08740234375,
|
|
"logits/rejected": -0.1708984375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0468,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -7.875,
|
|
"step": 849
|
|
},
|
|
{
|
|
"epoch": 1.0429447852760736,
|
|
"grad_norm": 7.4534731915625825,
|
|
"learning_rate": 6.59090909090909e-07,
|
|
"logits/chosen": -0.123046875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0253,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.0,
|
|
"step": 850
|
|
},
|
|
{
|
|
"epoch": 1.0441717791411043,
|
|
"grad_norm": 5.2289352654985475,
|
|
"learning_rate": 6.586776859504132e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0197,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -8.0,
|
|
"step": 851
|
|
},
|
|
{
|
|
"epoch": 1.045398773006135,
|
|
"grad_norm": 6.554354274429401,
|
|
"learning_rate": 6.582644628099173e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0294,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -7.375,
|
|
"step": 852
|
|
},
|
|
{
|
|
"epoch": 1.0466257668711656,
|
|
"grad_norm": 7.540480793038232,
|
|
"learning_rate": 6.578512396694215e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0337,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -7.3125,
|
|
"step": 853
|
|
},
|
|
{
|
|
"epoch": 1.0478527607361963,
|
|
"grad_norm": 16.233884522420407,
|
|
"learning_rate": 6.574380165289256e-07,
|
|
"logits/chosen": -0.06298828125,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0606,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -7.0625,
|
|
"step": 854
|
|
},
|
|
{
|
|
"epoch": 1.049079754601227,
|
|
"grad_norm": 10.737916852469844,
|
|
"learning_rate": 6.570247933884298e-07,
|
|
"logits/chosen": -0.0030975341796875,
|
|
"logits/rejected": -0.1455078125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0527,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -7.25,
|
|
"step": 855
|
|
},
|
|
{
|
|
"epoch": 1.0503067484662576,
|
|
"grad_norm": 15.766661594034929,
|
|
"learning_rate": 6.566115702479339e-07,
|
|
"logits/chosen": -0.10498046875,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0645,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -7.3125,
|
|
"step": 856
|
|
},
|
|
{
|
|
"epoch": 1.0515337423312883,
|
|
"grad_norm": 6.407676432361981,
|
|
"learning_rate": 6.56198347107438e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0364,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -7.75,
|
|
"step": 857
|
|
},
|
|
{
|
|
"epoch": 1.052760736196319,
|
|
"grad_norm": 6.0431977599113,
|
|
"learning_rate": 6.557851239669421e-07,
|
|
"logits/chosen": -0.0947265625,
|
|
"logits/rejected": -0.16015625,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.0322,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 858
|
|
},
|
|
{
|
|
"epoch": 1.0539877300613496,
|
|
"grad_norm": 9.301960440488351,
|
|
"learning_rate": 6.553719008264463e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.039,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.125,
|
|
"step": 859
|
|
},
|
|
{
|
|
"epoch": 1.0552147239263803,
|
|
"grad_norm": 9.970744151054586,
|
|
"learning_rate": 6.549586776859504e-07,
|
|
"logits/chosen": -0.0478515625,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0571,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.40625,
|
|
"rewards/margins": 7.84375,
|
|
"rewards/rejected": -7.4375,
|
|
"step": 860
|
|
},
|
|
{
|
|
"epoch": 1.056441717791411,
|
|
"grad_norm": 7.960815372398909,
|
|
"learning_rate": 6.545454545454546e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0389,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -7.96875,
|
|
"step": 861
|
|
},
|
|
{
|
|
"epoch": 1.0576687116564418,
|
|
"grad_norm": 9.866193192340742,
|
|
"learning_rate": 6.541322314049587e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0401,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.326171875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 862
|
|
},
|
|
{
|
|
"epoch": 1.0588957055214725,
|
|
"grad_norm": 4.394542482787353,
|
|
"learning_rate": 6.537190082644628e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0174,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 863
|
|
},
|
|
{
|
|
"epoch": 1.0601226993865032,
|
|
"grad_norm": 5.92275770140903,
|
|
"learning_rate": 6.53305785123967e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.032,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.125,
|
|
"step": 864
|
|
},
|
|
{
|
|
"epoch": 1.0613496932515338,
|
|
"grad_norm": 7.365814782107553,
|
|
"learning_rate": 6.52892561983471e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0276,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.375,
|
|
"step": 865
|
|
},
|
|
{
|
|
"epoch": 1.0625766871165645,
|
|
"grad_norm": 6.57470003109268,
|
|
"learning_rate": 6.524793388429751e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.029,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.87109375,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 866
|
|
},
|
|
{
|
|
"epoch": 1.0638036809815952,
|
|
"grad_norm": 13.828769179585446,
|
|
"learning_rate": 6.520661157024792e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0764,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.59765625,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 867
|
|
},
|
|
{
|
|
"epoch": 1.0650306748466258,
|
|
"grad_norm": 7.774577066294867,
|
|
"learning_rate": 6.516528925619834e-07,
|
|
"logits/chosen": 0.03515625,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.0474,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.380859375,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 868
|
|
},
|
|
{
|
|
"epoch": 1.0662576687116565,
|
|
"grad_norm": 6.055159550304544,
|
|
"learning_rate": 6.512396694214875e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.032,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 869
|
|
},
|
|
{
|
|
"epoch": 1.0674846625766872,
|
|
"grad_norm": 9.582774197578802,
|
|
"learning_rate": 6.508264462809917e-07,
|
|
"logits/chosen": -0.0341796875,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0416,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.71875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 870
|
|
},
|
|
{
|
|
"epoch": 1.0687116564417178,
|
|
"grad_norm": 19.609600220177253,
|
|
"learning_rate": 6.504132231404958e-07,
|
|
"logits/chosen": -0.037109375,
|
|
"logits/rejected": -0.2021484375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0693,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.283203125,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.375,
|
|
"step": 871
|
|
},
|
|
{
|
|
"epoch": 1.0699386503067485,
|
|
"grad_norm": 7.894961804284608,
|
|
"learning_rate": 6.5e-07,
|
|
"logits/chosen": -0.06103515625,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0343,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 872
|
|
},
|
|
{
|
|
"epoch": 1.0711656441717792,
|
|
"grad_norm": 8.583663392401974,
|
|
"learning_rate": 6.495867768595041e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0424,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 873
|
|
},
|
|
{
|
|
"epoch": 1.0723926380368098,
|
|
"grad_norm": 22.6036995812977,
|
|
"learning_rate": 6.491735537190083e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0788,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.125,
|
|
"step": 874
|
|
},
|
|
{
|
|
"epoch": 1.0736196319018405,
|
|
"grad_norm": 8.377642182505433,
|
|
"learning_rate": 6.487603305785123e-07,
|
|
"logits/chosen": -0.053955078125,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0367,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 875
|
|
},
|
|
{
|
|
"epoch": 1.0748466257668712,
|
|
"grad_norm": 8.79025541196024,
|
|
"learning_rate": 6.483471074380165e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0384,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 876
|
|
},
|
|
{
|
|
"epoch": 1.0760736196319018,
|
|
"grad_norm": 8.043418310108212,
|
|
"learning_rate": 6.479338842975206e-07,
|
|
"logits/chosen": 0.00104522705078125,
|
|
"logits/rejected": -0.1484375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0322,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.109375,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 877
|
|
},
|
|
{
|
|
"epoch": 1.0773006134969325,
|
|
"grad_norm": 12.824294807360966,
|
|
"learning_rate": 6.475206611570248e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0715,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.0230712890625,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -8.875,
|
|
"step": 878
|
|
},
|
|
{
|
|
"epoch": 1.0785276073619632,
|
|
"grad_norm": 7.194209272083771,
|
|
"learning_rate": 6.471074380165289e-07,
|
|
"logits/chosen": -0.042236328125,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0258,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 879
|
|
},
|
|
{
|
|
"epoch": 1.0797546012269938,
|
|
"grad_norm": 11.418531834795504,
|
|
"learning_rate": 6.46694214876033e-07,
|
|
"logits/chosen": -0.05126953125,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0485,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.1513671875,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 880
|
|
},
|
|
{
|
|
"epoch": 1.0809815950920245,
|
|
"grad_norm": 9.556772131268216,
|
|
"learning_rate": 6.462809917355372e-07,
|
|
"logits/chosen": -0.0159912109375,
|
|
"logits/rejected": -0.1552734375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.0536,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.51953125,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -7.53125,
|
|
"step": 881
|
|
},
|
|
{
|
|
"epoch": 1.0822085889570552,
|
|
"grad_norm": 18.827115566351896,
|
|
"learning_rate": 6.458677685950413e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0799,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -7.75,
|
|
"step": 882
|
|
},
|
|
{
|
|
"epoch": 1.0834355828220859,
|
|
"grad_norm": 7.135173761493492,
|
|
"learning_rate": 6.454545454545455e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0236,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.0625,
|
|
"step": 883
|
|
},
|
|
{
|
|
"epoch": 1.0846625766871165,
|
|
"grad_norm": 6.05359364559935,
|
|
"learning_rate": 6.450413223140495e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0245,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -7.875,
|
|
"step": 884
|
|
},
|
|
{
|
|
"epoch": 1.0858895705521472,
|
|
"grad_norm": 5.6232310801245005,
|
|
"learning_rate": 6.446280991735537e-07,
|
|
"logits/chosen": -0.1103515625,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0379,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.125,
|
|
"step": 885
|
|
},
|
|
{
|
|
"epoch": 1.0871165644171779,
|
|
"grad_norm": 6.935465868744795,
|
|
"learning_rate": 6.442148760330578e-07,
|
|
"logits/chosen": -0.01446533203125,
|
|
"logits/rejected": -0.2216796875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0364,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 8.125,
|
|
"rewards/rejected": -6.78125,
|
|
"step": 886
|
|
},
|
|
{
|
|
"epoch": 1.0883435582822085,
|
|
"grad_norm": 12.457264810968242,
|
|
"learning_rate": 6.43801652892562e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0531,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.125,
|
|
"step": 887
|
|
},
|
|
{
|
|
"epoch": 1.0895705521472392,
|
|
"grad_norm": 7.601309603795988,
|
|
"learning_rate": 6.433884297520661e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0288,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -7.75,
|
|
"step": 888
|
|
},
|
|
{
|
|
"epoch": 1.0907975460122699,
|
|
"grad_norm": 14.441348890944878,
|
|
"learning_rate": 6.429752066115703e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0736,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -7.5625,
|
|
"step": 889
|
|
},
|
|
{
|
|
"epoch": 1.0920245398773005,
|
|
"grad_norm": 8.025329021025417,
|
|
"learning_rate": 6.425619834710744e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0409,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -7.25,
|
|
"step": 890
|
|
},
|
|
{
|
|
"epoch": 1.0932515337423312,
|
|
"grad_norm": 7.226732081823533,
|
|
"learning_rate": 6.421487603305786e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0437,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -8.25,
|
|
"step": 891
|
|
},
|
|
{
|
|
"epoch": 1.0944785276073619,
|
|
"grad_norm": 6.766671772655921,
|
|
"learning_rate": 6.417355371900826e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0387,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.375,
|
|
"step": 892
|
|
},
|
|
{
|
|
"epoch": 1.0957055214723925,
|
|
"grad_norm": 11.204408599005943,
|
|
"learning_rate": 6.413223140495867e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.0656,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -7.59375,
|
|
"step": 893
|
|
},
|
|
{
|
|
"epoch": 1.0969325153374232,
|
|
"grad_norm": 16.672634853583897,
|
|
"learning_rate": 6.409090909090908e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0518,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 894
|
|
},
|
|
{
|
|
"epoch": 1.098159509202454,
|
|
"grad_norm": 10.576892299813201,
|
|
"learning_rate": 6.40495867768595e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0381,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -7.75,
|
|
"step": 895
|
|
},
|
|
{
|
|
"epoch": 1.0993865030674848,
|
|
"grad_norm": 9.283472414691788,
|
|
"learning_rate": 6.400826446280991e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0386,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 896
|
|
},
|
|
{
|
|
"epoch": 1.1006134969325154,
|
|
"grad_norm": 8.236402780568595,
|
|
"learning_rate": 6.396694214876032e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0353,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 897
|
|
},
|
|
{
|
|
"epoch": 1.101840490797546,
|
|
"grad_norm": 5.768384992257825,
|
|
"learning_rate": 6.392561983471074e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0283,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.125,
|
|
"step": 898
|
|
},
|
|
{
|
|
"epoch": 1.1030674846625768,
|
|
"grad_norm": 7.809914110956144,
|
|
"learning_rate": 6.388429752066115e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0367,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.09130859375,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 899
|
|
},
|
|
{
|
|
"epoch": 1.1042944785276074,
|
|
"grad_norm": 12.100772271435137,
|
|
"learning_rate": 6.384297520661157e-07,
|
|
"logits/chosen": -0.08203125,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.05,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.2578125,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -9.0,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.105521472392638,
|
|
"grad_norm": 10.462386125159243,
|
|
"learning_rate": 6.380165289256198e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -496.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.0395,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 901
|
|
},
|
|
{
|
|
"epoch": 1.1067484662576688,
|
|
"grad_norm": 12.728242995480565,
|
|
"learning_rate": 6.376033057851239e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0583,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.08251953125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.875,
|
|
"step": 902
|
|
},
|
|
{
|
|
"epoch": 1.1079754601226994,
|
|
"grad_norm": 10.08879572360814,
|
|
"learning_rate": 6.37190082644628e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0418,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.48046875,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 903
|
|
},
|
|
{
|
|
"epoch": 1.10920245398773,
|
|
"grad_norm": 5.729835019534765,
|
|
"learning_rate": 6.367768595041322e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0217,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 904
|
|
},
|
|
{
|
|
"epoch": 1.1104294478527608,
|
|
"grad_norm": 9.340557767143839,
|
|
"learning_rate": 6.363636363636363e-07,
|
|
"logits/chosen": -0.326171875,
|
|
"logits/rejected": -0.53515625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0322,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.25,
|
|
"step": 905
|
|
},
|
|
{
|
|
"epoch": 1.1116564417177914,
|
|
"grad_norm": 10.529203774813423,
|
|
"learning_rate": 6.359504132231405e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0497,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.53125,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 906
|
|
},
|
|
{
|
|
"epoch": 1.112883435582822,
|
|
"grad_norm": 10.47176282032967,
|
|
"learning_rate": 6.355371900826446e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0507,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.28125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 907
|
|
},
|
|
{
|
|
"epoch": 1.1141104294478528,
|
|
"grad_norm": 13.296041120798176,
|
|
"learning_rate": 6.351239669421488e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.028,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 908
|
|
},
|
|
{
|
|
"epoch": 1.1153374233128834,
|
|
"grad_norm": 12.388406412905939,
|
|
"learning_rate": 6.347107438016529e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -444.0,
|
|
"loss": 0.0529,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.72265625,
|
|
"rewards/margins": 11.625,
|
|
"rewards/rejected": -10.9375,
|
|
"step": 909
|
|
},
|
|
{
|
|
"epoch": 1.116564417177914,
|
|
"grad_norm": 13.15019677941546,
|
|
"learning_rate": 6.342975206611571e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0354,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 11.25,
|
|
"rewards/rejected": -10.125,
|
|
"step": 910
|
|
},
|
|
{
|
|
"epoch": 1.1177914110429448,
|
|
"grad_norm": 11.403214336902241,
|
|
"learning_rate": 6.338842975206611e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -468.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.0366,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 12.6875,
|
|
"rewards/rejected": -10.75,
|
|
"step": 911
|
|
},
|
|
{
|
|
"epoch": 1.1190184049079754,
|
|
"grad_norm": 16.126678956561374,
|
|
"learning_rate": 6.334710743801653e-07,
|
|
"logits/chosen": -0.0693359375,
|
|
"logits/rejected": -0.173828125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0496,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 912
|
|
},
|
|
{
|
|
"epoch": 1.120245398773006,
|
|
"grad_norm": 9.019433091214063,
|
|
"learning_rate": 6.330578512396694e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0465,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 913
|
|
},
|
|
{
|
|
"epoch": 1.1214723926380368,
|
|
"grad_norm": 16.963904925476527,
|
|
"learning_rate": 6.326446280991736e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0703,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.625,
|
|
"step": 914
|
|
},
|
|
{
|
|
"epoch": 1.1226993865030674,
|
|
"grad_norm": 12.995765624030382,
|
|
"learning_rate": 6.322314049586777e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0501,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.0,
|
|
"step": 915
|
|
},
|
|
{
|
|
"epoch": 1.123926380368098,
|
|
"grad_norm": 9.59256792153742,
|
|
"learning_rate": 6.318181818181818e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0404,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 916
|
|
},
|
|
{
|
|
"epoch": 1.1251533742331288,
|
|
"grad_norm": 14.293688400955162,
|
|
"learning_rate": 6.31404958677686e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0725,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.3125,
|
|
"step": 917
|
|
},
|
|
{
|
|
"epoch": 1.1263803680981594,
|
|
"grad_norm": 18.127127152101544,
|
|
"learning_rate": 6.3099173553719e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0716,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 8.6875,
|
|
"rewards/rejected": -7.65625,
|
|
"step": 918
|
|
},
|
|
{
|
|
"epoch": 1.1276073619631901,
|
|
"grad_norm": 10.161171251939434,
|
|
"learning_rate": 6.305785123966943e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0503,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 8.6875,
|
|
"rewards/rejected": -8.0625,
|
|
"step": 919
|
|
},
|
|
{
|
|
"epoch": 1.1288343558282208,
|
|
"grad_norm": 14.568103847834227,
|
|
"learning_rate": 6.301652892561982e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0433,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -7.75,
|
|
"step": 920
|
|
},
|
|
{
|
|
"epoch": 1.1300613496932514,
|
|
"grad_norm": 8.847430569562087,
|
|
"learning_rate": 6.297520661157024e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0261,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -7.6875,
|
|
"step": 921
|
|
},
|
|
{
|
|
"epoch": 1.1312883435582821,
|
|
"grad_norm": 8.616765181243158,
|
|
"learning_rate": 6.293388429752065e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0395,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.0,
|
|
"step": 922
|
|
},
|
|
{
|
|
"epoch": 1.132515337423313,
|
|
"grad_norm": 6.330290904369006,
|
|
"learning_rate": 6.289256198347107e-07,
|
|
"logits/chosen": -0.08349609375,
|
|
"logits/rejected": -0.177734375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0273,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -6.96875,
|
|
"step": 923
|
|
},
|
|
{
|
|
"epoch": 1.1337423312883437,
|
|
"grad_norm": 9.815829187551568,
|
|
"learning_rate": 6.285123966942148e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0437,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -7.78125,
|
|
"step": 924
|
|
},
|
|
{
|
|
"epoch": 1.1349693251533743,
|
|
"grad_norm": 11.784732478540608,
|
|
"learning_rate": 6.28099173553719e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0353,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 925
|
|
},
|
|
{
|
|
"epoch": 1.136196319018405,
|
|
"grad_norm": 17.203199300612283,
|
|
"learning_rate": 6.276859504132231e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0784,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -7.78125,
|
|
"step": 926
|
|
},
|
|
{
|
|
"epoch": 1.1374233128834357,
|
|
"grad_norm": 15.464919936179866,
|
|
"learning_rate": 6.272727272727273e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0463,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -7.875,
|
|
"step": 927
|
|
},
|
|
{
|
|
"epoch": 1.1386503067484663,
|
|
"grad_norm": 6.902255998247618,
|
|
"learning_rate": 6.268595041322314e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0351,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -7.84375,
|
|
"step": 928
|
|
},
|
|
{
|
|
"epoch": 1.139877300613497,
|
|
"grad_norm": 14.054680245437687,
|
|
"learning_rate": 6.264462809917355e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.0442,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -7.84375,
|
|
"step": 929
|
|
},
|
|
{
|
|
"epoch": 1.1411042944785277,
|
|
"grad_norm": 24.500256242909163,
|
|
"learning_rate": 6.260330578512396e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0651,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.75,
|
|
"step": 930
|
|
},
|
|
{
|
|
"epoch": 1.1423312883435583,
|
|
"grad_norm": 10.156003050487454,
|
|
"learning_rate": 6.256198347107438e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0505,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -7.71875,
|
|
"step": 931
|
|
},
|
|
{
|
|
"epoch": 1.143558282208589,
|
|
"grad_norm": 7.912842796144831,
|
|
"learning_rate": 6.252066115702479e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0388,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.5,
|
|
"step": 932
|
|
},
|
|
{
|
|
"epoch": 1.1447852760736197,
|
|
"grad_norm": 16.615422598723214,
|
|
"learning_rate": 6.24793388429752e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0471,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -8.3125,
|
|
"step": 933
|
|
},
|
|
{
|
|
"epoch": 1.1460122699386504,
|
|
"grad_norm": 14.649539698998428,
|
|
"learning_rate": 6.243801652892562e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0468,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -7.8125,
|
|
"step": 934
|
|
},
|
|
{
|
|
"epoch": 1.147239263803681,
|
|
"grad_norm": 7.488663396791275,
|
|
"learning_rate": 6.239669421487603e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0257,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 935
|
|
},
|
|
{
|
|
"epoch": 1.1484662576687117,
|
|
"grad_norm": 13.575757740475465,
|
|
"learning_rate": 6.235537190082645e-07,
|
|
"logits/chosen": -0.044189453125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0525,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.41796875,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.125,
|
|
"step": 936
|
|
},
|
|
{
|
|
"epoch": 1.1496932515337424,
|
|
"grad_norm": 10.619891086269176,
|
|
"learning_rate": 6.231404958677685e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0369,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.25,
|
|
"step": 937
|
|
},
|
|
{
|
|
"epoch": 1.150920245398773,
|
|
"grad_norm": 11.97746302237756,
|
|
"learning_rate": 6.227272727272727e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.169921875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0752,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 938
|
|
},
|
|
{
|
|
"epoch": 1.1521472392638037,
|
|
"grad_norm": 11.842135595518723,
|
|
"learning_rate": 6.223140495867768e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0274,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 939
|
|
},
|
|
{
|
|
"epoch": 1.1533742331288344,
|
|
"grad_norm": 6.3495393211670645,
|
|
"learning_rate": 6.21900826446281e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0242,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -9.875,
|
|
"step": 940
|
|
},
|
|
{
|
|
"epoch": 1.154601226993865,
|
|
"grad_norm": 17.940432875214363,
|
|
"learning_rate": 6.214876033057851e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0556,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 941
|
|
},
|
|
{
|
|
"epoch": 1.1558282208588957,
|
|
"grad_norm": 10.502645604068354,
|
|
"learning_rate": 6.210743801652893e-07,
|
|
"logits/chosen": -0.3203125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0336,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 942
|
|
},
|
|
{
|
|
"epoch": 1.1570552147239264,
|
|
"grad_norm": 6.526570566319107,
|
|
"learning_rate": 6.206611570247934e-07,
|
|
"logits/chosen": -0.0751953125,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.027,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.63671875,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.625,
|
|
"step": 943
|
|
},
|
|
{
|
|
"epoch": 1.158282208588957,
|
|
"grad_norm": 7.056786357514435,
|
|
"learning_rate": 6.202479338842976e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0264,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 944
|
|
},
|
|
{
|
|
"epoch": 1.1595092024539877,
|
|
"grad_norm": 10.73809466349613,
|
|
"learning_rate": 6.198347107438017e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0416,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.125,
|
|
"step": 945
|
|
},
|
|
{
|
|
"epoch": 1.1607361963190184,
|
|
"grad_norm": 7.022836532679102,
|
|
"learning_rate": 6.194214876033057e-07,
|
|
"logits/chosen": -0.09326171875,
|
|
"logits/rejected": -0.169921875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0327,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.62109375,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -7.875,
|
|
"step": 946
|
|
},
|
|
{
|
|
"epoch": 1.161963190184049,
|
|
"grad_norm": 11.545337320621988,
|
|
"learning_rate": 6.190082644628098e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0638,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -7.78125,
|
|
"step": 947
|
|
},
|
|
{
|
|
"epoch": 1.1631901840490797,
|
|
"grad_norm": 16.533338651381076,
|
|
"learning_rate": 6.18595041322314e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0553,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -7.625,
|
|
"step": 948
|
|
},
|
|
{
|
|
"epoch": 1.1644171779141104,
|
|
"grad_norm": 10.049032266591617,
|
|
"learning_rate": 6.181818181818181e-07,
|
|
"logits/chosen": -0.0888671875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0498,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -8.25,
|
|
"step": 949
|
|
},
|
|
{
|
|
"epoch": 1.165644171779141,
|
|
"grad_norm": 9.946858395715273,
|
|
"learning_rate": 6.177685950413223e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.0347,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -8.125,
|
|
"step": 950
|
|
},
|
|
{
|
|
"epoch": 1.1668711656441717,
|
|
"grad_norm": 8.78892412082473,
|
|
"learning_rate": 6.173553719008264e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0431,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.734375,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -7.875,
|
|
"step": 951
|
|
},
|
|
{
|
|
"epoch": 1.1680981595092024,
|
|
"grad_norm": 11.630136369956718,
|
|
"learning_rate": 6.169421487603305e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.57421875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0555,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.3125,
|
|
"step": 952
|
|
},
|
|
{
|
|
"epoch": 1.169325153374233,
|
|
"grad_norm": 10.447260099921499,
|
|
"learning_rate": 6.165289256198347e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.044,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -7.59375,
|
|
"step": 953
|
|
},
|
|
{
|
|
"epoch": 1.1705521472392637,
|
|
"grad_norm": 13.727338543385937,
|
|
"learning_rate": 6.161157024793388e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.0666,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.039306640625,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 954
|
|
},
|
|
{
|
|
"epoch": 1.1717791411042944,
|
|
"grad_norm": 14.923648123173107,
|
|
"learning_rate": 6.157024793388429e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.1767578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0515,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.125,
|
|
"step": 955
|
|
},
|
|
{
|
|
"epoch": 1.173006134969325,
|
|
"grad_norm": 7.9940420233037734,
|
|
"learning_rate": 6.15289256198347e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.0343,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -7.84375,
|
|
"step": 956
|
|
},
|
|
{
|
|
"epoch": 1.1742331288343557,
|
|
"grad_norm": 7.065002679153831,
|
|
"learning_rate": 6.148760330578512e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.031,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 957
|
|
},
|
|
{
|
|
"epoch": 1.1754601226993864,
|
|
"grad_norm": 10.90650359434899,
|
|
"learning_rate": 6.144628099173553e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0319,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -8.875,
|
|
"step": 958
|
|
},
|
|
{
|
|
"epoch": 1.1766871165644173,
|
|
"grad_norm": 7.826920221029951,
|
|
"learning_rate": 6.140495867768595e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0309,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.5,
|
|
"step": 959
|
|
},
|
|
{
|
|
"epoch": 1.177914110429448,
|
|
"grad_norm": 4.643420917768002,
|
|
"learning_rate": 6.136363636363636e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0221,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.03662109375,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -9.125,
|
|
"step": 960
|
|
},
|
|
{
|
|
"epoch": 1.1791411042944786,
|
|
"grad_norm": 6.952980511785281,
|
|
"learning_rate": 6.132231404958678e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0297,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 961
|
|
},
|
|
{
|
|
"epoch": 1.1803680981595093,
|
|
"grad_norm": 16.834157550822468,
|
|
"learning_rate": 6.128099173553719e-07,
|
|
"logits/chosen": -0.09130859375,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0445,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 962
|
|
},
|
|
{
|
|
"epoch": 1.18159509202454,
|
|
"grad_norm": 6.911866527354106,
|
|
"learning_rate": 6.123966942148761e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.029,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.625,
|
|
"step": 963
|
|
},
|
|
{
|
|
"epoch": 1.1828220858895706,
|
|
"grad_norm": 15.284583454778542,
|
|
"learning_rate": 6.119834710743801e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0325,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 964
|
|
},
|
|
{
|
|
"epoch": 1.1840490797546013,
|
|
"grad_norm": 14.626890231713732,
|
|
"learning_rate": 6.115702479338843e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0411,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.265625,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 965
|
|
},
|
|
{
|
|
"epoch": 1.185276073619632,
|
|
"grad_norm": 5.669188739019078,
|
|
"learning_rate": 6.111570247933884e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0308,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 966
|
|
},
|
|
{
|
|
"epoch": 1.1865030674846626,
|
|
"grad_norm": 5.096142341008849,
|
|
"learning_rate": 6.107438016528926e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0279,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -8.875,
|
|
"step": 967
|
|
},
|
|
{
|
|
"epoch": 1.1877300613496933,
|
|
"grad_norm": 6.118006395915871,
|
|
"learning_rate": 6.103305785123967e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0215,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -9.75,
|
|
"step": 968
|
|
},
|
|
{
|
|
"epoch": 1.188957055214724,
|
|
"grad_norm": 15.680989045225655,
|
|
"learning_rate": 6.099173553719008e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0555,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.71875,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 969
|
|
},
|
|
{
|
|
"epoch": 1.1901840490797546,
|
|
"grad_norm": 11.628185970418137,
|
|
"learning_rate": 6.09504132231405e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -494.0,
|
|
"loss": 0.0463,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.46875,
|
|
"rewards/margins": 22.75,
|
|
"rewards/rejected": -22.25,
|
|
"step": 970
|
|
},
|
|
{
|
|
"epoch": 1.1914110429447853,
|
|
"grad_norm": 9.816512004820881,
|
|
"learning_rate": 6.090909090909091e-07,
|
|
"logits/chosen": -0.1083984375,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0504,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.2392578125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -9.25,
|
|
"step": 971
|
|
},
|
|
{
|
|
"epoch": 1.192638036809816,
|
|
"grad_norm": 12.79931603977788,
|
|
"learning_rate": 6.086776859504133e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.0541,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.08056640625,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -8.625,
|
|
"step": 972
|
|
},
|
|
{
|
|
"epoch": 1.1938650306748466,
|
|
"grad_norm": 7.4966267701539575,
|
|
"learning_rate": 6.082644628099172e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.0236,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.5,
|
|
"step": 973
|
|
},
|
|
{
|
|
"epoch": 1.1950920245398773,
|
|
"grad_norm": 7.469181577002901,
|
|
"learning_rate": 6.078512396694214e-07,
|
|
"logits/chosen": -0.3203125,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0288,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 974
|
|
},
|
|
{
|
|
"epoch": 1.196319018404908,
|
|
"grad_norm": 10.73459043348965,
|
|
"learning_rate": 6.074380165289255e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.0346,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.51171875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 975
|
|
},
|
|
{
|
|
"epoch": 1.1975460122699386,
|
|
"grad_norm": 10.353571993398708,
|
|
"learning_rate": 6.070247933884297e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0449,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.25,
|
|
"step": 976
|
|
},
|
|
{
|
|
"epoch": 1.1987730061349693,
|
|
"grad_norm": 11.94762126827027,
|
|
"learning_rate": 6.066115702479338e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.021,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -9.75,
|
|
"step": 977
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"grad_norm": 14.933543180654313,
|
|
"learning_rate": 6.06198347107438e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0782,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.142578125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.25,
|
|
"step": 978
|
|
},
|
|
{
|
|
"epoch": 1.2012269938650306,
|
|
"grad_norm": 8.010892791144792,
|
|
"learning_rate": 6.057851239669421e-07,
|
|
"logits/chosen": -0.10498046875,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.03,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 979
|
|
},
|
|
{
|
|
"epoch": 1.2024539877300613,
|
|
"grad_norm": 3.2107623273524264,
|
|
"learning_rate": 6.053719008264463e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0158,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.125,
|
|
"step": 980
|
|
},
|
|
{
|
|
"epoch": 1.203680981595092,
|
|
"grad_norm": 12.258047105088115,
|
|
"learning_rate": 6.049586776859504e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0359,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 981
|
|
},
|
|
{
|
|
"epoch": 1.2049079754601226,
|
|
"grad_norm": 9.053722996745554,
|
|
"learning_rate": 6.045454545454545e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0361,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 982
|
|
},
|
|
{
|
|
"epoch": 1.2061349693251533,
|
|
"grad_norm": 17.32577717842402,
|
|
"learning_rate": 6.041322314049586e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0812,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.98046875,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -7.875,
|
|
"step": 983
|
|
},
|
|
{
|
|
"epoch": 1.207361963190184,
|
|
"grad_norm": 11.80473349256791,
|
|
"learning_rate": 6.037190082644628e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0559,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 984
|
|
},
|
|
{
|
|
"epoch": 1.2085889570552146,
|
|
"grad_norm": 11.330757544836207,
|
|
"learning_rate": 6.033057851239669e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0436,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.5,
|
|
"step": 985
|
|
},
|
|
{
|
|
"epoch": 1.2098159509202455,
|
|
"grad_norm": 8.85326566841804,
|
|
"learning_rate": 6.02892561983471e-07,
|
|
"logits/chosen": -0.025146484375,
|
|
"logits/rejected": -0.17578125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -7.9375,
|
|
"step": 986
|
|
},
|
|
{
|
|
"epoch": 1.2110429447852762,
|
|
"grad_norm": 9.318859526286015,
|
|
"learning_rate": 6.024793388429752e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0387,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -8.875,
|
|
"step": 987
|
|
},
|
|
{
|
|
"epoch": 1.2122699386503069,
|
|
"grad_norm": 6.388583557150669,
|
|
"learning_rate": 6.020661157024793e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0215,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.875,
|
|
"step": 988
|
|
},
|
|
{
|
|
"epoch": 1.2134969325153375,
|
|
"grad_norm": 12.51162569498335,
|
|
"learning_rate": 6.016528925619835e-07,
|
|
"logits/chosen": -0.0849609375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.0567,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.0625,
|
|
"step": 989
|
|
},
|
|
{
|
|
"epoch": 1.2147239263803682,
|
|
"grad_norm": 4.189958727493459,
|
|
"learning_rate": 6.012396694214876e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0241,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.5,
|
|
"step": 990
|
|
},
|
|
{
|
|
"epoch": 1.2159509202453989,
|
|
"grad_norm": 18.828022363797565,
|
|
"learning_rate": 6.008264462809917e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0424,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.30859375,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.875,
|
|
"step": 991
|
|
},
|
|
{
|
|
"epoch": 1.2171779141104295,
|
|
"grad_norm": 8.29806668215805,
|
|
"learning_rate": 6.004132231404958e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0359,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 992
|
|
},
|
|
{
|
|
"epoch": 1.2184049079754602,
|
|
"grad_norm": 12.31585651837839,
|
|
"learning_rate": 6e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0406,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -8.625,
|
|
"step": 993
|
|
},
|
|
{
|
|
"epoch": 1.2196319018404909,
|
|
"grad_norm": 12.63420819024809,
|
|
"learning_rate": 5.995867768595041e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0426,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.9453125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 994
|
|
},
|
|
{
|
|
"epoch": 1.2208588957055215,
|
|
"grad_norm": 12.719739688622587,
|
|
"learning_rate": 5.991735537190083e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0336,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.328125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 995
|
|
},
|
|
{
|
|
"epoch": 1.2220858895705522,
|
|
"grad_norm": 23.22282394116444,
|
|
"learning_rate": 5.987603305785124e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0613,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.30078125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 996
|
|
},
|
|
{
|
|
"epoch": 1.2233128834355829,
|
|
"grad_norm": 8.986531086985625,
|
|
"learning_rate": 5.983471074380166e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0299,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 997
|
|
},
|
|
{
|
|
"epoch": 1.2245398773006135,
|
|
"grad_norm": 9.853389441972034,
|
|
"learning_rate": 5.979338842975207e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0293,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 998
|
|
},
|
|
{
|
|
"epoch": 1.2257668711656442,
|
|
"grad_norm": 9.071764780193842,
|
|
"learning_rate": 5.975206611570249e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0317,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 999
|
|
},
|
|
{
|
|
"epoch": 1.2269938650306749,
|
|
"grad_norm": 10.178820112780551,
|
|
"learning_rate": 5.971074380165288e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.22265625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0527,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.2282208588957055,
|
|
"grad_norm": 11.96373096698322,
|
|
"learning_rate": 5.96694214876033e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0421,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1001
|
|
},
|
|
{
|
|
"epoch": 1.2294478527607362,
|
|
"grad_norm": 18.86424560815367,
|
|
"learning_rate": 5.962809917355371e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0593,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1002
|
|
},
|
|
{
|
|
"epoch": 1.2306748466257669,
|
|
"grad_norm": 10.998654938499559,
|
|
"learning_rate": 5.958677685950413e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0453,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1003
|
|
},
|
|
{
|
|
"epoch": 1.2319018404907975,
|
|
"grad_norm": 9.688517049359225,
|
|
"learning_rate": 5.954545454545454e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.0372,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 1004
|
|
},
|
|
{
|
|
"epoch": 1.2331288343558282,
|
|
"grad_norm": 7.568508066990736,
|
|
"learning_rate": 5.950413223140495e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0375,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1005
|
|
},
|
|
{
|
|
"epoch": 1.2343558282208589,
|
|
"grad_norm": 8.043981667739809,
|
|
"learning_rate": 5.946280991735537e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0343,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1006
|
|
},
|
|
{
|
|
"epoch": 1.2355828220858895,
|
|
"grad_norm": 9.076389557111685,
|
|
"learning_rate": 5.942148760330578e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0392,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1007
|
|
},
|
|
{
|
|
"epoch": 1.2368098159509202,
|
|
"grad_norm": 6.716133338993411,
|
|
"learning_rate": 5.93801652892562e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0358,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.30859375,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1008
|
|
},
|
|
{
|
|
"epoch": 1.2380368098159509,
|
|
"grad_norm": 14.243886624679257,
|
|
"learning_rate": 5.93388429752066e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0426,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1009
|
|
},
|
|
{
|
|
"epoch": 1.2392638036809815,
|
|
"grad_norm": 9.882437975002198,
|
|
"learning_rate": 5.929752066115702e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0465,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.1669921875,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1010
|
|
},
|
|
{
|
|
"epoch": 1.2404907975460122,
|
|
"grad_norm": 9.406774546540289,
|
|
"learning_rate": 5.925619834710743e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0333,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.486328125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1011
|
|
},
|
|
{
|
|
"epoch": 1.2417177914110429,
|
|
"grad_norm": 20.525112873834704,
|
|
"learning_rate": 5.921487603305785e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0875,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1012
|
|
},
|
|
{
|
|
"epoch": 1.2429447852760735,
|
|
"grad_norm": 7.641376288180867,
|
|
"learning_rate": 5.917355371900826e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.189453125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0401,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.2060546875,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1013
|
|
},
|
|
{
|
|
"epoch": 1.2441717791411042,
|
|
"grad_norm": 9.757416005438232,
|
|
"learning_rate": 5.913223140495868e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0377,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2353515625,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1014
|
|
},
|
|
{
|
|
"epoch": 1.2453987730061349,
|
|
"grad_norm": 10.878285072475228,
|
|
"learning_rate": 5.909090909090909e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0409,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1015
|
|
},
|
|
{
|
|
"epoch": 1.2466257668711656,
|
|
"grad_norm": 19.74212801251089,
|
|
"learning_rate": 5.904958677685951e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0846,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.3125,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -8.25,
|
|
"step": 1016
|
|
},
|
|
{
|
|
"epoch": 1.2478527607361962,
|
|
"grad_norm": 4.191293823067966,
|
|
"learning_rate": 5.900826446280992e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0156,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1017
|
|
},
|
|
{
|
|
"epoch": 1.2490797546012269,
|
|
"grad_norm": 11.578361944682054,
|
|
"learning_rate": 5.896694214876033e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0412,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.22265625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1018
|
|
},
|
|
{
|
|
"epoch": 1.2503067484662576,
|
|
"grad_norm": 7.246515197407325,
|
|
"learning_rate": 5.892561983471074e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.036,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5390625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1019
|
|
},
|
|
{
|
|
"epoch": 1.2515337423312882,
|
|
"grad_norm": 6.165996229294919,
|
|
"learning_rate": 5.888429752066116e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0328,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.59765625,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 1020
|
|
},
|
|
{
|
|
"epoch": 1.252760736196319,
|
|
"grad_norm": 10.267585001757054,
|
|
"learning_rate": 5.884297520661157e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0427,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.052490234375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1021
|
|
},
|
|
{
|
|
"epoch": 1.2539877300613496,
|
|
"grad_norm": 7.5599891875695855,
|
|
"learning_rate": 5.880165289256198e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0318,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1022
|
|
},
|
|
{
|
|
"epoch": 1.2552147239263804,
|
|
"grad_norm": 10.663359264262175,
|
|
"learning_rate": 5.87603305785124e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0434,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1023
|
|
},
|
|
{
|
|
"epoch": 1.2564417177914111,
|
|
"grad_norm": 11.563173659414957,
|
|
"learning_rate": 5.871900826446281e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0406,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.45703125,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1024
|
|
},
|
|
{
|
|
"epoch": 1.2576687116564418,
|
|
"grad_norm": 12.17374619972164,
|
|
"learning_rate": 5.867768595041323e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0562,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1025
|
|
},
|
|
{
|
|
"epoch": 1.2588957055214725,
|
|
"grad_norm": 9.83101212902539,
|
|
"learning_rate": 5.863636363636362e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0444,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.1669921875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1026
|
|
},
|
|
{
|
|
"epoch": 1.2601226993865031,
|
|
"grad_norm": 8.149400690675483,
|
|
"learning_rate": 5.859504132231404e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.49609375,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0292,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1027
|
|
},
|
|
{
|
|
"epoch": 1.2613496932515338,
|
|
"grad_norm": 13.750882151939985,
|
|
"learning_rate": 5.855371900826445e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0391,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1028
|
|
},
|
|
{
|
|
"epoch": 1.2625766871165645,
|
|
"grad_norm": 9.726306945673553,
|
|
"learning_rate": 5.851239669421487e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0359,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.44921875,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1029
|
|
},
|
|
{
|
|
"epoch": 1.2638036809815951,
|
|
"grad_norm": 7.652562840964626,
|
|
"learning_rate": 5.847107438016528e-07,
|
|
"logits/chosen": -0.3828125,
|
|
"logits/rejected": -0.470703125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0241,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1030
|
|
},
|
|
{
|
|
"epoch": 1.2650306748466258,
|
|
"grad_norm": 11.561043524760928,
|
|
"learning_rate": 5.84297520661157e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0708,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1031
|
|
},
|
|
{
|
|
"epoch": 1.2662576687116565,
|
|
"grad_norm": 4.254209446526944,
|
|
"learning_rate": 5.838842975206611e-07,
|
|
"logits/chosen": -0.0537109375,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0199,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1032
|
|
},
|
|
{
|
|
"epoch": 1.2674846625766871,
|
|
"grad_norm": 13.151571425602718,
|
|
"learning_rate": 5.834710743801653e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0567,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.462890625,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1033
|
|
},
|
|
{
|
|
"epoch": 1.2687116564417178,
|
|
"grad_norm": 4.560607426524578,
|
|
"learning_rate": 5.830578512396694e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -462.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0185,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1034
|
|
},
|
|
{
|
|
"epoch": 1.2699386503067485,
|
|
"grad_norm": 8.028525924524638,
|
|
"learning_rate": 5.826446280991735e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0256,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1035
|
|
},
|
|
{
|
|
"epoch": 1.2711656441717791,
|
|
"grad_norm": 21.358423278389765,
|
|
"learning_rate": 5.822314049586776e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0752,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1036
|
|
},
|
|
{
|
|
"epoch": 1.2723926380368098,
|
|
"grad_norm": 6.0826820143061875,
|
|
"learning_rate": 5.818181818181818e-07,
|
|
"logits/chosen": -0.349609375,
|
|
"logits/rejected": -0.48828125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0243,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1037
|
|
},
|
|
{
|
|
"epoch": 1.2736196319018405,
|
|
"grad_norm": 14.998343449845411,
|
|
"learning_rate": 5.814049586776859e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0378,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1038
|
|
},
|
|
{
|
|
"epoch": 1.2748466257668711,
|
|
"grad_norm": 13.369802004135122,
|
|
"learning_rate": 5.809917355371901e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0575,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1039
|
|
},
|
|
{
|
|
"epoch": 1.2760736196319018,
|
|
"grad_norm": 9.714563154915252,
|
|
"learning_rate": 5.805785123966942e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.54296875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0337,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1040
|
|
},
|
|
{
|
|
"epoch": 1.2773006134969325,
|
|
"grad_norm": 3.5220505940476023,
|
|
"learning_rate": 5.801652892561983e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0133,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1041
|
|
},
|
|
{
|
|
"epoch": 1.2785276073619631,
|
|
"grad_norm": 8.554988076729757,
|
|
"learning_rate": 5.797520661157025e-07,
|
|
"logits/chosen": -0.36328125,
|
|
"logits/rejected": -0.56640625,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0274,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1042
|
|
},
|
|
{
|
|
"epoch": 1.2797546012269938,
|
|
"grad_norm": 20.659630475510642,
|
|
"learning_rate": 5.793388429752066e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0819,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": -0.0654296875,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1043
|
|
},
|
|
{
|
|
"epoch": 1.2809815950920245,
|
|
"grad_norm": 5.917507321013971,
|
|
"learning_rate": 5.789256198347107e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0194,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.8515625,
|
|
"rewards/margins": 11.25,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1044
|
|
},
|
|
{
|
|
"epoch": 1.2822085889570551,
|
|
"grad_norm": 7.485314427905281,
|
|
"learning_rate": 5.785123966942148e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.58203125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0264,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1045
|
|
},
|
|
{
|
|
"epoch": 1.2834355828220858,
|
|
"grad_norm": 12.428392386378055,
|
|
"learning_rate": 5.78099173553719e-07,
|
|
"logits/chosen": -0.07861328125,
|
|
"logits/rejected": -0.234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0553,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1046
|
|
},
|
|
{
|
|
"epoch": 1.2846625766871167,
|
|
"grad_norm": 11.865177600712363,
|
|
"learning_rate": 5.776859504132231e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0402,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1047
|
|
},
|
|
{
|
|
"epoch": 1.2858895705521474,
|
|
"grad_norm": 6.121866387536907,
|
|
"learning_rate": 5.772727272727273e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0298,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1048
|
|
},
|
|
{
|
|
"epoch": 1.287116564417178,
|
|
"grad_norm": 8.637135190386745,
|
|
"learning_rate": 5.768595041322314e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0325,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1049
|
|
},
|
|
{
|
|
"epoch": 1.2883435582822087,
|
|
"grad_norm": 10.620633006209482,
|
|
"learning_rate": 5.764462809917356e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0493,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1050
|
|
},
|
|
{
|
|
"epoch": 1.2895705521472394,
|
|
"grad_norm": 10.362409008928314,
|
|
"learning_rate": 5.760330578512397e-07,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0338,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 11.5,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1051
|
|
},
|
|
{
|
|
"epoch": 1.29079754601227,
|
|
"grad_norm": 14.131368731885177,
|
|
"learning_rate": 5.756198347107439e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0653,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1052
|
|
},
|
|
{
|
|
"epoch": 1.2920245398773007,
|
|
"grad_norm": 14.407504352342405,
|
|
"learning_rate": 5.752066115702478e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0566,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1053
|
|
},
|
|
{
|
|
"epoch": 1.2932515337423314,
|
|
"grad_norm": 9.117800906524298,
|
|
"learning_rate": 5.74793388429752e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0331,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1054
|
|
},
|
|
{
|
|
"epoch": 1.294478527607362,
|
|
"grad_norm": 7.675278699888434,
|
|
"learning_rate": 5.743801652892561e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0246,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1055
|
|
},
|
|
{
|
|
"epoch": 1.2957055214723927,
|
|
"grad_norm": 3.067944816790383,
|
|
"learning_rate": 5.739669421487603e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0106,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.55859375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1056
|
|
},
|
|
{
|
|
"epoch": 1.2969325153374234,
|
|
"grad_norm": 12.782027941945014,
|
|
"learning_rate": 5.735537190082644e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.035,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1057
|
|
},
|
|
{
|
|
"epoch": 1.298159509202454,
|
|
"grad_norm": 15.736064132192451,
|
|
"learning_rate": 5.731404958677685e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0579,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1058
|
|
},
|
|
{
|
|
"epoch": 1.2993865030674847,
|
|
"grad_norm": 7.712747947073451,
|
|
"learning_rate": 5.727272727272727e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0396,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1059
|
|
},
|
|
{
|
|
"epoch": 1.3006134969325154,
|
|
"grad_norm": 9.603537828951685,
|
|
"learning_rate": 5.723140495867768e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0329,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1060
|
|
},
|
|
{
|
|
"epoch": 1.301840490797546,
|
|
"grad_norm": 9.662822016970589,
|
|
"learning_rate": 5.71900826446281e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0306,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.5390625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1061
|
|
},
|
|
{
|
|
"epoch": 1.3030674846625767,
|
|
"grad_norm": 14.609973288414059,
|
|
"learning_rate": 5.71487603305785e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.061,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1062
|
|
},
|
|
{
|
|
"epoch": 1.3042944785276074,
|
|
"grad_norm": 14.693317698121852,
|
|
"learning_rate": 5.710743801652892e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0673,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.357421875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1063
|
|
},
|
|
{
|
|
"epoch": 1.305521472392638,
|
|
"grad_norm": 8.05390792569244,
|
|
"learning_rate": 5.706611570247933e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.033,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1064
|
|
},
|
|
{
|
|
"epoch": 1.3067484662576687,
|
|
"grad_norm": 13.435103341979305,
|
|
"learning_rate": 5.702479338842975e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0483,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 11.625,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1065
|
|
},
|
|
{
|
|
"epoch": 1.3079754601226994,
|
|
"grad_norm": 11.67763310390295,
|
|
"learning_rate": 5.698347107438016e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.53125,
|
|
"logps/chosen": -460.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0331,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1066
|
|
},
|
|
{
|
|
"epoch": 1.30920245398773,
|
|
"grad_norm": 11.458119014088348,
|
|
"learning_rate": 5.694214876033058e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0451,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.205078125,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1067
|
|
},
|
|
{
|
|
"epoch": 1.3104294478527607,
|
|
"grad_norm": 12.122904851371764,
|
|
"learning_rate": 5.690082644628099e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0572,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1068
|
|
},
|
|
{
|
|
"epoch": 1.3116564417177914,
|
|
"grad_norm": 13.90827772396844,
|
|
"learning_rate": 5.685950413223141e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0636,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1069
|
|
},
|
|
{
|
|
"epoch": 1.312883435582822,
|
|
"grad_norm": 9.035804413971455,
|
|
"learning_rate": 5.681818181818182e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0257,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1070
|
|
},
|
|
{
|
|
"epoch": 1.3141104294478527,
|
|
"grad_norm": 15.897823579022534,
|
|
"learning_rate": 5.677685950413223e-07,
|
|
"logits/chosen": -0.310546875,
|
|
"logits/rejected": -0.49609375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0676,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1071
|
|
},
|
|
{
|
|
"epoch": 1.3153374233128834,
|
|
"grad_norm": 11.094124927576006,
|
|
"learning_rate": 5.673553719008264e-07,
|
|
"logits/chosen": -0.10693359375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0384,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1072
|
|
},
|
|
{
|
|
"epoch": 1.316564417177914,
|
|
"grad_norm": 6.165465266289324,
|
|
"learning_rate": 5.669421487603306e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.376953125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1073
|
|
},
|
|
{
|
|
"epoch": 1.3177914110429447,
|
|
"grad_norm": 7.146179133263634,
|
|
"learning_rate": 5.665289256198347e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0283,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1074
|
|
},
|
|
{
|
|
"epoch": 1.3190184049079754,
|
|
"grad_norm": 13.935494997647524,
|
|
"learning_rate": 5.661157024793388e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0375,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1075
|
|
},
|
|
{
|
|
"epoch": 1.320245398773006,
|
|
"grad_norm": 8.713084237897876,
|
|
"learning_rate": 5.65702479338843e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.035,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1076
|
|
},
|
|
{
|
|
"epoch": 1.3214723926380367,
|
|
"grad_norm": 9.94252039217701,
|
|
"learning_rate": 5.652892561983471e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -296.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.046,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.21875,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1077
|
|
},
|
|
{
|
|
"epoch": 1.3226993865030674,
|
|
"grad_norm": 6.777120253972798,
|
|
"learning_rate": 5.648760330578513e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0373,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.275390625,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1078
|
|
},
|
|
{
|
|
"epoch": 1.323926380368098,
|
|
"grad_norm": 11.901471778566595,
|
|
"learning_rate": 5.644628099173554e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0492,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1079
|
|
},
|
|
{
|
|
"epoch": 1.3251533742331287,
|
|
"grad_norm": 11.984770537411984,
|
|
"learning_rate": 5.640495867768594e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0537,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 1080
|
|
},
|
|
{
|
|
"epoch": 1.3263803680981594,
|
|
"grad_norm": 8.083487691388004,
|
|
"learning_rate": 5.636363636363635e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0362,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1081
|
|
},
|
|
{
|
|
"epoch": 1.32760736196319,
|
|
"grad_norm": 8.097180732642633,
|
|
"learning_rate": 5.632231404958677e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0345,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1082
|
|
},
|
|
{
|
|
"epoch": 1.3288343558282207,
|
|
"grad_norm": 9.41069278000276,
|
|
"learning_rate": 5.628099173553718e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.0417,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.54296875,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.0625,
|
|
"step": 1083
|
|
},
|
|
{
|
|
"epoch": 1.3300613496932514,
|
|
"grad_norm": 16.537222089616133,
|
|
"learning_rate": 5.62396694214876e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0741,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.66796875,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.3125,
|
|
"step": 1084
|
|
},
|
|
{
|
|
"epoch": 1.331288343558282,
|
|
"grad_norm": 13.530650671519162,
|
|
"learning_rate": 5.619834710743801e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.041,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.42578125,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.1875,
|
|
"step": 1085
|
|
},
|
|
{
|
|
"epoch": 1.332515337423313,
|
|
"grad_norm": 10.340910667933214,
|
|
"learning_rate": 5.615702479338843e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.0516,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1086
|
|
},
|
|
{
|
|
"epoch": 1.3337423312883436,
|
|
"grad_norm": 11.074217538800664,
|
|
"learning_rate": 5.611570247933884e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0304,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.1357421875,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1087
|
|
},
|
|
{
|
|
"epoch": 1.3349693251533743,
|
|
"grad_norm": 10.043414597825459,
|
|
"learning_rate": 5.607438016528926e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0358,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1088
|
|
},
|
|
{
|
|
"epoch": 1.336196319018405,
|
|
"grad_norm": 8.188705525008332,
|
|
"learning_rate": 5.603305785123966e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0421,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.1826171875,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1089
|
|
},
|
|
{
|
|
"epoch": 1.3374233128834356,
|
|
"grad_norm": 10.920187153056888,
|
|
"learning_rate": 5.599173553719008e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0406,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.12158203125,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1090
|
|
},
|
|
{
|
|
"epoch": 1.3386503067484663,
|
|
"grad_norm": 16.571836670852875,
|
|
"learning_rate": 5.595041322314049e-07,
|
|
"logits/chosen": -0.06787109375,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0548,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1091
|
|
},
|
|
{
|
|
"epoch": 1.339877300613497,
|
|
"grad_norm": 9.408368180232918,
|
|
"learning_rate": 5.590909090909091e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0364,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1092
|
|
},
|
|
{
|
|
"epoch": 1.3411042944785276,
|
|
"grad_norm": 10.48194056376553,
|
|
"learning_rate": 5.586776859504132e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0409,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.59765625,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1093
|
|
},
|
|
{
|
|
"epoch": 1.3423312883435583,
|
|
"grad_norm": 11.285665679958338,
|
|
"learning_rate": 5.582644628099173e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0419,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1094
|
|
},
|
|
{
|
|
"epoch": 1.343558282208589,
|
|
"grad_norm": 7.642179045492436,
|
|
"learning_rate": 5.578512396694215e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0211,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1095
|
|
},
|
|
{
|
|
"epoch": 1.3447852760736196,
|
|
"grad_norm": 23.401618912710124,
|
|
"learning_rate": 5.574380165289256e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0753,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.4453125,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1096
|
|
},
|
|
{
|
|
"epoch": 1.3460122699386503,
|
|
"grad_norm": 10.531709636211138,
|
|
"learning_rate": 5.570247933884298e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0337,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1097
|
|
},
|
|
{
|
|
"epoch": 1.347239263803681,
|
|
"grad_norm": 7.967487413880986,
|
|
"learning_rate": 5.566115702479338e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0256,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.50390625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1098
|
|
},
|
|
{
|
|
"epoch": 1.3484662576687116,
|
|
"grad_norm": 16.694151675253067,
|
|
"learning_rate": 5.56198347107438e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0718,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.158203125,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1099
|
|
},
|
|
{
|
|
"epoch": 1.3496932515337423,
|
|
"grad_norm": 8.445759709333183,
|
|
"learning_rate": 5.557851239669421e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0309,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.350920245398773,
|
|
"grad_norm": 12.485526733661077,
|
|
"learning_rate": 5.553719008264463e-07,
|
|
"logits/chosen": -0.05224609375,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0531,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1101
|
|
},
|
|
{
|
|
"epoch": 1.3521472392638036,
|
|
"grad_norm": 15.796919260156438,
|
|
"learning_rate": 5.549586776859504e-07,
|
|
"logits/chosen": -0.328125,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0546,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1102
|
|
},
|
|
{
|
|
"epoch": 1.3533742331288343,
|
|
"grad_norm": 15.342911659562413,
|
|
"learning_rate": 5.545454545454546e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0609,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1103
|
|
},
|
|
{
|
|
"epoch": 1.354601226993865,
|
|
"grad_norm": 13.61755323012672,
|
|
"learning_rate": 5.541322314049587e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0537,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1104
|
|
},
|
|
{
|
|
"epoch": 1.3558282208588956,
|
|
"grad_norm": 5.348773232223692,
|
|
"learning_rate": 5.537190082644629e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0196,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1105
|
|
},
|
|
{
|
|
"epoch": 1.3570552147239263,
|
|
"grad_norm": 9.23688326147907,
|
|
"learning_rate": 5.533057851239669e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0376,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.2158203125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1106
|
|
},
|
|
{
|
|
"epoch": 1.358282208588957,
|
|
"grad_norm": 10.337127278268197,
|
|
"learning_rate": 5.52892561983471e-07,
|
|
"logits/chosen": -0.0634765625,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0325,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1107
|
|
},
|
|
{
|
|
"epoch": 1.3595092024539877,
|
|
"grad_norm": 8.048172315481981,
|
|
"learning_rate": 5.524793388429751e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1108
|
|
},
|
|
{
|
|
"epoch": 1.3607361963190185,
|
|
"grad_norm": 9.951965164399498,
|
|
"learning_rate": 5.520661157024793e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.046,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1109
|
|
},
|
|
{
|
|
"epoch": 1.3619631901840492,
|
|
"grad_norm": 15.713244319910332,
|
|
"learning_rate": 5.516528925619834e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.045,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.62109375,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1110
|
|
},
|
|
{
|
|
"epoch": 1.3631901840490799,
|
|
"grad_norm": 6.6445097309291645,
|
|
"learning_rate": 5.512396694214875e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0239,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1111
|
|
},
|
|
{
|
|
"epoch": 1.3644171779141105,
|
|
"grad_norm": 10.698810133699197,
|
|
"learning_rate": 5.508264462809917e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0444,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1112
|
|
},
|
|
{
|
|
"epoch": 1.3656441717791412,
|
|
"grad_norm": 8.75625222872145,
|
|
"learning_rate": 5.504132231404958e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0389,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1113
|
|
},
|
|
{
|
|
"epoch": 1.3668711656441719,
|
|
"grad_norm": 14.820626389823108,
|
|
"learning_rate": 5.5e-07,
|
|
"logits/chosen": -0.33203125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0522,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.0311279296875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1114
|
|
},
|
|
{
|
|
"epoch": 1.3680981595092025,
|
|
"grad_norm": 10.806464298425707,
|
|
"learning_rate": 5.49586776859504e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0399,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1115
|
|
},
|
|
{
|
|
"epoch": 1.3693251533742332,
|
|
"grad_norm": 12.398755004150619,
|
|
"learning_rate": 5.491735537190082e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0584,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.51171875,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1116
|
|
},
|
|
{
|
|
"epoch": 1.3705521472392639,
|
|
"grad_norm": 7.926695771662054,
|
|
"learning_rate": 5.487603305785123e-07,
|
|
"logits/chosen": -0.328125,
|
|
"logits/rejected": -0.546875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0332,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1117
|
|
},
|
|
{
|
|
"epoch": 1.3717791411042946,
|
|
"grad_norm": 12.967005998204884,
|
|
"learning_rate": 5.483471074380165e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0647,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.1982421875,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1118
|
|
},
|
|
{
|
|
"epoch": 1.3730061349693252,
|
|
"grad_norm": 8.690846709616252,
|
|
"learning_rate": 5.479338842975206e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0439,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.314453125,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1119
|
|
},
|
|
{
|
|
"epoch": 1.3742331288343559,
|
|
"grad_norm": 12.612636966465232,
|
|
"learning_rate": 5.475206611570248e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0329,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1120
|
|
},
|
|
{
|
|
"epoch": 1.3754601226993866,
|
|
"grad_norm": 19.147713553971546,
|
|
"learning_rate": 5.471074380165289e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0934,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1121
|
|
},
|
|
{
|
|
"epoch": 1.3766871165644172,
|
|
"grad_norm": 10.685319730151653,
|
|
"learning_rate": 5.466942148760331e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0339,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.10986328125,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1122
|
|
},
|
|
{
|
|
"epoch": 1.377914110429448,
|
|
"grad_norm": 6.823747170287133,
|
|
"learning_rate": 5.462809917355372e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.1943359375,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1123
|
|
},
|
|
{
|
|
"epoch": 1.3791411042944786,
|
|
"grad_norm": 9.373358623542401,
|
|
"learning_rate": 5.458677685950413e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0364,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.45703125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1124
|
|
},
|
|
{
|
|
"epoch": 1.3803680981595092,
|
|
"grad_norm": 7.45084009741824,
|
|
"learning_rate": 5.454545454545454e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0305,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.2060546875,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1125
|
|
},
|
|
{
|
|
"epoch": 1.38159509202454,
|
|
"grad_norm": 7.2859452576082875,
|
|
"learning_rate": 5.450413223140496e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0311,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.36328125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1126
|
|
},
|
|
{
|
|
"epoch": 1.3828220858895706,
|
|
"grad_norm": 11.336660386449822,
|
|
"learning_rate": 5.446280991735537e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0491,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1127
|
|
},
|
|
{
|
|
"epoch": 1.3840490797546012,
|
|
"grad_norm": 14.677111428331852,
|
|
"learning_rate": 5.442148760330579e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0602,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.1298828125,
|
|
"rewards/margins": 8.6875,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1128
|
|
},
|
|
{
|
|
"epoch": 1.385276073619632,
|
|
"grad_norm": 6.03614944922332,
|
|
"learning_rate": 5.43801652892562e-07,
|
|
"logits/chosen": -0.0859375,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0196,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.66796875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1129
|
|
},
|
|
{
|
|
"epoch": 1.3865030674846626,
|
|
"grad_norm": 14.584337506543694,
|
|
"learning_rate": 5.433884297520661e-07,
|
|
"logits/chosen": -0.03369140625,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0467,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.78515625,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1130
|
|
},
|
|
{
|
|
"epoch": 1.3877300613496932,
|
|
"grad_norm": 7.936510681518484,
|
|
"learning_rate": 5.429752066115703e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.57421875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0305,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1131
|
|
},
|
|
{
|
|
"epoch": 1.388957055214724,
|
|
"grad_norm": 16.7491951172574,
|
|
"learning_rate": 5.425619834710744e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -486.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0446,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1132
|
|
},
|
|
{
|
|
"epoch": 1.3901840490797546,
|
|
"grad_norm": 11.205645603928188,
|
|
"learning_rate": 5.421487603305785e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.036,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1133
|
|
},
|
|
{
|
|
"epoch": 1.3914110429447852,
|
|
"grad_norm": 9.552272909962287,
|
|
"learning_rate": 5.417355371900825e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.0505,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.275390625,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1134
|
|
},
|
|
{
|
|
"epoch": 1.392638036809816,
|
|
"grad_norm": 7.685740650856805,
|
|
"learning_rate": 5.413223140495867e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0368,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1135
|
|
},
|
|
{
|
|
"epoch": 1.3938650306748466,
|
|
"grad_norm": 12.176746728587242,
|
|
"learning_rate": 5.409090909090908e-07,
|
|
"logits/chosen": -0.34375,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -458.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.038,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.8359375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1136
|
|
},
|
|
{
|
|
"epoch": 1.3950920245398772,
|
|
"grad_norm": 15.407220989147827,
|
|
"learning_rate": 5.40495867768595e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.53515625,
|
|
"logps/chosen": -458.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0725,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 1137
|
|
},
|
|
{
|
|
"epoch": 1.396319018404908,
|
|
"grad_norm": 11.595740842607116,
|
|
"learning_rate": 5.400826446280991e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0388,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.68359375,
|
|
"rewards/margins": 8.375,
|
|
"rewards/rejected": -7.71875,
|
|
"step": 1138
|
|
},
|
|
{
|
|
"epoch": 1.3975460122699386,
|
|
"grad_norm": 10.334632077667756,
|
|
"learning_rate": 5.396694214876033e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0395,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1139
|
|
},
|
|
{
|
|
"epoch": 1.3987730061349692,
|
|
"grad_norm": 11.12675728199778,
|
|
"learning_rate": 5.392561983471074e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0403,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1140
|
|
},
|
|
{
|
|
"epoch": 1.4,
|
|
"grad_norm": 8.2530815585702,
|
|
"learning_rate": 5.388429752066116e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0267,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1141
|
|
},
|
|
{
|
|
"epoch": 1.4012269938650306,
|
|
"grad_norm": 12.407791022701804,
|
|
"learning_rate": 5.384297520661156e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.041,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1142
|
|
},
|
|
{
|
|
"epoch": 1.4024539877300612,
|
|
"grad_norm": 13.426430614683477,
|
|
"learning_rate": 5.380165289256198e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0466,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.36328125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1143
|
|
},
|
|
{
|
|
"epoch": 1.403680981595092,
|
|
"grad_norm": 7.549793043721441,
|
|
"learning_rate": 5.376033057851239e-07,
|
|
"logits/chosen": -0.3203125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0251,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1144
|
|
},
|
|
{
|
|
"epoch": 1.4049079754601226,
|
|
"grad_norm": 17.724354463632945,
|
|
"learning_rate": 5.371900826446281e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0436,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1145
|
|
},
|
|
{
|
|
"epoch": 1.4061349693251532,
|
|
"grad_norm": 11.163232075653351,
|
|
"learning_rate": 5.367768595041322e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0389,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.34375,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1146
|
|
},
|
|
{
|
|
"epoch": 1.407361963190184,
|
|
"grad_norm": 5.048479771147863,
|
|
"learning_rate": 5.363636363636363e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0217,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1147
|
|
},
|
|
{
|
|
"epoch": 1.4085889570552146,
|
|
"grad_norm": 7.392908485762525,
|
|
"learning_rate": 5.359504132231405e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0259,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.59765625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1148
|
|
},
|
|
{
|
|
"epoch": 1.4098159509202455,
|
|
"grad_norm": 6.886214116353296,
|
|
"learning_rate": 5.355371900826446e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0229,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1149
|
|
},
|
|
{
|
|
"epoch": 1.4110429447852761,
|
|
"grad_norm": 8.17141221388661,
|
|
"learning_rate": 5.351239669421488e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.031,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.51953125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1150
|
|
},
|
|
{
|
|
"epoch": 1.4122699386503068,
|
|
"grad_norm": 7.150017999811418,
|
|
"learning_rate": 5.347107438016528e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0268,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.34765625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1151
|
|
},
|
|
{
|
|
"epoch": 1.4134969325153375,
|
|
"grad_norm": 8.591304255967401,
|
|
"learning_rate": 5.34297520661157e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0239,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1152
|
|
},
|
|
{
|
|
"epoch": 1.4147239263803681,
|
|
"grad_norm": 13.347174248835012,
|
|
"learning_rate": 5.338842975206611e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0527,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1153
|
|
},
|
|
{
|
|
"epoch": 1.4159509202453988,
|
|
"grad_norm": 10.014740664300271,
|
|
"learning_rate": 5.334710743801653e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0275,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1154
|
|
},
|
|
{
|
|
"epoch": 1.4171779141104295,
|
|
"grad_norm": 10.057526488489339,
|
|
"learning_rate": 5.330578512396694e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0378,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.0546875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1155
|
|
},
|
|
{
|
|
"epoch": 1.4184049079754601,
|
|
"grad_norm": 10.244804107297995,
|
|
"learning_rate": 5.326446280991736e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0267,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1156
|
|
},
|
|
{
|
|
"epoch": 1.4196319018404908,
|
|
"grad_norm": 21.212005914641182,
|
|
"learning_rate": 5.322314049586777e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0509,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1157
|
|
},
|
|
{
|
|
"epoch": 1.4208588957055215,
|
|
"grad_norm": 8.301949109409644,
|
|
"learning_rate": 5.318181818181819e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.036,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.69921875,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1158
|
|
},
|
|
{
|
|
"epoch": 1.4220858895705522,
|
|
"grad_norm": 16.99643166865997,
|
|
"learning_rate": 5.31404958677686e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0731,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.10546875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1159
|
|
},
|
|
{
|
|
"epoch": 1.4233128834355828,
|
|
"grad_norm": 12.794485629789506,
|
|
"learning_rate": 5.309917355371901e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.04,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1160
|
|
},
|
|
{
|
|
"epoch": 1.4245398773006135,
|
|
"grad_norm": 11.691918713848954,
|
|
"learning_rate": 5.305785123966941e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0653,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1161
|
|
},
|
|
{
|
|
"epoch": 1.4257668711656442,
|
|
"grad_norm": 7.94726211955386,
|
|
"learning_rate": 5.301652892561983e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0334,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.466796875,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1162
|
|
},
|
|
{
|
|
"epoch": 1.4269938650306748,
|
|
"grad_norm": 6.923425395788315,
|
|
"learning_rate": 5.297520661157024e-07,
|
|
"logits/chosen": -0.337890625,
|
|
"logits/rejected": -0.5859375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0243,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 1163
|
|
},
|
|
{
|
|
"epoch": 1.4282208588957055,
|
|
"grad_norm": 20.88597213226608,
|
|
"learning_rate": 5.293388429752065e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0767,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.271484375,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1164
|
|
},
|
|
{
|
|
"epoch": 1.4294478527607362,
|
|
"grad_norm": 9.162918526540627,
|
|
"learning_rate": 5.289256198347107e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0265,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.21484375,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1165
|
|
},
|
|
{
|
|
"epoch": 1.4306748466257668,
|
|
"grad_norm": 15.19106236090889,
|
|
"learning_rate": 5.285123966942148e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0432,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.166015625,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1166
|
|
},
|
|
{
|
|
"epoch": 1.4319018404907975,
|
|
"grad_norm": 5.9734292073664434,
|
|
"learning_rate": 5.28099173553719e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.019,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.38671875,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1167
|
|
},
|
|
{
|
|
"epoch": 1.4331288343558282,
|
|
"grad_norm": 18.69741109983899,
|
|
"learning_rate": 5.276859504132231e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0606,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.0205078125,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1168
|
|
},
|
|
{
|
|
"epoch": 1.4343558282208588,
|
|
"grad_norm": 11.177560877009283,
|
|
"learning_rate": 5.272727272727272e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0368,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.4296875,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1169
|
|
},
|
|
{
|
|
"epoch": 1.4355828220858895,
|
|
"grad_norm": 16.224444034226,
|
|
"learning_rate": 5.268595041322313e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0446,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.1083984375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1170
|
|
},
|
|
{
|
|
"epoch": 1.4368098159509202,
|
|
"grad_norm": 8.814427959339382,
|
|
"learning_rate": 5.264462809917355e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0253,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.83984375,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1171
|
|
},
|
|
{
|
|
"epoch": 1.438036809815951,
|
|
"grad_norm": 16.110987616417514,
|
|
"learning_rate": 5.260330578512396e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0639,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.0147705078125,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1172
|
|
},
|
|
{
|
|
"epoch": 1.4392638036809817,
|
|
"grad_norm": 11.295152778410886,
|
|
"learning_rate": 5.256198347107438e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0414,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.435546875,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1173
|
|
},
|
|
{
|
|
"epoch": 1.4404907975460124,
|
|
"grad_norm": 8.198100524653398,
|
|
"learning_rate": 5.252066115702479e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0229,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2021484375,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1174
|
|
},
|
|
{
|
|
"epoch": 1.441717791411043,
|
|
"grad_norm": 6.468793057035204,
|
|
"learning_rate": 5.247933884297521e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0328,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.007080078125,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1175
|
|
},
|
|
{
|
|
"epoch": 1.4429447852760737,
|
|
"grad_norm": 7.821621764342707,
|
|
"learning_rate": 5.243801652892562e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0356,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.0257568359375,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1176
|
|
},
|
|
{
|
|
"epoch": 1.4441717791411044,
|
|
"grad_norm": 8.99027020675616,
|
|
"learning_rate": 5.239669421487604e-07,
|
|
"logits/chosen": -0.123046875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0349,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.51953125,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1177
|
|
},
|
|
{
|
|
"epoch": 1.445398773006135,
|
|
"grad_norm": 9.760481269717861,
|
|
"learning_rate": 5.235537190082644e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0351,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1178
|
|
},
|
|
{
|
|
"epoch": 1.4466257668711657,
|
|
"grad_norm": 9.85395321742729,
|
|
"learning_rate": 5.231404958677686e-07,
|
|
"logits/chosen": -0.30078125,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0425,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1179
|
|
},
|
|
{
|
|
"epoch": 1.4478527607361964,
|
|
"grad_norm": 9.491024820250248,
|
|
"learning_rate": 5.227272727272727e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0389,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.474609375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1180
|
|
},
|
|
{
|
|
"epoch": 1.449079754601227,
|
|
"grad_norm": 8.628955288461011,
|
|
"learning_rate": 5.223140495867769e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.027,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -8.25,
|
|
"step": 1181
|
|
},
|
|
{
|
|
"epoch": 1.4503067484662577,
|
|
"grad_norm": 12.330343362663173,
|
|
"learning_rate": 5.21900826446281e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0423,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.2080078125,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1182
|
|
},
|
|
{
|
|
"epoch": 1.4515337423312884,
|
|
"grad_norm": 18.311745403244508,
|
|
"learning_rate": 5.214876033057851e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.066,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.30859375,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.3125,
|
|
"step": 1183
|
|
},
|
|
{
|
|
"epoch": 1.452760736196319,
|
|
"grad_norm": 9.930008847693435,
|
|
"learning_rate": 5.210743801652893e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.044,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.7734375,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1184
|
|
},
|
|
{
|
|
"epoch": 1.4539877300613497,
|
|
"grad_norm": 10.599628472788549,
|
|
"learning_rate": 5.206611570247934e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0381,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1185
|
|
},
|
|
{
|
|
"epoch": 1.4552147239263804,
|
|
"grad_norm": 9.24364465937099,
|
|
"learning_rate": 5.202479338842976e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0501,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1186
|
|
},
|
|
{
|
|
"epoch": 1.456441717791411,
|
|
"grad_norm": 8.564500000609362,
|
|
"learning_rate": 5.198347107438016e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0344,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -7.96875,
|
|
"step": 1187
|
|
},
|
|
{
|
|
"epoch": 1.4576687116564417,
|
|
"grad_norm": 10.020038829678471,
|
|
"learning_rate": 5.194214876033058e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0295,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1188
|
|
},
|
|
{
|
|
"epoch": 1.4588957055214724,
|
|
"grad_norm": 12.283792640414937,
|
|
"learning_rate": 5.190082644628098e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0626,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1189
|
|
},
|
|
{
|
|
"epoch": 1.460122699386503,
|
|
"grad_norm": 6.5665974869898145,
|
|
"learning_rate": 5.18595041322314e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0261,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1190
|
|
},
|
|
{
|
|
"epoch": 1.4613496932515337,
|
|
"grad_norm": 13.640277779908661,
|
|
"learning_rate": 5.181818181818181e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.220703125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0433,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1191
|
|
},
|
|
{
|
|
"epoch": 1.4625766871165644,
|
|
"grad_norm": 11.234457361782162,
|
|
"learning_rate": 5.177685950413223e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0419,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1192
|
|
},
|
|
{
|
|
"epoch": 1.463803680981595,
|
|
"grad_norm": 10.870360637609034,
|
|
"learning_rate": 5.173553719008264e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0267,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1193
|
|
},
|
|
{
|
|
"epoch": 1.4650306748466257,
|
|
"grad_norm": 6.441651379375814,
|
|
"learning_rate": 5.169421487603306e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0148,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1194
|
|
},
|
|
{
|
|
"epoch": 1.4662576687116564,
|
|
"grad_norm": 9.295817211422534,
|
|
"learning_rate": 5.165289256198346e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0392,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1195
|
|
},
|
|
{
|
|
"epoch": 1.467484662576687,
|
|
"grad_norm": 10.168242508246577,
|
|
"learning_rate": 5.161157024793388e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0385,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.359375,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1196
|
|
},
|
|
{
|
|
"epoch": 1.4687116564417177,
|
|
"grad_norm": 21.858565498383,
|
|
"learning_rate": 5.157024793388429e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0658,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.56640625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1197
|
|
},
|
|
{
|
|
"epoch": 1.4699386503067484,
|
|
"grad_norm": 14.67724383961791,
|
|
"learning_rate": 5.152892561983471e-07,
|
|
"logits/chosen": -0.08544921875,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0526,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1198
|
|
},
|
|
{
|
|
"epoch": 1.471165644171779,
|
|
"grad_norm": 17.65593615760072,
|
|
"learning_rate": 5.148760330578512e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0403,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1199
|
|
},
|
|
{
|
|
"epoch": 1.4723926380368098,
|
|
"grad_norm": 6.391043811028244,
|
|
"learning_rate": 5.144628099173553e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0221,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.388671875,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 1.4736196319018404,
|
|
"grad_norm": 5.833463699107399,
|
|
"learning_rate": 5.140495867768595e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0195,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.56640625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1201
|
|
},
|
|
{
|
|
"epoch": 1.474846625766871,
|
|
"grad_norm": 10.331582181573117,
|
|
"learning_rate": 5.136363636363636e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0325,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1202
|
|
},
|
|
{
|
|
"epoch": 1.4760736196319018,
|
|
"grad_norm": 11.36366878280935,
|
|
"learning_rate": 5.132231404958678e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.041,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1203
|
|
},
|
|
{
|
|
"epoch": 1.4773006134969324,
|
|
"grad_norm": 11.124998893181886,
|
|
"learning_rate": 5.128099173553718e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0386,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1204
|
|
},
|
|
{
|
|
"epoch": 1.478527607361963,
|
|
"grad_norm": 15.878680519356722,
|
|
"learning_rate": 5.12396694214876e-07,
|
|
"logits/chosen": -0.0673828125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.062,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1205
|
|
},
|
|
{
|
|
"epoch": 1.4797546012269938,
|
|
"grad_norm": 5.925858246827769,
|
|
"learning_rate": 5.119834710743801e-07,
|
|
"logits/chosen": -0.12060546875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0295,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1206
|
|
},
|
|
{
|
|
"epoch": 1.4809815950920244,
|
|
"grad_norm": 13.479825569113405,
|
|
"learning_rate": 5.115702479338843e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0424,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.296875,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1207
|
|
},
|
|
{
|
|
"epoch": 1.482208588957055,
|
|
"grad_norm": 13.480469485836236,
|
|
"learning_rate": 5.111570247933884e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0591,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.26953125,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1208
|
|
},
|
|
{
|
|
"epoch": 1.4834355828220858,
|
|
"grad_norm": 8.932445062757273,
|
|
"learning_rate": 5.107438016528926e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0442,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.455078125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1209
|
|
},
|
|
{
|
|
"epoch": 1.4846625766871164,
|
|
"grad_norm": 15.208159495097217,
|
|
"learning_rate": 5.103305785123967e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0309,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1210
|
|
},
|
|
{
|
|
"epoch": 1.4858895705521473,
|
|
"grad_norm": 4.0538490215592295,
|
|
"learning_rate": 5.099173553719009e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0149,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1211
|
|
},
|
|
{
|
|
"epoch": 1.487116564417178,
|
|
"grad_norm": 7.366657527338233,
|
|
"learning_rate": 5.09504132231405e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0336,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.27734375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1212
|
|
},
|
|
{
|
|
"epoch": 1.4883435582822087,
|
|
"grad_norm": 6.362487185558626,
|
|
"learning_rate": 5.090909090909091e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0276,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1213
|
|
},
|
|
{
|
|
"epoch": 1.4895705521472393,
|
|
"grad_norm": 12.994115633971985,
|
|
"learning_rate": 5.086776859504132e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0499,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.0216064453125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1214
|
|
},
|
|
{
|
|
"epoch": 1.49079754601227,
|
|
"grad_norm": 12.980563427542194,
|
|
"learning_rate": 5.082644628099174e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.4765625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0324,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1215
|
|
},
|
|
{
|
|
"epoch": 1.4920245398773007,
|
|
"grad_norm": 10.536318852600628,
|
|
"learning_rate": 5.078512396694214e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0379,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.04443359375,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1216
|
|
},
|
|
{
|
|
"epoch": 1.4932515337423313,
|
|
"grad_norm": 8.26021938537469,
|
|
"learning_rate": 5.074380165289256e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0268,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1217
|
|
},
|
|
{
|
|
"epoch": 1.494478527607362,
|
|
"grad_norm": 7.323360369517363,
|
|
"learning_rate": 5.070247933884297e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0259,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.1298828125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1218
|
|
},
|
|
{
|
|
"epoch": 1.4957055214723927,
|
|
"grad_norm": 18.24885381124581,
|
|
"learning_rate": 5.066115702479338e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0589,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.3359375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1219
|
|
},
|
|
{
|
|
"epoch": 1.4969325153374233,
|
|
"grad_norm": 15.42823881469555,
|
|
"learning_rate": 5.06198347107438e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0584,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.08642578125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1220
|
|
},
|
|
{
|
|
"epoch": 1.498159509202454,
|
|
"grad_norm": 11.450116625722911,
|
|
"learning_rate": 5.057851239669421e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0396,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.578125,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1221
|
|
},
|
|
{
|
|
"epoch": 1.4993865030674847,
|
|
"grad_norm": 11.04611444061931,
|
|
"learning_rate": 5.053719008264462e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0441,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.474609375,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1222
|
|
},
|
|
{
|
|
"epoch": 1.5006134969325153,
|
|
"grad_norm": 7.1619021316985245,
|
|
"learning_rate": 5.049586776859503e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0317,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1223
|
|
},
|
|
{
|
|
"epoch": 1.501840490797546,
|
|
"grad_norm": 9.585839726779575,
|
|
"learning_rate": 5.045454545454545e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0476,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.50390625,
|
|
"rewards/margins": 8.5625,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1224
|
|
},
|
|
{
|
|
"epoch": 1.5030674846625767,
|
|
"grad_norm": 5.127132425101296,
|
|
"learning_rate": 5.041322314049586e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0216,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2177734375,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1225
|
|
},
|
|
{
|
|
"epoch": 1.5042944785276073,
|
|
"grad_norm": 15.259210492730748,
|
|
"learning_rate": 5.037190082644628e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0396,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.228515625,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1226
|
|
},
|
|
{
|
|
"epoch": 1.505521472392638,
|
|
"grad_norm": 6.575684304084051,
|
|
"learning_rate": 5.033057851239669e-07,
|
|
"logits/chosen": -0.373046875,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0212,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1227
|
|
},
|
|
{
|
|
"epoch": 1.5067484662576687,
|
|
"grad_norm": 25.557879267365085,
|
|
"learning_rate": 5.028925619834711e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0996,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.26171875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1228
|
|
},
|
|
{
|
|
"epoch": 1.5079754601226993,
|
|
"grad_norm": 6.532362298257727,
|
|
"learning_rate": 5.024793388429752e-07,
|
|
"logits/chosen": -0.302734375,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0238,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.16796875,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1229
|
|
},
|
|
{
|
|
"epoch": 1.50920245398773,
|
|
"grad_norm": 9.494715021563923,
|
|
"learning_rate": 5.020661157024794e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.041,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1230
|
|
},
|
|
{
|
|
"epoch": 1.510429447852761,
|
|
"grad_norm": 9.0476570200281,
|
|
"learning_rate": 5.016528925619834e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.0297,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1231
|
|
},
|
|
{
|
|
"epoch": 1.5116564417177916,
|
|
"grad_norm": 8.203274654266307,
|
|
"learning_rate": 5.012396694214876e-07,
|
|
"logits/chosen": -0.35546875,
|
|
"logits/rejected": -0.55859375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0223,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1232
|
|
},
|
|
{
|
|
"epoch": 1.5128834355828222,
|
|
"grad_norm": 7.3468816241633075,
|
|
"learning_rate": 5.008264462809917e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0234,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1233
|
|
},
|
|
{
|
|
"epoch": 1.514110429447853,
|
|
"grad_norm": 22.21410237119622,
|
|
"learning_rate": 5.004132231404959e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0622,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1234
|
|
},
|
|
{
|
|
"epoch": 1.5153374233128836,
|
|
"grad_norm": 9.572826892921766,
|
|
"learning_rate": 5e-07,
|
|
"logits/chosen": -0.12109375,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0349,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.06640625,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1235
|
|
},
|
|
{
|
|
"epoch": 1.5165644171779142,
|
|
"grad_norm": 15.526612231629745,
|
|
"learning_rate": 4.995867768595041e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0456,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.25,
|
|
"step": 1236
|
|
},
|
|
{
|
|
"epoch": 1.517791411042945,
|
|
"grad_norm": 15.456982845798143,
|
|
"learning_rate": 4.991735537190083e-07,
|
|
"logits/chosen": -0.310546875,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0545,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1237
|
|
},
|
|
{
|
|
"epoch": 1.5190184049079756,
|
|
"grad_norm": 6.907460274932075,
|
|
"learning_rate": 4.987603305785124e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0224,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1238
|
|
},
|
|
{
|
|
"epoch": 1.5202453987730062,
|
|
"grad_norm": 14.894657531734518,
|
|
"learning_rate": 4.983471074380165e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.4765625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0561,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.87890625,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1239
|
|
},
|
|
{
|
|
"epoch": 1.521472392638037,
|
|
"grad_norm": 9.38095700557469,
|
|
"learning_rate": 4.979338842975207e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0413,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.7109375,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1240
|
|
},
|
|
{
|
|
"epoch": 1.5226993865030676,
|
|
"grad_norm": 12.191743774010911,
|
|
"learning_rate": 4.975206611570248e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0552,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1241
|
|
},
|
|
{
|
|
"epoch": 1.5239263803680982,
|
|
"grad_norm": 19.69037826903355,
|
|
"learning_rate": 4.971074380165288e-07,
|
|
"logits/chosen": -0.322265625,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0494,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.37109375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1242
|
|
},
|
|
{
|
|
"epoch": 1.525153374233129,
|
|
"grad_norm": 9.945102116470355,
|
|
"learning_rate": 4.96694214876033e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0326,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.1494140625,
|
|
"rewards/margins": 8.75,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1243
|
|
},
|
|
{
|
|
"epoch": 1.5263803680981596,
|
|
"grad_norm": 5.733787444240008,
|
|
"learning_rate": 4.962809917355371e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.24609375,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0212,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.263671875,
|
|
"rewards/margins": 8.6875,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 1244
|
|
},
|
|
{
|
|
"epoch": 1.5276073619631902,
|
|
"grad_norm": 12.673613847436323,
|
|
"learning_rate": 4.958677685950413e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0558,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.7734375,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1245
|
|
},
|
|
{
|
|
"epoch": 1.528834355828221,
|
|
"grad_norm": 4.522415021420316,
|
|
"learning_rate": 4.954545454545454e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.024,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1246
|
|
},
|
|
{
|
|
"epoch": 1.5300613496932516,
|
|
"grad_norm": 3.7381046178457242,
|
|
"learning_rate": 4.950413223140495e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0158,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1247
|
|
},
|
|
{
|
|
"epoch": 1.5312883435582823,
|
|
"grad_norm": 13.945477935019138,
|
|
"learning_rate": 4.946280991735537e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1248
|
|
},
|
|
{
|
|
"epoch": 1.532515337423313,
|
|
"grad_norm": 16.432552097040393,
|
|
"learning_rate": 4.942148760330578e-07,
|
|
"logits/chosen": -0.031494140625,
|
|
"logits/rejected": -0.1376953125,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0685,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.052734375,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1249
|
|
},
|
|
{
|
|
"epoch": 1.5337423312883436,
|
|
"grad_norm": 8.401137914561525,
|
|
"learning_rate": 4.93801652892562e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -446.0,
|
|
"loss": 0.0277,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1250
|
|
},
|
|
{
|
|
"epoch": 1.5349693251533743,
|
|
"grad_norm": 17.707461229750546,
|
|
"learning_rate": 4.933884297520661e-07,
|
|
"logits/chosen": -0.34765625,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0528,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1251
|
|
},
|
|
{
|
|
"epoch": 1.536196319018405,
|
|
"grad_norm": 13.506598380063398,
|
|
"learning_rate": 4.929752066115702e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0413,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1252
|
|
},
|
|
{
|
|
"epoch": 1.5374233128834356,
|
|
"grad_norm": 14.504145082644028,
|
|
"learning_rate": 4.925619834710743e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0452,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1253
|
|
},
|
|
{
|
|
"epoch": 1.5386503067484663,
|
|
"grad_norm": 13.492438510341476,
|
|
"learning_rate": 4.921487603305785e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0503,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.66796875,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1254
|
|
},
|
|
{
|
|
"epoch": 1.539877300613497,
|
|
"grad_norm": 8.820977295546955,
|
|
"learning_rate": 4.917355371900826e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0296,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1255
|
|
},
|
|
{
|
|
"epoch": 1.5411042944785276,
|
|
"grad_norm": 10.020492300796558,
|
|
"learning_rate": 4.913223140495867e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0361,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1256
|
|
},
|
|
{
|
|
"epoch": 1.5423312883435583,
|
|
"grad_norm": 13.431484369727785,
|
|
"learning_rate": 4.909090909090909e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0469,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.470703125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1257
|
|
},
|
|
{
|
|
"epoch": 1.543558282208589,
|
|
"grad_norm": 8.035190397652592,
|
|
"learning_rate": 4.90495867768595e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0297,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1258
|
|
},
|
|
{
|
|
"epoch": 1.5447852760736196,
|
|
"grad_norm": 7.248588905954608,
|
|
"learning_rate": 4.900826446280992e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0256,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1259
|
|
},
|
|
{
|
|
"epoch": 1.5460122699386503,
|
|
"grad_norm": 15.937446922392892,
|
|
"learning_rate": 4.896694214876033e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0368,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1260
|
|
},
|
|
{
|
|
"epoch": 1.547239263803681,
|
|
"grad_norm": 7.599195460785998,
|
|
"learning_rate": 4.892561983471074e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0242,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1261
|
|
},
|
|
{
|
|
"epoch": 1.5484662576687116,
|
|
"grad_norm": 37.85052519895869,
|
|
"learning_rate": 4.888429752066116e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0639,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.322265625,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1262
|
|
},
|
|
{
|
|
"epoch": 1.5496932515337423,
|
|
"grad_norm": 14.14107221218377,
|
|
"learning_rate": 4.884297520661157e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0419,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1263
|
|
},
|
|
{
|
|
"epoch": 1.550920245398773,
|
|
"grad_norm": 14.069819821126352,
|
|
"learning_rate": 4.880165289256198e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0664,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.38671875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1264
|
|
},
|
|
{
|
|
"epoch": 1.5521472392638036,
|
|
"grad_norm": 16.843456872134922,
|
|
"learning_rate": 4.87603305785124e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0411,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1265
|
|
},
|
|
{
|
|
"epoch": 1.5533742331288343,
|
|
"grad_norm": 10.667014161685005,
|
|
"learning_rate": 4.871900826446281e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0361,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.185546875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1266
|
|
},
|
|
{
|
|
"epoch": 1.554601226993865,
|
|
"grad_norm": 16.452210048241056,
|
|
"learning_rate": 4.867768595041323e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0943,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1267
|
|
},
|
|
{
|
|
"epoch": 1.5558282208588956,
|
|
"grad_norm": 10.954318677494932,
|
|
"learning_rate": 4.863636363636364e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -496.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0296,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1268
|
|
},
|
|
{
|
|
"epoch": 1.5570552147239263,
|
|
"grad_norm": 13.826760593553331,
|
|
"learning_rate": 4.859504132231405e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0464,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1269
|
|
},
|
|
{
|
|
"epoch": 1.558282208588957,
|
|
"grad_norm": 12.172332899130659,
|
|
"learning_rate": 4.855371900826447e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0331,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1270
|
|
},
|
|
{
|
|
"epoch": 1.5595092024539876,
|
|
"grad_norm": 8.436571867819609,
|
|
"learning_rate": 4.851239669421487e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0296,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1271
|
|
},
|
|
{
|
|
"epoch": 1.5607361963190183,
|
|
"grad_norm": 5.472725546442262,
|
|
"learning_rate": 4.847107438016528e-07,
|
|
"logits/chosen": -0.302734375,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0185,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1272
|
|
},
|
|
{
|
|
"epoch": 1.561963190184049,
|
|
"grad_norm": 17.9929195637627,
|
|
"learning_rate": 4.842975206611569e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0839,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.54296875,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1273
|
|
},
|
|
{
|
|
"epoch": 1.5631901840490796,
|
|
"grad_norm": 7.914532444843044,
|
|
"learning_rate": 4.838842975206611e-07,
|
|
"logits/chosen": -0.361328125,
|
|
"logits/rejected": -0.55859375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.021,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1274
|
|
},
|
|
{
|
|
"epoch": 1.5644171779141103,
|
|
"grad_norm": 13.680874559343534,
|
|
"learning_rate": 4.834710743801652e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.063,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.25,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1275
|
|
},
|
|
{
|
|
"epoch": 1.565644171779141,
|
|
"grad_norm": 8.520481992182615,
|
|
"learning_rate": 4.830578512396694e-07,
|
|
"logits/chosen": -0.0830078125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0239,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1276
|
|
},
|
|
{
|
|
"epoch": 1.5668711656441716,
|
|
"grad_norm": 20.687758972058443,
|
|
"learning_rate": 4.826446280991735e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0732,
|
|
"rewards/accuracies": 0.9453125,
|
|
"rewards/chosen": 0.37890625,
|
|
"rewards/margins": 9.0,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1277
|
|
},
|
|
{
|
|
"epoch": 1.5680981595092025,
|
|
"grad_norm": 6.214727457063051,
|
|
"learning_rate": 4.822314049586776e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0211,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1278
|
|
},
|
|
{
|
|
"epoch": 1.5693251533742332,
|
|
"grad_norm": 15.82891980095153,
|
|
"learning_rate": 4.818181818181818e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0612,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1279
|
|
},
|
|
{
|
|
"epoch": 1.5705521472392638,
|
|
"grad_norm": 15.084041080731549,
|
|
"learning_rate": 4.814049586776859e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0557,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1280
|
|
},
|
|
{
|
|
"epoch": 1.5717791411042945,
|
|
"grad_norm": 11.611649573437312,
|
|
"learning_rate": 4.809917355371901e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.0456,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1281
|
|
},
|
|
{
|
|
"epoch": 1.5730061349693252,
|
|
"grad_norm": 9.047220041907309,
|
|
"learning_rate": 4.805785123966942e-07,
|
|
"logits/chosen": -0.0111083984375,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0318,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1282
|
|
},
|
|
{
|
|
"epoch": 1.5742331288343558,
|
|
"grad_norm": 12.723617187572223,
|
|
"learning_rate": 4.801652892561983e-07,
|
|
"logits/chosen": -0.06494140625,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0504,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.0478515625,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1283
|
|
},
|
|
{
|
|
"epoch": 1.5754601226993865,
|
|
"grad_norm": 16.789305184554657,
|
|
"learning_rate": 4.797520661157025e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.072,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1284
|
|
},
|
|
{
|
|
"epoch": 1.5766871165644172,
|
|
"grad_norm": 17.232596988140553,
|
|
"learning_rate": 4.793388429752066e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0493,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1285
|
|
},
|
|
{
|
|
"epoch": 1.5779141104294478,
|
|
"grad_norm": 10.096699395706557,
|
|
"learning_rate": 4.789256198347108e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0377,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.09912109375,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1286
|
|
},
|
|
{
|
|
"epoch": 1.5791411042944785,
|
|
"grad_norm": 14.336573693251026,
|
|
"learning_rate": 4.785123966942149e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -320.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0517,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.42578125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1287
|
|
},
|
|
{
|
|
"epoch": 1.5803680981595092,
|
|
"grad_norm": 12.880034581469118,
|
|
"learning_rate": 4.78099173553719e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0504,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.59375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1288
|
|
},
|
|
{
|
|
"epoch": 1.5815950920245399,
|
|
"grad_norm": 8.245102498870926,
|
|
"learning_rate": 4.776859504132231e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.0418,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.25390625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1289
|
|
},
|
|
{
|
|
"epoch": 1.5828220858895705,
|
|
"grad_norm": 7.068021899768858,
|
|
"learning_rate": 4.772727272727273e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0189,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.0024261474609375,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1290
|
|
},
|
|
{
|
|
"epoch": 1.5840490797546012,
|
|
"grad_norm": 9.743383518312836,
|
|
"learning_rate": 4.768595041322314e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0292,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.01275634765625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1291
|
|
},
|
|
{
|
|
"epoch": 1.5852760736196319,
|
|
"grad_norm": 6.209713888247011,
|
|
"learning_rate": 4.764462809917355e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0249,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.125,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1292
|
|
},
|
|
{
|
|
"epoch": 1.5865030674846625,
|
|
"grad_norm": 14.80409692246538,
|
|
"learning_rate": 4.7603305785123966e-07,
|
|
"logits/chosen": -0.08740234375,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0269,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.23828125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1293
|
|
},
|
|
{
|
|
"epoch": 1.5877300613496934,
|
|
"grad_norm": 5.456155704440144,
|
|
"learning_rate": 4.756198347107438e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0198,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.240234375,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1294
|
|
},
|
|
{
|
|
"epoch": 1.588957055214724,
|
|
"grad_norm": 8.72340128371093,
|
|
"learning_rate": 4.7520661157024796e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0447,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.25390625,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1295
|
|
},
|
|
{
|
|
"epoch": 1.5901840490797547,
|
|
"grad_norm": 19.421110765065105,
|
|
"learning_rate": 4.7479338842975205e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.1032,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": -0.7265625,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1296
|
|
},
|
|
{
|
|
"epoch": 1.5914110429447854,
|
|
"grad_norm": 13.177535617402823,
|
|
"learning_rate": 4.743801652892562e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0482,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.11767578125,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1297
|
|
},
|
|
{
|
|
"epoch": 1.592638036809816,
|
|
"grad_norm": 13.467051816946709,
|
|
"learning_rate": 4.739669421487603e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0554,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1298
|
|
},
|
|
{
|
|
"epoch": 1.5938650306748468,
|
|
"grad_norm": 13.633399146997593,
|
|
"learning_rate": 4.7355371900826444e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0448,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.1689453125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1299
|
|
},
|
|
{
|
|
"epoch": 1.5950920245398774,
|
|
"grad_norm": 10.239023412942723,
|
|
"learning_rate": 4.7314049586776853e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0321,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.54296875,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.596319018404908,
|
|
"grad_norm": 8.41274981521433,
|
|
"learning_rate": 4.727272727272727e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0274,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.126953125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1301
|
|
},
|
|
{
|
|
"epoch": 1.5975460122699388,
|
|
"grad_norm": 9.317632271393505,
|
|
"learning_rate": 4.7231404958677683e-07,
|
|
"logits/chosen": -0.333984375,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0439,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.78515625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1302
|
|
},
|
|
{
|
|
"epoch": 1.5987730061349694,
|
|
"grad_norm": 13.137223593433058,
|
|
"learning_rate": 4.71900826446281e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0384,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1303
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"grad_norm": 6.281757480148536,
|
|
"learning_rate": 4.7148760330578507e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0227,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1304
|
|
},
|
|
{
|
|
"epoch": 1.6012269938650308,
|
|
"grad_norm": 7.3065474985773,
|
|
"learning_rate": 4.710743801652892e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0379,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1305
|
|
},
|
|
{
|
|
"epoch": 1.6024539877300614,
|
|
"grad_norm": 8.228030393276795,
|
|
"learning_rate": 4.7066115702479336e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.69921875,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1306
|
|
},
|
|
{
|
|
"epoch": 1.603680981595092,
|
|
"grad_norm": 8.110612340051029,
|
|
"learning_rate": 4.702479338842975e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0225,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1307
|
|
},
|
|
{
|
|
"epoch": 1.6049079754601228,
|
|
"grad_norm": 6.18502649696746,
|
|
"learning_rate": 4.6983471074380166e-07,
|
|
"logits/chosen": -0.302734375,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0208,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.9140625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1308
|
|
},
|
|
{
|
|
"epoch": 1.6061349693251534,
|
|
"grad_norm": 7.213253884175224,
|
|
"learning_rate": 4.6942148760330575e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0282,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.50390625,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1309
|
|
},
|
|
{
|
|
"epoch": 1.607361963190184,
|
|
"grad_norm": 9.092736992318676,
|
|
"learning_rate": 4.690082644628099e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0338,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1310
|
|
},
|
|
{
|
|
"epoch": 1.6085889570552148,
|
|
"grad_norm": 8.974238019147444,
|
|
"learning_rate": 4.6859504132231405e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0474,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5,
|
|
"rewards/margins": 8.625,
|
|
"rewards/rejected": -8.125,
|
|
"step": 1311
|
|
},
|
|
{
|
|
"epoch": 1.6098159509202454,
|
|
"grad_norm": 7.743650713864273,
|
|
"learning_rate": 4.681818181818182e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.48828125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0239,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1312
|
|
},
|
|
{
|
|
"epoch": 1.611042944785276,
|
|
"grad_norm": 6.653895084192328,
|
|
"learning_rate": 4.677685950413223e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0301,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6484375,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1313
|
|
},
|
|
{
|
|
"epoch": 1.6122699386503068,
|
|
"grad_norm": 5.718149967619655,
|
|
"learning_rate": 4.6735537190082644e-07,
|
|
"logits/chosen": -0.3515625,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0163,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1314
|
|
},
|
|
{
|
|
"epoch": 1.6134969325153374,
|
|
"grad_norm": 9.683315531046246,
|
|
"learning_rate": 4.669421487603306e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0366,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1315
|
|
},
|
|
{
|
|
"epoch": 1.614723926380368,
|
|
"grad_norm": 10.858818646193528,
|
|
"learning_rate": 4.665289256198347e-07,
|
|
"logits/chosen": -0.06884765625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0366,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.400390625,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1316
|
|
},
|
|
{
|
|
"epoch": 1.6159509202453988,
|
|
"grad_norm": 7.172302137792857,
|
|
"learning_rate": 4.661157024793388e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0284,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.431640625,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1317
|
|
},
|
|
{
|
|
"epoch": 1.6171779141104294,
|
|
"grad_norm": 9.41448963392097,
|
|
"learning_rate": 4.657024793388429e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0216,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.271484375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1318
|
|
},
|
|
{
|
|
"epoch": 1.61840490797546,
|
|
"grad_norm": 9.252537190793298,
|
|
"learning_rate": 4.6528925619834707e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.022,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1319
|
|
},
|
|
{
|
|
"epoch": 1.6196319018404908,
|
|
"grad_norm": 17.13507082983842,
|
|
"learning_rate": 4.648760330578512e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0512,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1320
|
|
},
|
|
{
|
|
"epoch": 1.6208588957055214,
|
|
"grad_norm": 19.682226431049997,
|
|
"learning_rate": 4.6446280991735536e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0486,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1321
|
|
},
|
|
{
|
|
"epoch": 1.622085889570552,
|
|
"grad_norm": 16.418057590683848,
|
|
"learning_rate": 4.6404958677685945e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0403,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.197265625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1322
|
|
},
|
|
{
|
|
"epoch": 1.6233128834355828,
|
|
"grad_norm": 4.021950027608913,
|
|
"learning_rate": 4.636363636363636e-07,
|
|
"logits/chosen": -0.058837890625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0165,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.50390625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1323
|
|
},
|
|
{
|
|
"epoch": 1.6245398773006134,
|
|
"grad_norm": 12.437847508441237,
|
|
"learning_rate": 4.6322314049586775e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0547,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1324
|
|
},
|
|
{
|
|
"epoch": 1.6257668711656441,
|
|
"grad_norm": 10.730686903178869,
|
|
"learning_rate": 4.628099173553719e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0442,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.1064453125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1325
|
|
},
|
|
{
|
|
"epoch": 1.6269938650306748,
|
|
"grad_norm": 9.534136764757534,
|
|
"learning_rate": 4.6239669421487604e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0273,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.435546875,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 1326
|
|
},
|
|
{
|
|
"epoch": 1.6282208588957054,
|
|
"grad_norm": 20.348608830967564,
|
|
"learning_rate": 4.6198347107438014e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0771,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.26953125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1327
|
|
},
|
|
{
|
|
"epoch": 1.6294478527607361,
|
|
"grad_norm": 16.467726513797313,
|
|
"learning_rate": 4.615702479338843e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0509,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.2265625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1328
|
|
},
|
|
{
|
|
"epoch": 1.6306748466257668,
|
|
"grad_norm": 14.50216473422956,
|
|
"learning_rate": 4.6115702479338843e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0508,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.042724609375,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1329
|
|
},
|
|
{
|
|
"epoch": 1.6319018404907975,
|
|
"grad_norm": 17.609332819885694,
|
|
"learning_rate": 4.607438016528926e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0593,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.01068115234375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1330
|
|
},
|
|
{
|
|
"epoch": 1.6331288343558281,
|
|
"grad_norm": 9.648923642616328,
|
|
"learning_rate": 4.6033057851239667e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0324,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.50390625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -11.0,
|
|
"step": 1331
|
|
},
|
|
{
|
|
"epoch": 1.6343558282208588,
|
|
"grad_norm": 11.932983005476947,
|
|
"learning_rate": 4.599173553719008e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0277,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.083984375,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1332
|
|
},
|
|
{
|
|
"epoch": 1.6355828220858895,
|
|
"grad_norm": 14.449256139382465,
|
|
"learning_rate": 4.595041322314049e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -422.0,
|
|
"loss": 0.0479,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.02294921875,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1333
|
|
},
|
|
{
|
|
"epoch": 1.6368098159509201,
|
|
"grad_norm": 8.041299488095149,
|
|
"learning_rate": 4.5909090909090906e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0278,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.83984375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1334
|
|
},
|
|
{
|
|
"epoch": 1.6380368098159508,
|
|
"grad_norm": 21.666093835744284,
|
|
"learning_rate": 4.586776859504132e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0554,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1335
|
|
},
|
|
{
|
|
"epoch": 1.6392638036809815,
|
|
"grad_norm": 12.443768139743762,
|
|
"learning_rate": 4.582644628099173e-07,
|
|
"logits/chosen": -0.400390625,
|
|
"logits/rejected": -0.5703125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0438,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.35546875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1336
|
|
},
|
|
{
|
|
"epoch": 1.6404907975460121,
|
|
"grad_norm": 6.130885243369341,
|
|
"learning_rate": 4.5785123966942145e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0168,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1337
|
|
},
|
|
{
|
|
"epoch": 1.6417177914110428,
|
|
"grad_norm": 7.893252550575079,
|
|
"learning_rate": 4.574380165289256e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0289,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1338
|
|
},
|
|
{
|
|
"epoch": 1.6429447852760735,
|
|
"grad_norm": 14.67821861843715,
|
|
"learning_rate": 4.5702479338842974e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.0462,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1339
|
|
},
|
|
{
|
|
"epoch": 1.6441717791411041,
|
|
"grad_norm": 16.25414995251407,
|
|
"learning_rate": 4.5661157024793384e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0781,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.72265625,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1340
|
|
},
|
|
{
|
|
"epoch": 1.645398773006135,
|
|
"grad_norm": 5.665767495720483,
|
|
"learning_rate": 4.56198347107438e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0193,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1341
|
|
},
|
|
{
|
|
"epoch": 1.6466257668711657,
|
|
"grad_norm": 6.7505668149236815,
|
|
"learning_rate": 4.5578512396694213e-07,
|
|
"logits/chosen": -0.296875,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0356,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1342
|
|
},
|
|
{
|
|
"epoch": 1.6478527607361964,
|
|
"grad_norm": 29.828942096440358,
|
|
"learning_rate": 4.553719008264463e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0802,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.2138671875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1343
|
|
},
|
|
{
|
|
"epoch": 1.649079754601227,
|
|
"grad_norm": 4.919641836025328,
|
|
"learning_rate": 4.5495867768595037e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0135,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1344
|
|
},
|
|
{
|
|
"epoch": 1.6503067484662577,
|
|
"grad_norm": 6.909150892571506,
|
|
"learning_rate": 4.545454545454545e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0264,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1345
|
|
},
|
|
{
|
|
"epoch": 1.6515337423312884,
|
|
"grad_norm": 19.450004740007984,
|
|
"learning_rate": 4.5413223140495867e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.054,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1346
|
|
},
|
|
{
|
|
"epoch": 1.652760736196319,
|
|
"grad_norm": 22.592366646625774,
|
|
"learning_rate": 4.537190082644628e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.1187,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1347
|
|
},
|
|
{
|
|
"epoch": 1.6539877300613497,
|
|
"grad_norm": 9.508485427813358,
|
|
"learning_rate": 4.5330578512396696e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0426,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.2578125,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1348
|
|
},
|
|
{
|
|
"epoch": 1.6552147239263804,
|
|
"grad_norm": 7.907526912278326,
|
|
"learning_rate": 4.5289256198347106e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0334,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1349
|
|
},
|
|
{
|
|
"epoch": 1.656441717791411,
|
|
"grad_norm": 11.995630429638908,
|
|
"learning_rate": 4.524793388429752e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0366,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1350
|
|
},
|
|
{
|
|
"epoch": 1.6576687116564417,
|
|
"grad_norm": 20.03302009325522,
|
|
"learning_rate": 4.520661157024793e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0959,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.48828125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1351
|
|
},
|
|
{
|
|
"epoch": 1.6588957055214724,
|
|
"grad_norm": 7.588890304865535,
|
|
"learning_rate": 4.5165289256198344e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0262,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1352
|
|
},
|
|
{
|
|
"epoch": 1.660122699386503,
|
|
"grad_norm": 17.636418772329574,
|
|
"learning_rate": 4.5123966942148754e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.2158203125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0469,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.00933837890625,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1353
|
|
},
|
|
{
|
|
"epoch": 1.6613496932515337,
|
|
"grad_norm": 5.770439441479,
|
|
"learning_rate": 4.508264462809917e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0281,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.318359375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1354
|
|
},
|
|
{
|
|
"epoch": 1.6625766871165644,
|
|
"grad_norm": 14.51368947816898,
|
|
"learning_rate": 4.5041322314049583e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0538,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.043212890625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1355
|
|
},
|
|
{
|
|
"epoch": 1.6638036809815953,
|
|
"grad_norm": 11.607022681375701,
|
|
"learning_rate": 4.5e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0419,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.330078125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1356
|
|
},
|
|
{
|
|
"epoch": 1.665030674846626,
|
|
"grad_norm": 6.201925304480739,
|
|
"learning_rate": 4.4958677685950413e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0151,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2734375,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1357
|
|
},
|
|
{
|
|
"epoch": 1.6662576687116566,
|
|
"grad_norm": 22.69395879001461,
|
|
"learning_rate": 4.491735537190082e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.5,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0691,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.0242919921875,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1358
|
|
},
|
|
{
|
|
"epoch": 1.6674846625766873,
|
|
"grad_norm": 13.473649796271712,
|
|
"learning_rate": 4.4876033057851237e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.032,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1359
|
|
},
|
|
{
|
|
"epoch": 1.668711656441718,
|
|
"grad_norm": 15.956836423582544,
|
|
"learning_rate": 4.483471074380165e-07,
|
|
"logits/chosen": -0.435546875,
|
|
"logits/rejected": -0.6640625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0543,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.546875,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1360
|
|
},
|
|
{
|
|
"epoch": 1.6699386503067486,
|
|
"grad_norm": 10.52492902643604,
|
|
"learning_rate": 4.4793388429752066e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0367,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.2265625,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1361
|
|
},
|
|
{
|
|
"epoch": 1.6711656441717793,
|
|
"grad_norm": 12.553784249961321,
|
|
"learning_rate": 4.4752066115702476e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0429,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.455078125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1362
|
|
},
|
|
{
|
|
"epoch": 1.67239263803681,
|
|
"grad_norm": 4.1059888676009235,
|
|
"learning_rate": 4.471074380165289e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0136,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.37109375,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1363
|
|
},
|
|
{
|
|
"epoch": 1.6736196319018406,
|
|
"grad_norm": 6.426592860391038,
|
|
"learning_rate": 4.4669421487603305e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0177,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1364
|
|
},
|
|
{
|
|
"epoch": 1.6748466257668713,
|
|
"grad_norm": 14.59538416768274,
|
|
"learning_rate": 4.462809917355372e-07,
|
|
"logits/chosen": -0.09130859375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0686,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1365
|
|
},
|
|
{
|
|
"epoch": 1.676073619631902,
|
|
"grad_norm": 9.111205296618245,
|
|
"learning_rate": 4.4586776859504135e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0284,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1366
|
|
},
|
|
{
|
|
"epoch": 1.6773006134969326,
|
|
"grad_norm": 11.707668054119575,
|
|
"learning_rate": 4.4545454545454544e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0541,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1367
|
|
},
|
|
{
|
|
"epoch": 1.6785276073619633,
|
|
"grad_norm": 15.80964855399164,
|
|
"learning_rate": 4.450413223140496e-07,
|
|
"logits/chosen": -0.35546875,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0463,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.2890625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1368
|
|
},
|
|
{
|
|
"epoch": 1.679754601226994,
|
|
"grad_norm": 15.404669004215867,
|
|
"learning_rate": 4.446280991735537e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0727,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.046875,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1369
|
|
},
|
|
{
|
|
"epoch": 1.6809815950920246,
|
|
"grad_norm": 11.767198454193279,
|
|
"learning_rate": 4.4421487603305783e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0307,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1370
|
|
},
|
|
{
|
|
"epoch": 1.6822085889570553,
|
|
"grad_norm": 15.203848393950707,
|
|
"learning_rate": 4.438016528925619e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -462.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0446,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1371
|
|
},
|
|
{
|
|
"epoch": 1.683435582822086,
|
|
"grad_norm": 7.699957082384994,
|
|
"learning_rate": 4.4338842975206607e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0303,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1372
|
|
},
|
|
{
|
|
"epoch": 1.6846625766871166,
|
|
"grad_norm": 14.759671942315824,
|
|
"learning_rate": 4.429752066115702e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0544,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1373
|
|
},
|
|
{
|
|
"epoch": 1.6858895705521473,
|
|
"grad_norm": 13.766883929453416,
|
|
"learning_rate": 4.4256198347107436e-07,
|
|
"logits/chosen": -0.09228515625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0567,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.4375,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1374
|
|
},
|
|
{
|
|
"epoch": 1.687116564417178,
|
|
"grad_norm": 13.485718445028985,
|
|
"learning_rate": 4.421487603305785e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0588,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.4453125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1375
|
|
},
|
|
{
|
|
"epoch": 1.6883435582822086,
|
|
"grad_norm": 6.195212386545783,
|
|
"learning_rate": 4.417355371900826e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0235,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1376
|
|
},
|
|
{
|
|
"epoch": 1.6895705521472393,
|
|
"grad_norm": 11.43635424727631,
|
|
"learning_rate": 4.4132231404958675e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0333,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1377
|
|
},
|
|
{
|
|
"epoch": 1.69079754601227,
|
|
"grad_norm": 10.39713135155569,
|
|
"learning_rate": 4.409090909090909e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0319,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.1572265625,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1378
|
|
},
|
|
{
|
|
"epoch": 1.6920245398773006,
|
|
"grad_norm": 8.367380437077856,
|
|
"learning_rate": 4.4049586776859505e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0413,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1379
|
|
},
|
|
{
|
|
"epoch": 1.6932515337423313,
|
|
"grad_norm": 6.305067427064263,
|
|
"learning_rate": 4.4008264462809914e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.022,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.75,
|
|
"step": 1380
|
|
},
|
|
{
|
|
"epoch": 1.694478527607362,
|
|
"grad_norm": 12.587916291736752,
|
|
"learning_rate": 4.396694214876033e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.1416015625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0527,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.4375,
|
|
"step": 1381
|
|
},
|
|
{
|
|
"epoch": 1.6957055214723926,
|
|
"grad_norm": 16.965413619004877,
|
|
"learning_rate": 4.3925619834710743e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0457,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.28515625,
|
|
"rewards/margins": 9.1875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1382
|
|
},
|
|
{
|
|
"epoch": 1.6969325153374233,
|
|
"grad_norm": 7.563921574034532,
|
|
"learning_rate": 4.388429752066116e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0216,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1383
|
|
},
|
|
{
|
|
"epoch": 1.698159509202454,
|
|
"grad_norm": 12.216087776822986,
|
|
"learning_rate": 4.3842975206611573e-07,
|
|
"logits/chosen": -0.06640625,
|
|
"logits/rejected": -0.1806640625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0425,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.44921875,
|
|
"rewards/margins": 8.875,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1384
|
|
},
|
|
{
|
|
"epoch": 1.6993865030674846,
|
|
"grad_norm": 11.215413816204737,
|
|
"learning_rate": 4.380165289256198e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.0529,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.77734375,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.5,
|
|
"step": 1385
|
|
},
|
|
{
|
|
"epoch": 1.7006134969325153,
|
|
"grad_norm": 10.008620457696347,
|
|
"learning_rate": 4.3760330578512397e-07,
|
|
"logits/chosen": -0.33984375,
|
|
"logits/rejected": -0.5546875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0442,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1386
|
|
},
|
|
{
|
|
"epoch": 1.701840490797546,
|
|
"grad_norm": 7.455196987203626,
|
|
"learning_rate": 4.3719008264462806e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -426.0,
|
|
"loss": 0.0248,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1387
|
|
},
|
|
{
|
|
"epoch": 1.7030674846625766,
|
|
"grad_norm": 8.662478879591731,
|
|
"learning_rate": 4.367768595041322e-07,
|
|
"logits/chosen": -0.0830078125,
|
|
"logits/rejected": -0.21875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.031,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.48046875,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -8.375,
|
|
"step": 1388
|
|
},
|
|
{
|
|
"epoch": 1.7042944785276073,
|
|
"grad_norm": 7.6361137281028615,
|
|
"learning_rate": 4.363636363636363e-07,
|
|
"logits/chosen": -0.345703125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0317,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1389
|
|
},
|
|
{
|
|
"epoch": 1.705521472392638,
|
|
"grad_norm": 18.332957888708503,
|
|
"learning_rate": 4.3595041322314045e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0492,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.6796875,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1390
|
|
},
|
|
{
|
|
"epoch": 1.7067484662576686,
|
|
"grad_norm": 7.997617467329183,
|
|
"learning_rate": 4.355371900826446e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0348,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -8.625,
|
|
"step": 1391
|
|
},
|
|
{
|
|
"epoch": 1.7079754601226993,
|
|
"grad_norm": 8.389199197881604,
|
|
"learning_rate": 4.3512396694214875e-07,
|
|
"logits/chosen": -0.0654296875,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0357,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.27734375,
|
|
"rewards/margins": 8.4375,
|
|
"rewards/rejected": -8.125,
|
|
"step": 1392
|
|
},
|
|
{
|
|
"epoch": 1.70920245398773,
|
|
"grad_norm": 11.292267274642116,
|
|
"learning_rate": 4.3471074380165284e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.04,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -8.125,
|
|
"step": 1393
|
|
},
|
|
{
|
|
"epoch": 1.7104294478527606,
|
|
"grad_norm": 14.8096908295346,
|
|
"learning_rate": 4.34297520661157e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0714,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.70703125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1394
|
|
},
|
|
{
|
|
"epoch": 1.7116564417177913,
|
|
"grad_norm": 4.560186843363393,
|
|
"learning_rate": 4.3388429752066114e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0181,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -8.5625,
|
|
"step": 1395
|
|
},
|
|
{
|
|
"epoch": 1.712883435582822,
|
|
"grad_norm": 3.555779696153153,
|
|
"learning_rate": 4.334710743801653e-07,
|
|
"logits/chosen": -0.058349609375,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0154,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1396
|
|
},
|
|
{
|
|
"epoch": 1.7141104294478526,
|
|
"grad_norm": 6.320037630953501,
|
|
"learning_rate": 4.3305785123966943e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0227,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1397
|
|
},
|
|
{
|
|
"epoch": 1.7153374233128833,
|
|
"grad_norm": 12.587020497649968,
|
|
"learning_rate": 4.326446280991735e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0381,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1398
|
|
},
|
|
{
|
|
"epoch": 1.716564417177914,
|
|
"grad_norm": 8.21734381269449,
|
|
"learning_rate": 4.3223140495867767e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0304,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.439453125,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1399
|
|
},
|
|
{
|
|
"epoch": 1.7177914110429446,
|
|
"grad_norm": 16.708540461160506,
|
|
"learning_rate": 4.318181818181818e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.061767578125,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.7190184049079753,
|
|
"grad_norm": 6.018088564359161,
|
|
"learning_rate": 4.3140495867768597e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -422.0,
|
|
"loss": 0.0257,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1401
|
|
},
|
|
{
|
|
"epoch": 1.720245398773006,
|
|
"grad_norm": 9.64716082744446,
|
|
"learning_rate": 4.3099173553719006e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0384,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.63671875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1402
|
|
},
|
|
{
|
|
"epoch": 1.7214723926380369,
|
|
"grad_norm": 10.130578341464988,
|
|
"learning_rate": 4.305785123966942e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0278,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.490234375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1403
|
|
},
|
|
{
|
|
"epoch": 1.7226993865030675,
|
|
"grad_norm": 11.039748507876151,
|
|
"learning_rate": 4.3016528925619835e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.036,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1404
|
|
},
|
|
{
|
|
"epoch": 1.7239263803680982,
|
|
"grad_norm": 14.523874384158516,
|
|
"learning_rate": 4.2975206611570245e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0472,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.578125,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1405
|
|
},
|
|
{
|
|
"epoch": 1.7251533742331289,
|
|
"grad_norm": 9.063176566475141,
|
|
"learning_rate": 4.293388429752066e-07,
|
|
"logits/chosen": -0.37109375,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0238,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.0693359375,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1406
|
|
},
|
|
{
|
|
"epoch": 1.7263803680981595,
|
|
"grad_norm": 15.952092507687093,
|
|
"learning_rate": 4.289256198347107e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0578,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1407
|
|
},
|
|
{
|
|
"epoch": 1.7276073619631902,
|
|
"grad_norm": 10.98220621553548,
|
|
"learning_rate": 4.2851239669421484e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -468.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0479,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1408
|
|
},
|
|
{
|
|
"epoch": 1.7288343558282209,
|
|
"grad_norm": 12.984611394573623,
|
|
"learning_rate": 4.28099173553719e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0459,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.380859375,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1409
|
|
},
|
|
{
|
|
"epoch": 1.7300613496932515,
|
|
"grad_norm": 11.734603858184595,
|
|
"learning_rate": 4.2768595041322313e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.06,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": -0.0009307861328125,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1410
|
|
},
|
|
{
|
|
"epoch": 1.7312883435582822,
|
|
"grad_norm": 8.211070448578011,
|
|
"learning_rate": 4.272727272727272e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0301,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.031494140625,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1411
|
|
},
|
|
{
|
|
"epoch": 1.7325153374233129,
|
|
"grad_norm": 13.425279256537092,
|
|
"learning_rate": 4.2685950413223137e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0574,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.75390625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1412
|
|
},
|
|
{
|
|
"epoch": 1.7337423312883435,
|
|
"grad_norm": 12.01341319433186,
|
|
"learning_rate": 4.264462809917355e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0513,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.0439453125,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1413
|
|
},
|
|
{
|
|
"epoch": 1.7349693251533742,
|
|
"grad_norm": 8.496894022138902,
|
|
"learning_rate": 4.2603305785123967e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0355,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1414
|
|
},
|
|
{
|
|
"epoch": 1.7361963190184049,
|
|
"grad_norm": 11.478176301169068,
|
|
"learning_rate": 4.256198347107438e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0468,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.28515625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1415
|
|
},
|
|
{
|
|
"epoch": 1.7374233128834355,
|
|
"grad_norm": 12.073139914109973,
|
|
"learning_rate": 4.252066115702479e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0407,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1416
|
|
},
|
|
{
|
|
"epoch": 1.7386503067484662,
|
|
"grad_norm": 8.60220265310358,
|
|
"learning_rate": 4.2479338842975205e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0291,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.40625,
|
|
"rewards/margins": 10.3125,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1417
|
|
},
|
|
{
|
|
"epoch": 1.7398773006134969,
|
|
"grad_norm": 13.270979885719917,
|
|
"learning_rate": 4.243801652892562e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0386,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.412109375,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1418
|
|
},
|
|
{
|
|
"epoch": 1.7411042944785278,
|
|
"grad_norm": 9.195409586615746,
|
|
"learning_rate": 4.2396694214876035e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.045,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.1591796875,
|
|
"rewards/margins": 9.0625,
|
|
"rewards/rejected": -9.25,
|
|
"step": 1419
|
|
},
|
|
{
|
|
"epoch": 1.7423312883435584,
|
|
"grad_norm": 12.730546721565815,
|
|
"learning_rate": 4.2355371900826444e-07,
|
|
"logits/chosen": -0.373046875,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.0347,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.470703125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1420
|
|
},
|
|
{
|
|
"epoch": 1.743558282208589,
|
|
"grad_norm": 6.95704360931474,
|
|
"learning_rate": 4.231404958677686e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.470703125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0262,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.5390625,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1421
|
|
},
|
|
{
|
|
"epoch": 1.7447852760736198,
|
|
"grad_norm": 20.61017669645841,
|
|
"learning_rate": 4.227272727272727e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0702,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.26953125,
|
|
"rewards/margins": 9.6875,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1422
|
|
},
|
|
{
|
|
"epoch": 1.7460122699386504,
|
|
"grad_norm": 8.24095024686675,
|
|
"learning_rate": 4.2231404958677683e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0341,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.52734375,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -11.3125,
|
|
"step": 1423
|
|
},
|
|
{
|
|
"epoch": 1.747239263803681,
|
|
"grad_norm": 7.056980615335467,
|
|
"learning_rate": 4.21900826446281e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.5390625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.291015625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1424
|
|
},
|
|
{
|
|
"epoch": 1.7484662576687118,
|
|
"grad_norm": 11.181146787442435,
|
|
"learning_rate": 4.2148760330578507e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0407,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.337890625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1425
|
|
},
|
|
{
|
|
"epoch": 1.7496932515337424,
|
|
"grad_norm": 12.135481836534872,
|
|
"learning_rate": 4.210743801652892e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0463,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.404296875,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1426
|
|
},
|
|
{
|
|
"epoch": 1.7509202453987731,
|
|
"grad_norm": 18.47606307566779,
|
|
"learning_rate": 4.2066115702479337e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0716,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.56640625,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1427
|
|
},
|
|
{
|
|
"epoch": 1.7521472392638038,
|
|
"grad_norm": 11.883217515506011,
|
|
"learning_rate": 4.202479338842975e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.048,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1428
|
|
},
|
|
{
|
|
"epoch": 1.7533742331288344,
|
|
"grad_norm": 9.359450285608279,
|
|
"learning_rate": 4.198347107438016e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0363,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1429
|
|
},
|
|
{
|
|
"epoch": 1.7546012269938651,
|
|
"grad_norm": 5.403328894023132,
|
|
"learning_rate": 4.1942148760330576e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0207,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.447265625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1430
|
|
},
|
|
{
|
|
"epoch": 1.7558282208588958,
|
|
"grad_norm": 13.88374823643348,
|
|
"learning_rate": 4.190082644628099e-07,
|
|
"logits/chosen": -0.08837890625,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0454,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.2451171875,
|
|
"rewards/margins": 9.25,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1431
|
|
},
|
|
{
|
|
"epoch": 1.7570552147239265,
|
|
"grad_norm": 5.97038097853659,
|
|
"learning_rate": 4.1859504132231405e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0129,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.49609375,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1432
|
|
},
|
|
{
|
|
"epoch": 1.7582822085889571,
|
|
"grad_norm": 17.90462476723502,
|
|
"learning_rate": 4.1818181818181814e-07,
|
|
"logits/chosen": -0.06298828125,
|
|
"logits/rejected": -0.173828125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.082,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.0341796875,
|
|
"rewards/margins": 8.9375,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1433
|
|
},
|
|
{
|
|
"epoch": 1.7595092024539878,
|
|
"grad_norm": 12.219362600170777,
|
|
"learning_rate": 4.177685950413223e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.185546875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0437,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.2470703125,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1434
|
|
},
|
|
{
|
|
"epoch": 1.7607361963190185,
|
|
"grad_norm": 8.236105423146372,
|
|
"learning_rate": 4.1735537190082644e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0309,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.33203125,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1435
|
|
},
|
|
{
|
|
"epoch": 1.7619631901840491,
|
|
"grad_norm": 4.350023627691537,
|
|
"learning_rate": 4.169421487603306e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0149,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1436
|
|
},
|
|
{
|
|
"epoch": 1.7631901840490798,
|
|
"grad_norm": 13.979708771545518,
|
|
"learning_rate": 4.1652892561983473e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0501,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.400390625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1437
|
|
},
|
|
{
|
|
"epoch": 1.7644171779141105,
|
|
"grad_norm": 6.8908623313404656,
|
|
"learning_rate": 4.161157024793388e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0252,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2158203125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1438
|
|
},
|
|
{
|
|
"epoch": 1.7656441717791411,
|
|
"grad_norm": 6.601019632509093,
|
|
"learning_rate": 4.1570247933884297e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.4765625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0215,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.25390625,
|
|
"rewards/margins": 9.5,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1439
|
|
},
|
|
{
|
|
"epoch": 1.7668711656441718,
|
|
"grad_norm": 6.7024300095645275,
|
|
"learning_rate": 4.1528925619834707e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -468.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.0189,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1440
|
|
},
|
|
{
|
|
"epoch": 1.7680981595092025,
|
|
"grad_norm": 11.101280506573113,
|
|
"learning_rate": 4.148760330578512e-07,
|
|
"logits/chosen": -0.3515625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0453,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.1845703125,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1441
|
|
},
|
|
{
|
|
"epoch": 1.7693251533742331,
|
|
"grad_norm": 18.77005914648779,
|
|
"learning_rate": 4.144628099173553e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0532,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.27734375,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1442
|
|
},
|
|
{
|
|
"epoch": 1.7705521472392638,
|
|
"grad_norm": 7.507931073067611,
|
|
"learning_rate": 4.1404958677685946e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0316,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1443
|
|
},
|
|
{
|
|
"epoch": 1.7717791411042945,
|
|
"grad_norm": 17.854965876535875,
|
|
"learning_rate": 4.136363636363636e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0365,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.828125,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1444
|
|
},
|
|
{
|
|
"epoch": 1.7730061349693251,
|
|
"grad_norm": 9.924019883131129,
|
|
"learning_rate": 4.1322314049586775e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0282,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.07470703125,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1445
|
|
},
|
|
{
|
|
"epoch": 1.7742331288343558,
|
|
"grad_norm": 14.644045237262645,
|
|
"learning_rate": 4.128099173553719e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0437,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.248046875,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.9375,
|
|
"step": 1446
|
|
},
|
|
{
|
|
"epoch": 1.7754601226993865,
|
|
"grad_norm": 7.0908605556645865,
|
|
"learning_rate": 4.12396694214876e-07,
|
|
"logits/chosen": -0.302734375,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0345,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.01043701171875,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1447
|
|
},
|
|
{
|
|
"epoch": 1.7766871165644171,
|
|
"grad_norm": 10.894561808859045,
|
|
"learning_rate": 4.1198347107438014e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0386,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.435546875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1448
|
|
},
|
|
{
|
|
"epoch": 1.7779141104294478,
|
|
"grad_norm": 10.71583529637343,
|
|
"learning_rate": 4.115702479338843e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0385,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.162109375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1449
|
|
},
|
|
{
|
|
"epoch": 1.7791411042944785,
|
|
"grad_norm": 14.482137623286459,
|
|
"learning_rate": 4.1115702479338843e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0505,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.31640625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1450
|
|
},
|
|
{
|
|
"epoch": 1.7803680981595091,
|
|
"grad_norm": 7.3249919433337105,
|
|
"learning_rate": 4.1074380165289253e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0276,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.62890625,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1451
|
|
},
|
|
{
|
|
"epoch": 1.7815950920245398,
|
|
"grad_norm": 11.251957500630867,
|
|
"learning_rate": 4.103305785123967e-07,
|
|
"logits/chosen": -0.349609375,
|
|
"logits/rejected": -0.6015625,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0185,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 12.0,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1452
|
|
},
|
|
{
|
|
"epoch": 1.7828220858895705,
|
|
"grad_norm": 6.442063193778657,
|
|
"learning_rate": 4.099173553719008e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0186,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1453
|
|
},
|
|
{
|
|
"epoch": 1.7840490797546011,
|
|
"grad_norm": 7.1845818764252725,
|
|
"learning_rate": 4.0950413223140497e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0241,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.376953125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1454
|
|
},
|
|
{
|
|
"epoch": 1.7852760736196318,
|
|
"grad_norm": 7.446575211100864,
|
|
"learning_rate": 4.090909090909091e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0286,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6171875,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1455
|
|
},
|
|
{
|
|
"epoch": 1.7865030674846625,
|
|
"grad_norm": 14.069147561510022,
|
|
"learning_rate": 4.086776859504132e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0471,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.326171875,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1456
|
|
},
|
|
{
|
|
"epoch": 1.7877300613496931,
|
|
"grad_norm": 11.293899268365516,
|
|
"learning_rate": 4.0826446280991736e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0346,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1457
|
|
},
|
|
{
|
|
"epoch": 1.7889570552147238,
|
|
"grad_norm": 16.691884123502405,
|
|
"learning_rate": 4.0785123966942145e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.19921875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0533,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.1015625,
|
|
"rewards/margins": 8.8125,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1458
|
|
},
|
|
{
|
|
"epoch": 1.7901840490797545,
|
|
"grad_norm": 6.148425377141088,
|
|
"learning_rate": 4.074380165289256e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0184,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1459
|
|
},
|
|
{
|
|
"epoch": 1.7914110429447851,
|
|
"grad_norm": 7.338788152445361,
|
|
"learning_rate": 4.070247933884297e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0285,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.0625,
|
|
"step": 1460
|
|
},
|
|
{
|
|
"epoch": 1.7926380368098158,
|
|
"grad_norm": 14.029223166941092,
|
|
"learning_rate": 4.0661157024793384e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.049,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1461
|
|
},
|
|
{
|
|
"epoch": 1.7938650306748465,
|
|
"grad_norm": 11.42908739485831,
|
|
"learning_rate": 4.06198347107438e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0405,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1462
|
|
},
|
|
{
|
|
"epoch": 1.7950920245398772,
|
|
"grad_norm": 5.963109363337343,
|
|
"learning_rate": 4.0578512396694213e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0194,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.0,
|
|
"step": 1463
|
|
},
|
|
{
|
|
"epoch": 1.7963190184049078,
|
|
"grad_norm": 14.956820658115301,
|
|
"learning_rate": 4.053719008264463e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0501,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.435546875,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1464
|
|
},
|
|
{
|
|
"epoch": 1.7975460122699385,
|
|
"grad_norm": 11.354471146566448,
|
|
"learning_rate": 4.049586776859504e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0342,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1465
|
|
},
|
|
{
|
|
"epoch": 1.7987730061349694,
|
|
"grad_norm": 4.766899810723856,
|
|
"learning_rate": 4.045454545454545e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0126,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.33203125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1466
|
|
},
|
|
{
|
|
"epoch": 1.8,
|
|
"grad_norm": 7.212165516394141,
|
|
"learning_rate": 4.0413223140495867e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.024,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1467
|
|
},
|
|
{
|
|
"epoch": 1.8012269938650307,
|
|
"grad_norm": 8.338857694062831,
|
|
"learning_rate": 4.037190082644628e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0224,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1468
|
|
},
|
|
{
|
|
"epoch": 1.8024539877300614,
|
|
"grad_norm": 13.100083487085746,
|
|
"learning_rate": 4.033057851239669e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0498,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.404296875,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1469
|
|
},
|
|
{
|
|
"epoch": 1.803680981595092,
|
|
"grad_norm": 10.916988776212177,
|
|
"learning_rate": 4.0289256198347106e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0578,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.0654296875,
|
|
"rewards/margins": 9.4375,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1470
|
|
},
|
|
{
|
|
"epoch": 1.8049079754601227,
|
|
"grad_norm": 7.143802702613123,
|
|
"learning_rate": 4.024793388429752e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0259,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -8.9375,
|
|
"step": 1471
|
|
},
|
|
{
|
|
"epoch": 1.8061349693251534,
|
|
"grad_norm": 11.957785888173838,
|
|
"learning_rate": 4.0206611570247935e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0331,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1472
|
|
},
|
|
{
|
|
"epoch": 1.807361963190184,
|
|
"grad_norm": 9.172708593888022,
|
|
"learning_rate": 4.0165289256198345e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0343,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.75390625,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1473
|
|
},
|
|
{
|
|
"epoch": 1.8085889570552147,
|
|
"grad_norm": 20.42633762489965,
|
|
"learning_rate": 4.012396694214876e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0481,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1474
|
|
},
|
|
{
|
|
"epoch": 1.8098159509202454,
|
|
"grad_norm": 9.30371862964098,
|
|
"learning_rate": 4.0082644628099174e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0329,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1475
|
|
},
|
|
{
|
|
"epoch": 1.811042944785276,
|
|
"grad_norm": 8.50334602386536,
|
|
"learning_rate": 4.0041322314049583e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0192,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.875,
|
|
"step": 1476
|
|
},
|
|
{
|
|
"epoch": 1.8122699386503067,
|
|
"grad_norm": 10.967449160243186,
|
|
"learning_rate": 4e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.044,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.291015625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1477
|
|
},
|
|
{
|
|
"epoch": 1.8134969325153374,
|
|
"grad_norm": 12.523588802562843,
|
|
"learning_rate": 3.995867768595041e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0407,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1478
|
|
},
|
|
{
|
|
"epoch": 1.814723926380368,
|
|
"grad_norm": 17.617293853198372,
|
|
"learning_rate": 3.991735537190082e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0449,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1479
|
|
},
|
|
{
|
|
"epoch": 1.8159509202453987,
|
|
"grad_norm": 16.725078336197416,
|
|
"learning_rate": 3.9876033057851237e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0556,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.6875,
|
|
"step": 1480
|
|
},
|
|
{
|
|
"epoch": 1.8171779141104294,
|
|
"grad_norm": 14.873599945559999,
|
|
"learning_rate": 3.983471074380165e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0314,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.3671875,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1481
|
|
},
|
|
{
|
|
"epoch": 1.8184049079754603,
|
|
"grad_norm": 10.769303978492216,
|
|
"learning_rate": 3.979338842975206e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0671,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.51171875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1482
|
|
},
|
|
{
|
|
"epoch": 1.819631901840491,
|
|
"grad_norm": 14.575937453093236,
|
|
"learning_rate": 3.9752066115702476e-07,
|
|
"logits/chosen": -0.054443359375,
|
|
"logits/rejected": -0.2080078125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.0595,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.263671875,
|
|
"rewards/margins": 9.5625,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1483
|
|
},
|
|
{
|
|
"epoch": 1.8208588957055216,
|
|
"grad_norm": 14.8789085654287,
|
|
"learning_rate": 3.971074380165289e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0387,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.10107421875,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1484
|
|
},
|
|
{
|
|
"epoch": 1.8220858895705523,
|
|
"grad_norm": 6.030017337982837,
|
|
"learning_rate": 3.9669421487603305e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -406.0,
|
|
"loss": 0.0167,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.09619140625,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1485
|
|
},
|
|
{
|
|
"epoch": 1.823312883435583,
|
|
"grad_norm": 15.152829519992128,
|
|
"learning_rate": 3.962809917355372e-07,
|
|
"logits/chosen": -0.3359375,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0597,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.265625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1486
|
|
},
|
|
{
|
|
"epoch": 1.8245398773006136,
|
|
"grad_norm": 6.213481541736639,
|
|
"learning_rate": 3.958677685950413e-07,
|
|
"logits/chosen": -0.38671875,
|
|
"logits/rejected": -0.5390625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0196,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1487
|
|
},
|
|
{
|
|
"epoch": 1.8257668711656443,
|
|
"grad_norm": 13.57994141249716,
|
|
"learning_rate": 3.9545454545454544e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0264,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1488
|
|
},
|
|
{
|
|
"epoch": 1.826993865030675,
|
|
"grad_norm": 9.928534941983283,
|
|
"learning_rate": 3.950413223140496e-07,
|
|
"logits/chosen": -0.373046875,
|
|
"logits/rejected": -0.57421875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0345,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1489
|
|
},
|
|
{
|
|
"epoch": 1.8282208588957056,
|
|
"grad_norm": 9.667869262637105,
|
|
"learning_rate": 3.9462809917355374e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0303,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1490
|
|
},
|
|
{
|
|
"epoch": 1.8294478527607363,
|
|
"grad_norm": 14.604470050902151,
|
|
"learning_rate": 3.9421487603305783e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -480.0,
|
|
"logps/rejected": -426.0,
|
|
"loss": 0.0453,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 11.4375,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1491
|
|
},
|
|
{
|
|
"epoch": 1.830674846625767,
|
|
"grad_norm": 11.486448484848255,
|
|
"learning_rate": 3.93801652892562e-07,
|
|
"logits/chosen": -0.390625,
|
|
"logits/rejected": -0.53515625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0388,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.1787109375,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1492
|
|
},
|
|
{
|
|
"epoch": 1.8319018404907976,
|
|
"grad_norm": 9.225775675956832,
|
|
"learning_rate": 3.933884297520661e-07,
|
|
"logits/chosen": -0.34375,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.037,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.56640625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1493
|
|
},
|
|
{
|
|
"epoch": 1.8331288343558283,
|
|
"grad_norm": 7.2985591147140205,
|
|
"learning_rate": 3.929752066115702e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.0318,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.181640625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1494
|
|
},
|
|
{
|
|
"epoch": 1.834355828220859,
|
|
"grad_norm": 9.350892300250125,
|
|
"learning_rate": 3.9256198347107437e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0313,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1495
|
|
},
|
|
{
|
|
"epoch": 1.8355828220858896,
|
|
"grad_norm": 8.057530987361305,
|
|
"learning_rate": 3.9214876033057846e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0202,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1496
|
|
},
|
|
{
|
|
"epoch": 1.8368098159509203,
|
|
"grad_norm": 5.551190334543901,
|
|
"learning_rate": 3.917355371900826e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0191,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1497
|
|
},
|
|
{
|
|
"epoch": 1.838036809815951,
|
|
"grad_norm": 15.809682440160046,
|
|
"learning_rate": 3.9132231404958675e-07,
|
|
"logits/chosen": -0.35546875,
|
|
"logits/rejected": -0.53125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.0582,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.337890625,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1498
|
|
},
|
|
{
|
|
"epoch": 1.8392638036809816,
|
|
"grad_norm": 5.390092854825261,
|
|
"learning_rate": 3.909090909090909e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0202,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.625,
|
|
"step": 1499
|
|
},
|
|
{
|
|
"epoch": 1.8404907975460123,
|
|
"grad_norm": 12.262156805884102,
|
|
"learning_rate": 3.90495867768595e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.053,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.1875,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.841717791411043,
|
|
"grad_norm": 9.278805331932283,
|
|
"learning_rate": 3.9008264462809914e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0293,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.9921875,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1501
|
|
},
|
|
{
|
|
"epoch": 1.8429447852760736,
|
|
"grad_norm": 11.34165239212778,
|
|
"learning_rate": 3.896694214876033e-07,
|
|
"logits/chosen": -0.11865234375,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0369,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.337890625,
|
|
"rewards/margins": 9.625,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1502
|
|
},
|
|
{
|
|
"epoch": 1.8441717791411043,
|
|
"grad_norm": 13.06187531213226,
|
|
"learning_rate": 3.8925619834710744e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -440.0,
|
|
"loss": 0.0554,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1503
|
|
},
|
|
{
|
|
"epoch": 1.845398773006135,
|
|
"grad_norm": 8.15269649689573,
|
|
"learning_rate": 3.888429752066116e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0248,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.5,
|
|
"step": 1504
|
|
},
|
|
{
|
|
"epoch": 1.8466257668711656,
|
|
"grad_norm": 8.48140297192026,
|
|
"learning_rate": 3.884297520661157e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0335,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.36328125,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1505
|
|
},
|
|
{
|
|
"epoch": 1.8478527607361963,
|
|
"grad_norm": 5.976136545769875,
|
|
"learning_rate": 3.880165289256198e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0223,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.03271484375,
|
|
"rewards/margins": 9.9375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1506
|
|
},
|
|
{
|
|
"epoch": 1.849079754601227,
|
|
"grad_norm": 12.032245428101259,
|
|
"learning_rate": 3.8760330578512397e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0443,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -8.875,
|
|
"step": 1507
|
|
},
|
|
{
|
|
"epoch": 1.8503067484662576,
|
|
"grad_norm": 10.199797501179388,
|
|
"learning_rate": 3.871900826446281e-07,
|
|
"logits/chosen": -0.10546875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0315,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.150390625,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1508
|
|
},
|
|
{
|
|
"epoch": 1.8515337423312883,
|
|
"grad_norm": 6.5671804486968774,
|
|
"learning_rate": 3.867768595041322e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0182,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1509
|
|
},
|
|
{
|
|
"epoch": 1.852760736196319,
|
|
"grad_norm": 11.311752995448698,
|
|
"learning_rate": 3.8636363636363636e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0299,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.7578125,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1510
|
|
},
|
|
{
|
|
"epoch": 1.8539877300613496,
|
|
"grad_norm": 5.595085227865536,
|
|
"learning_rate": 3.8595041322314045e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -442.0,
|
|
"loss": 0.0144,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1511
|
|
},
|
|
{
|
|
"epoch": 1.8552147239263803,
|
|
"grad_norm": 6.453774365694717,
|
|
"learning_rate": 3.855371900826446e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.0234,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 11.5,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 1512
|
|
},
|
|
{
|
|
"epoch": 1.856441717791411,
|
|
"grad_norm": 11.243496057329885,
|
|
"learning_rate": 3.8512396694214875e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.0355,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.353515625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1513
|
|
},
|
|
{
|
|
"epoch": 1.8576687116564417,
|
|
"grad_norm": 7.864665267298382,
|
|
"learning_rate": 3.8471074380165284e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0293,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.578125,
|
|
"rewards/margins": 11.75,
|
|
"rewards/rejected": -11.1875,
|
|
"step": 1514
|
|
},
|
|
{
|
|
"epoch": 1.8588957055214723,
|
|
"grad_norm": 14.98455530314727,
|
|
"learning_rate": 3.84297520661157e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0471,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1515
|
|
},
|
|
{
|
|
"epoch": 1.860122699386503,
|
|
"grad_norm": 9.73663053420357,
|
|
"learning_rate": 3.8388429752066114e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.037,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.3984375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1516
|
|
},
|
|
{
|
|
"epoch": 1.8613496932515337,
|
|
"grad_norm": 5.410046220332471,
|
|
"learning_rate": 3.834710743801653e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0189,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1517
|
|
},
|
|
{
|
|
"epoch": 1.8625766871165643,
|
|
"grad_norm": 8.031460913772204,
|
|
"learning_rate": 3.830578512396694e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0349,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.10546875,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.875,
|
|
"step": 1518
|
|
},
|
|
{
|
|
"epoch": 1.863803680981595,
|
|
"grad_norm": 12.750454028160657,
|
|
"learning_rate": 3.826446280991735e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0334,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.671875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1519
|
|
},
|
|
{
|
|
"epoch": 1.8650306748466257,
|
|
"grad_norm": 6.944061399154593,
|
|
"learning_rate": 3.8223140495867767e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.0195,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.01513671875,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1520
|
|
},
|
|
{
|
|
"epoch": 1.8662576687116563,
|
|
"grad_norm": 15.457898244887168,
|
|
"learning_rate": 3.818181818181818e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0429,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.056396484375,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1521
|
|
},
|
|
{
|
|
"epoch": 1.867484662576687,
|
|
"grad_norm": 10.979401286631072,
|
|
"learning_rate": 3.814049586776859e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0385,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.498046875,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1522
|
|
},
|
|
{
|
|
"epoch": 1.8687116564417177,
|
|
"grad_norm": 7.649817660124712,
|
|
"learning_rate": 3.8099173553719006e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0241,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.384765625,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1523
|
|
},
|
|
{
|
|
"epoch": 1.8699386503067483,
|
|
"grad_norm": 16.44611991464789,
|
|
"learning_rate": 3.805785123966942e-07,
|
|
"logits/chosen": -0.12109375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0604,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.5,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1524
|
|
},
|
|
{
|
|
"epoch": 1.871165644171779,
|
|
"grad_norm": 17.771708021322524,
|
|
"learning_rate": 3.8016528925619836e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0502,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.62890625,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1525
|
|
},
|
|
{
|
|
"epoch": 1.8723926380368097,
|
|
"grad_norm": 16.63663539415726,
|
|
"learning_rate": 3.797520661157025e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0696,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.80078125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.5625,
|
|
"step": 1526
|
|
},
|
|
{
|
|
"epoch": 1.8736196319018403,
|
|
"grad_norm": 12.827827932259845,
|
|
"learning_rate": 3.793388429752066e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0427,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.71875,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1527
|
|
},
|
|
{
|
|
"epoch": 1.874846625766871,
|
|
"grad_norm": 9.415527723517455,
|
|
"learning_rate": 3.7892561983471074e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.0489,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1528
|
|
},
|
|
{
|
|
"epoch": 1.876073619631902,
|
|
"grad_norm": 6.670343790682242,
|
|
"learning_rate": 3.7851239669421484e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.48828125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0193,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 12.3125,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1529
|
|
},
|
|
{
|
|
"epoch": 1.8773006134969326,
|
|
"grad_norm": 10.626627973552237,
|
|
"learning_rate": 3.78099173553719e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0456,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.408203125,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1530
|
|
},
|
|
{
|
|
"epoch": 1.8785276073619632,
|
|
"grad_norm": 11.56652832934513,
|
|
"learning_rate": 3.776859504132231e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0515,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.291015625,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1531
|
|
},
|
|
{
|
|
"epoch": 1.879754601226994,
|
|
"grad_norm": 15.226780143354295,
|
|
"learning_rate": 3.7727272727272723e-07,
|
|
"logits/chosen": -0.369140625,
|
|
"logits/rejected": -0.5390625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -428.0,
|
|
"loss": 0.052,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1532
|
|
},
|
|
{
|
|
"epoch": 1.8809815950920246,
|
|
"grad_norm": 7.944046538944383,
|
|
"learning_rate": 3.768595041322314e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0296,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.41796875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1533
|
|
},
|
|
{
|
|
"epoch": 1.8822085889570552,
|
|
"grad_norm": 8.04958912618652,
|
|
"learning_rate": 3.764462809917355e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0269,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.044921875,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1534
|
|
},
|
|
{
|
|
"epoch": 1.883435582822086,
|
|
"grad_norm": 15.263159339847686,
|
|
"learning_rate": 3.7603305785123967e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.0574,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1535
|
|
},
|
|
{
|
|
"epoch": 1.8846625766871166,
|
|
"grad_norm": 18.347643088880886,
|
|
"learning_rate": 3.7561983471074376e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.0722,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1536
|
|
},
|
|
{
|
|
"epoch": 1.8858895705521472,
|
|
"grad_norm": 4.610478591105244,
|
|
"learning_rate": 3.752066115702479e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0145,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -10.1875,
|
|
"step": 1537
|
|
},
|
|
{
|
|
"epoch": 1.887116564417178,
|
|
"grad_norm": 9.31460509302408,
|
|
"learning_rate": 3.7479338842975206e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.016,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -9.75,
|
|
"step": 1538
|
|
},
|
|
{
|
|
"epoch": 1.8883435582822086,
|
|
"grad_norm": 7.656908175355756,
|
|
"learning_rate": 3.743801652892562e-07,
|
|
"logits/chosen": -0.09130859375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0295,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1539
|
|
},
|
|
{
|
|
"epoch": 1.8895705521472392,
|
|
"grad_norm": 10.474289089874233,
|
|
"learning_rate": 3.739669421487603e-07,
|
|
"logits/chosen": -0.365234375,
|
|
"logits/rejected": -0.49609375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -446.0,
|
|
"loss": 0.0353,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1540
|
|
},
|
|
{
|
|
"epoch": 1.89079754601227,
|
|
"grad_norm": 11.653041843490755,
|
|
"learning_rate": 3.7355371900826445e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0467,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -9.8125,
|
|
"step": 1541
|
|
},
|
|
{
|
|
"epoch": 1.8920245398773006,
|
|
"grad_norm": 6.90819622680087,
|
|
"learning_rate": 3.731404958677686e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0172,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1542
|
|
},
|
|
{
|
|
"epoch": 1.8932515337423312,
|
|
"grad_norm": 6.33852479078536,
|
|
"learning_rate": 3.7272727272727274e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.0229,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1543
|
|
},
|
|
{
|
|
"epoch": 1.8944785276073621,
|
|
"grad_norm": 7.96846603898992,
|
|
"learning_rate": 3.723140495867769e-07,
|
|
"logits/chosen": -0.302734375,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -428.0,
|
|
"loss": 0.0271,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1544
|
|
},
|
|
{
|
|
"epoch": 1.8957055214723928,
|
|
"grad_norm": 7.051890288163619,
|
|
"learning_rate": 3.71900826446281e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0245,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.2138671875,
|
|
"rewards/margins": 9.3125,
|
|
"rewards/rejected": -9.125,
|
|
"step": 1545
|
|
},
|
|
{
|
|
"epoch": 1.8969325153374235,
|
|
"grad_norm": 15.805752169858186,
|
|
"learning_rate": 3.7148760330578513e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0441,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5546875,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1546
|
|
},
|
|
{
|
|
"epoch": 1.8981595092024541,
|
|
"grad_norm": 10.178224306999683,
|
|
"learning_rate": 3.710743801652892e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.5390625,
|
|
"logps/chosen": -482.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0259,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1547
|
|
},
|
|
{
|
|
"epoch": 1.8993865030674848,
|
|
"grad_norm": 25.993309160540278,
|
|
"learning_rate": 3.7066115702479337e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0997,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1548
|
|
},
|
|
{
|
|
"epoch": 1.9006134969325155,
|
|
"grad_norm": 7.667616913932587,
|
|
"learning_rate": 3.7024793388429746e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.0379,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.453125,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -9.3125,
|
|
"step": 1549
|
|
},
|
|
{
|
|
"epoch": 1.9018404907975461,
|
|
"grad_norm": 13.167443325757498,
|
|
"learning_rate": 3.698347107438016e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0415,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -10.4375,
|
|
"step": 1550
|
|
},
|
|
{
|
|
"epoch": 1.9030674846625768,
|
|
"grad_norm": 15.637823338670781,
|
|
"learning_rate": 3.6942148760330576e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0506,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.0849609375,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -9.9375,
|
|
"step": 1551
|
|
},
|
|
{
|
|
"epoch": 1.9042944785276075,
|
|
"grad_norm": 14.70210007647305,
|
|
"learning_rate": 3.690082644628099e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.0658,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.365234375,
|
|
"rewards/margins": 9.125,
|
|
"rewards/rejected": -8.8125,
|
|
"step": 1552
|
|
},
|
|
{
|
|
"epoch": 1.9055214723926381,
|
|
"grad_norm": 11.691206237837736,
|
|
"learning_rate": 3.6859504132231405e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0288,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.369140625,
|
|
"rewards/margins": 10.0625,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1553
|
|
},
|
|
{
|
|
"epoch": 1.9067484662576688,
|
|
"grad_norm": 18.284716294660406,
|
|
"learning_rate": 3.6818181818181815e-07,
|
|
"logits/chosen": -0.34765625,
|
|
"logits/rejected": -0.60546875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -388.0,
|
|
"loss": 0.048,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.423828125,
|
|
"rewards/margins": 10.6875,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1554
|
|
},
|
|
{
|
|
"epoch": 1.9079754601226995,
|
|
"grad_norm": 9.24129216670663,
|
|
"learning_rate": 3.677685950413223e-07,
|
|
"logits/chosen": -0.3671875,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.03,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.416015625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1555
|
|
},
|
|
{
|
|
"epoch": 1.9092024539877301,
|
|
"grad_norm": 7.704932745320588,
|
|
"learning_rate": 3.6735537190082644e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.0294,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.73828125,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -9.6875,
|
|
"step": 1556
|
|
},
|
|
{
|
|
"epoch": 1.9104294478527608,
|
|
"grad_norm": 13.710236829789524,
|
|
"learning_rate": 3.669421487603306e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0433,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.44921875,
|
|
"rewards/margins": 9.875,
|
|
"rewards/rejected": -9.375,
|
|
"step": 1557
|
|
},
|
|
{
|
|
"epoch": 1.9116564417177915,
|
|
"grad_norm": 6.318068543623311,
|
|
"learning_rate": 3.665289256198347e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -416.0,
|
|
"loss": 0.0166,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6171875,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1558
|
|
},
|
|
{
|
|
"epoch": 1.9128834355828221,
|
|
"grad_norm": 13.79104557098978,
|
|
"learning_rate": 3.6611570247933883e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0582,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.404296875,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -9.4375,
|
|
"step": 1559
|
|
},
|
|
{
|
|
"epoch": 1.9141104294478528,
|
|
"grad_norm": 5.97530423830877,
|
|
"learning_rate": 3.65702479338843e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0259,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.61328125,
|
|
"rewards/margins": 10.0,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1560
|
|
},
|
|
{
|
|
"epoch": 1.9153374233128835,
|
|
"grad_norm": 11.147232531508921,
|
|
"learning_rate": 3.652892561983471e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.5390625,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0337,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1561
|
|
},
|
|
{
|
|
"epoch": 1.9165644171779141,
|
|
"grad_norm": 3.795404161803849,
|
|
"learning_rate": 3.648760330578512e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -402.0,
|
|
"loss": 0.0112,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 11.625,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1562
|
|
},
|
|
{
|
|
"epoch": 1.9177914110429448,
|
|
"grad_norm": 7.950607336247216,
|
|
"learning_rate": 3.6446280991735536e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.48046875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0267,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.91796875,
|
|
"rewards/margins": 11.5,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1563
|
|
},
|
|
{
|
|
"epoch": 1.9190184049079755,
|
|
"grad_norm": 6.750098939471797,
|
|
"learning_rate": 3.640495867768595e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0229,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.10302734375,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1564
|
|
},
|
|
{
|
|
"epoch": 1.9202453987730062,
|
|
"grad_norm": 4.13433249219941,
|
|
"learning_rate": 3.636363636363636e-07,
|
|
"logits/chosen": -0.37109375,
|
|
"logits/rejected": -0.55859375,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0195,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.333984375,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1565
|
|
},
|
|
{
|
|
"epoch": 1.9214723926380368,
|
|
"grad_norm": 22.18088920794604,
|
|
"learning_rate": 3.6322314049586775e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0589,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 11.625,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1566
|
|
},
|
|
{
|
|
"epoch": 1.9226993865030675,
|
|
"grad_norm": 10.561203616639823,
|
|
"learning_rate": 3.6280991735537185e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0375,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.1279296875,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1567
|
|
},
|
|
{
|
|
"epoch": 1.9239263803680982,
|
|
"grad_norm": 12.75740234673262,
|
|
"learning_rate": 3.62396694214876e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1568
|
|
},
|
|
{
|
|
"epoch": 1.9251533742331288,
|
|
"grad_norm": 5.026166618428857,
|
|
"learning_rate": 3.6198347107438014e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0153,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.1103515625,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1569
|
|
},
|
|
{
|
|
"epoch": 1.9263803680981595,
|
|
"grad_norm": 9.832323660191497,
|
|
"learning_rate": 3.615702479338843e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0342,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.1416015625,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1570
|
|
},
|
|
{
|
|
"epoch": 1.9276073619631902,
|
|
"grad_norm": 13.920385434409814,
|
|
"learning_rate": 3.611570247933884e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.0492,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.515625,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1571
|
|
},
|
|
{
|
|
"epoch": 1.9288343558282208,
|
|
"grad_norm": 3.4134224721088477,
|
|
"learning_rate": 3.6074380165289253e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0112,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.0245361328125,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -11.5,
|
|
"step": 1572
|
|
},
|
|
{
|
|
"epoch": 1.9300613496932515,
|
|
"grad_norm": 11.116255169152767,
|
|
"learning_rate": 3.603305785123967e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0452,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.515625,
|
|
"rewards/margins": 10.125,
|
|
"rewards/rejected": -10.625,
|
|
"step": 1573
|
|
},
|
|
{
|
|
"epoch": 1.9312883435582822,
|
|
"grad_norm": 10.02896608166906,
|
|
"learning_rate": 3.599173553719008e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0321,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.1025390625,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1574
|
|
},
|
|
{
|
|
"epoch": 1.9325153374233128,
|
|
"grad_norm": 17.40130304083344,
|
|
"learning_rate": 3.5950413223140497e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.07,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": 0.2890625,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1575
|
|
},
|
|
{
|
|
"epoch": 1.9337423312883435,
|
|
"grad_norm": 5.781863141417342,
|
|
"learning_rate": 3.5909090909090907e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0178,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.734375,
|
|
"rewards/margins": 10.625,
|
|
"rewards/rejected": -11.375,
|
|
"step": 1576
|
|
},
|
|
{
|
|
"epoch": 1.9349693251533742,
|
|
"grad_norm": 8.131365031995118,
|
|
"learning_rate": 3.586776859504132e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.0291,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 11.5,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1577
|
|
},
|
|
{
|
|
"epoch": 1.9361963190184048,
|
|
"grad_norm": 10.709218652851176,
|
|
"learning_rate": 3.5826446280991736e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0347,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.26953125,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -11.25,
|
|
"step": 1578
|
|
},
|
|
{
|
|
"epoch": 1.9374233128834355,
|
|
"grad_norm": 7.249673210758022,
|
|
"learning_rate": 3.578512396694215e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.027,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.10595703125,
|
|
"rewards/margins": 10.8125,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1579
|
|
},
|
|
{
|
|
"epoch": 1.9386503067484662,
|
|
"grad_norm": 7.675554169427657,
|
|
"learning_rate": 3.574380165289256e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -410.0,
|
|
"loss": 0.0204,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -11.0,
|
|
"step": 1580
|
|
},
|
|
{
|
|
"epoch": 1.9398773006134968,
|
|
"grad_norm": 6.9165090446756015,
|
|
"learning_rate": 3.5702479338842975e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0205,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.65625,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.6875,
|
|
"step": 1581
|
|
},
|
|
{
|
|
"epoch": 1.9411042944785275,
|
|
"grad_norm": 9.108540875480172,
|
|
"learning_rate": 3.5661157024793384e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0489,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.177734375,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1582
|
|
},
|
|
{
|
|
"epoch": 1.9423312883435582,
|
|
"grad_norm": 13.715595140114255,
|
|
"learning_rate": 3.56198347107438e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -392.0,
|
|
"loss": 0.0456,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1583
|
|
},
|
|
{
|
|
"epoch": 1.9435582822085888,
|
|
"grad_norm": 13.734215043255173,
|
|
"learning_rate": 3.5578512396694214e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0557,
|
|
"rewards/accuracies": 0.9609375,
|
|
"rewards/chosen": 0.2275390625,
|
|
"rewards/margins": 10.1875,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1584
|
|
},
|
|
{
|
|
"epoch": 1.9447852760736195,
|
|
"grad_norm": 9.856633180140339,
|
|
"learning_rate": 3.5537190082644623e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -424.0,
|
|
"loss": 0.0279,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 12.3125,
|
|
"rewards/rejected": -11.375,
|
|
"step": 1585
|
|
},
|
|
{
|
|
"epoch": 1.9460122699386502,
|
|
"grad_norm": 9.741335079817274,
|
|
"learning_rate": 3.549586776859504e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0336,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.375,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.0,
|
|
"step": 1586
|
|
},
|
|
{
|
|
"epoch": 1.9472392638036808,
|
|
"grad_norm": 9.931830276819845,
|
|
"learning_rate": 3.545454545454545e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.0392,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.013671875,
|
|
"rewards/margins": 10.25,
|
|
"rewards/rejected": -10.3125,
|
|
"step": 1587
|
|
},
|
|
{
|
|
"epoch": 1.9484662576687115,
|
|
"grad_norm": 5.077544005638982,
|
|
"learning_rate": 3.5413223140495867e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.028,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.3828125,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.0625,
|
|
"step": 1588
|
|
},
|
|
{
|
|
"epoch": 1.9496932515337422,
|
|
"grad_norm": 5.99251352113674,
|
|
"learning_rate": 3.5371900826446277e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.0198,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.1103515625,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -11.1875,
|
|
"step": 1589
|
|
},
|
|
{
|
|
"epoch": 1.9509202453987728,
|
|
"grad_norm": 8.666944032081426,
|
|
"learning_rate": 3.533057851239669e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0255,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.58984375,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1590
|
|
},
|
|
{
|
|
"epoch": 1.9521472392638037,
|
|
"grad_norm": 5.466349867166155,
|
|
"learning_rate": 3.5289256198347106e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0166,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.0732421875,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1591
|
|
},
|
|
{
|
|
"epoch": 1.9533742331288344,
|
|
"grad_norm": 8.401550026450433,
|
|
"learning_rate": 3.524793388429752e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0336,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.4140625,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -10.875,
|
|
"step": 1592
|
|
},
|
|
{
|
|
"epoch": 1.954601226993865,
|
|
"grad_norm": 7.339335945531898,
|
|
"learning_rate": 3.5206611570247935e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -390.0,
|
|
"loss": 0.0231,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.251953125,
|
|
"rewards/margins": 11.25,
|
|
"rewards/rejected": -11.5625,
|
|
"step": 1593
|
|
},
|
|
{
|
|
"epoch": 1.9558282208588957,
|
|
"grad_norm": 11.870910323143484,
|
|
"learning_rate": 3.5165289256198345e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.0634,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.0673828125,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1594
|
|
},
|
|
{
|
|
"epoch": 1.9570552147239264,
|
|
"grad_norm": 13.816013175513662,
|
|
"learning_rate": 3.512396694214876e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0389,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.58203125,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -12.0,
|
|
"step": 1595
|
|
},
|
|
{
|
|
"epoch": 1.958282208588957,
|
|
"grad_norm": 6.9710320429295844,
|
|
"learning_rate": 3.5082644628099174e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0213,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.29296875,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -11.25,
|
|
"step": 1596
|
|
},
|
|
{
|
|
"epoch": 1.9595092024539877,
|
|
"grad_norm": 9.747700068565688,
|
|
"learning_rate": 3.504132231404959e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0296,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.10009765625,
|
|
"rewards/margins": 11.75,
|
|
"rewards/rejected": -11.6875,
|
|
"step": 1597
|
|
},
|
|
{
|
|
"epoch": 1.9607361963190184,
|
|
"grad_norm": 18.638853579643484,
|
|
"learning_rate": 3.5e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -418.0,
|
|
"loss": 0.078,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.51953125,
|
|
"rewards/margins": 10.75,
|
|
"rewards/rejected": -11.25,
|
|
"step": 1598
|
|
},
|
|
{
|
|
"epoch": 1.961963190184049,
|
|
"grad_norm": 6.898933701538754,
|
|
"learning_rate": 3.4958677685950413e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.0232,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.07177734375,
|
|
"rewards/margins": 11.0,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1599
|
|
},
|
|
{
|
|
"epoch": 1.9631901840490797,
|
|
"grad_norm": 9.79067027138803,
|
|
"learning_rate": 3.491735537190082e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0294,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.11328125,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.75,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.9644171779141104,
|
|
"grad_norm": 10.685805672567945,
|
|
"learning_rate": 3.4876033057851237e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.0484,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.62109375,
|
|
"rewards/margins": 9.75,
|
|
"rewards/rejected": -10.375,
|
|
"step": 1601
|
|
},
|
|
{
|
|
"epoch": 1.965644171779141,
|
|
"grad_norm": 11.740531113887924,
|
|
"learning_rate": 3.483471074380165e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.0346,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 11.125,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1602
|
|
},
|
|
{
|
|
"epoch": 1.9668711656441717,
|
|
"grad_norm": 8.669620191532285,
|
|
"learning_rate": 3.479338842975206e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -470.0,
|
|
"logps/rejected": -422.0,
|
|
"loss": 0.0311,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.09033203125,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -11.875,
|
|
"step": 1603
|
|
},
|
|
{
|
|
"epoch": 1.9680981595092024,
|
|
"grad_norm": 12.965515385461424,
|
|
"learning_rate": 3.4752066115702476e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.0412,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.25390625,
|
|
"rewards/margins": 10.375,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1604
|
|
},
|
|
{
|
|
"epoch": 1.969325153374233,
|
|
"grad_norm": 13.974528204714584,
|
|
"learning_rate": 3.471074380165289e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.058,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.4375,
|
|
"rewards/margins": 10.4375,
|
|
"rewards/rejected": -10.875,
|
|
"step": 1605
|
|
},
|
|
{
|
|
"epoch": 1.9705521472392638,
|
|
"grad_norm": 7.746586397313187,
|
|
"learning_rate": 3.4669421487603306e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -424.0,
|
|
"loss": 0.0225,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -11.125,
|
|
"step": 1606
|
|
},
|
|
{
|
|
"epoch": 1.9717791411042946,
|
|
"grad_norm": 7.748575507990979,
|
|
"learning_rate": 3.4628099173553715e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.0272,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.412109375,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -10.9375,
|
|
"step": 1607
|
|
},
|
|
{
|
|
"epoch": 1.9730061349693253,
|
|
"grad_norm": 13.435867421827172,
|
|
"learning_rate": 3.458677685950413e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -398.0,
|
|
"loss": 0.0317,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.287109375,
|
|
"rewards/margins": 11.375,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1608
|
|
},
|
|
{
|
|
"epoch": 1.974233128834356,
|
|
"grad_norm": 15.769895113467898,
|
|
"learning_rate": 3.4545454545454544e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -386.0,
|
|
"loss": 0.0527,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.453125,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -11.375,
|
|
"step": 1609
|
|
},
|
|
{
|
|
"epoch": 1.9754601226993866,
|
|
"grad_norm": 27.285867342336143,
|
|
"learning_rate": 3.450413223140496e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.0378,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.0216064453125,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.9375,
|
|
"step": 1610
|
|
},
|
|
{
|
|
"epoch": 1.9766871165644173,
|
|
"grad_norm": 12.445604660791277,
|
|
"learning_rate": 3.446280991735537e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.0573,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.75390625,
|
|
"rewards/margins": 9.375,
|
|
"rewards/rejected": -10.125,
|
|
"step": 1611
|
|
},
|
|
{
|
|
"epoch": 1.977914110429448,
|
|
"grad_norm": 12.10779950431408,
|
|
"learning_rate": 3.4421487603305783e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -450.0,
|
|
"loss": 0.0382,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.5390625,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -11.125,
|
|
"step": 1612
|
|
},
|
|
{
|
|
"epoch": 1.9791411042944786,
|
|
"grad_norm": 8.047966924278876,
|
|
"learning_rate": 3.43801652892562e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -408.0,
|
|
"loss": 0.0214,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.20703125,
|
|
"rewards/margins": 12.25,
|
|
"rewards/rejected": -12.0625,
|
|
"step": 1613
|
|
},
|
|
{
|
|
"epoch": 1.9803680981595093,
|
|
"grad_norm": 13.116491381616166,
|
|
"learning_rate": 3.433884297520661e-07,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.53125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -420.0,
|
|
"loss": 0.0587,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.054931640625,
|
|
"rewards/margins": 11.6875,
|
|
"rewards/rejected": -11.625,
|
|
"step": 1614
|
|
},
|
|
{
|
|
"epoch": 1.98159509202454,
|
|
"grad_norm": 16.786202882853118,
|
|
"learning_rate": 3.429752066115703e-07,
|
|
"logits/chosen": -0.359375,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -404.0,
|
|
"loss": 0.0693,
|
|
"rewards/accuracies": 0.96875,
|
|
"rewards/chosen": 0.1005859375,
|
|
"rewards/margins": 11.8125,
|
|
"rewards/rejected": -11.6875,
|
|
"step": 1615
|
|
},
|
|
{
|
|
"epoch": 1.9828220858895707,
|
|
"grad_norm": 12.829409413140326,
|
|
"learning_rate": 3.4256198347107437e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.0385,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": -0.5078125,
|
|
"rewards/margins": 10.5,
|
|
"rewards/rejected": -11.0,
|
|
"step": 1616
|
|
},
|
|
{
|
|
"epoch": 1.9840490797546013,
|
|
"grad_norm": 18.122092647645268,
|
|
"learning_rate": 3.421487603305785e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.0511,
|
|
"rewards/accuracies": 0.9765625,
|
|
"rewards/chosen": -0.400390625,
|
|
"rewards/margins": 9.8125,
|
|
"rewards/rejected": -10.25,
|
|
"step": 1617
|
|
},
|
|
{
|
|
"epoch": 1.985276073619632,
|
|
"grad_norm": 8.195211426183509,
|
|
"learning_rate": 3.417355371900826e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.0299,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.000885009765625,
|
|
"rewards/margins": 10.875,
|
|
"rewards/rejected": -10.875,
|
|
"step": 1618
|
|
},
|
|
{
|
|
"epoch": 1.9865030674846627,
|
|
"grad_norm": 8.464344293060527,
|
|
"learning_rate": 3.4132231404958676e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.0338,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": -0.22265625,
|
|
"rewards/margins": 10.9375,
|
|
"rewards/rejected": -11.125,
|
|
"step": 1619
|
|
},
|
|
{
|
|
"epoch": 1.9877300613496933,
|
|
"grad_norm": 11.021039310577015,
|
|
"learning_rate": 3.4090909090909085e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.0253,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.00183868408203125,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1620
|
|
},
|
|
{
|
|
"epoch": 1.988957055214724,
|
|
"grad_norm": 6.287061732273872,
|
|
"learning_rate": 3.40495867768595e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0227,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.216796875,
|
|
"rewards/margins": 11.3125,
|
|
"rewards/rejected": -11.0625,
|
|
"step": 1621
|
|
},
|
|
{
|
|
"epoch": 1.9901840490797547,
|
|
"grad_norm": 9.17892638805544,
|
|
"learning_rate": 3.4008264462809914e-07,
|
|
"logits/chosen": -0.4453125,
|
|
"logits/rejected": -0.578125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -436.0,
|
|
"loss": 0.034,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": -0.369140625,
|
|
"rewards/margins": 11.0625,
|
|
"rewards/rejected": -11.4375,
|
|
"step": 1622
|
|
},
|
|
{
|
|
"epoch": 1.9914110429447853,
|
|
"grad_norm": 9.092422311744626,
|
|
"learning_rate": 3.396694214876033e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0203,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.330078125,
|
|
"rewards/margins": 11.625,
|
|
"rewards/rejected": -11.3125,
|
|
"step": 1623
|
|
},
|
|
{
|
|
"epoch": 1.992638036809816,
|
|
"grad_norm": 5.684286344549792,
|
|
"learning_rate": 3.3925619834710744e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.546875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.0194,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.56640625,
|
|
"rewards/margins": 11.75,
|
|
"rewards/rejected": -11.1875,
|
|
"step": 1624
|
|
},
|
|
{
|
|
"epoch": 1.9938650306748467,
|
|
"grad_norm": 7.850956245380302,
|
|
"learning_rate": 3.3884297520661153e-07,
|
|
"logits/chosen": -0.1201171875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -400.0,
|
|
"loss": 0.024,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 11.5,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1625
|
|
},
|
|
{
|
|
"epoch": 1.9950920245398773,
|
|
"grad_norm": 17.273452397613575,
|
|
"learning_rate": 3.384297520661157e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -414.0,
|
|
"loss": 0.0586,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.032958984375,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.5,
|
|
"step": 1626
|
|
},
|
|
{
|
|
"epoch": 1.996319018404908,
|
|
"grad_norm": 9.958407286584684,
|
|
"learning_rate": 3.3801652892561983e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.0374,
|
|
"rewards/accuracies": 0.984375,
|
|
"rewards/chosen": 0.416015625,
|
|
"rewards/margins": 11.1875,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1627
|
|
},
|
|
{
|
|
"epoch": 1.9975460122699387,
|
|
"grad_norm": 10.140409437858745,
|
|
"learning_rate": 3.37603305785124e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.0394,
|
|
"rewards/accuracies": 0.9921875,
|
|
"rewards/chosen": 0.7109375,
|
|
"rewards/margins": 11.5625,
|
|
"rewards/rejected": -10.8125,
|
|
"step": 1628
|
|
},
|
|
{
|
|
"epoch": 1.9987730061349693,
|
|
"grad_norm": 11.087425946230793,
|
|
"learning_rate": 3.3719008264462807e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -396.0,
|
|
"loss": 0.0392,
|
|
"rewards/accuracies": 1.0,
|
|
"rewards/chosen": 0.028564453125,
|
|
"rewards/margins": 10.5625,
|
|
"rewards/rejected": -10.5625,
|
|
"step": 1629
|
|
},
|
|
{
|
|
"epoch": 2.0,
|
|
"grad_norm": 9.7441598484212,
|
|
"learning_rate": 3.367768595041322e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -412.0,
|
|
"loss": 0.0308,
|
|
"rewards/accuracies": 0.9915966987609863,
|
|
"rewards/chosen": 0.333984375,
|
|
"rewards/margins": 11.75,
|
|
"rewards/rejected": -11.4375,
|
|
"step": 1630
|
|
}
|
|
],
|
|
"logging_steps": 1,
|
|
"max_steps": 2445,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 3,
|
|
"save_steps": 500,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": false
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 0.0,
|
|
"train_batch_size": 4,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|