71cf44d51f
Model: shisa-ai/ablation-145-a128.dpo.armorm.rp.tl.8e7-shisa-v2-llama-3.1-8b Source: Original Platform
12260 lines
359 KiB
JSON
12260 lines
359 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 1.0,
|
|
"eval_steps": 500,
|
|
"global_step": 815,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.001226993865030675,
|
|
"grad_norm": 151.42932274482283,
|
|
"learning_rate": 3.2e-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": 162.73344539002164,
|
|
"learning_rate": 6.4e-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": 111.8042794987884,
|
|
"learning_rate": 9.599999999999999e-08,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.6886,
|
|
"rewards/accuracies": 0.3046875,
|
|
"rewards/chosen": 0.01531982421875,
|
|
"rewards/margins": 0.01361083984375,
|
|
"rewards/rejected": 0.0016632080078125,
|
|
"step": 3
|
|
},
|
|
{
|
|
"epoch": 0.0049079754601227,
|
|
"grad_norm": 155.41451533543852,
|
|
"learning_rate": 1.28e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -242.0,
|
|
"loss": 0.7131,
|
|
"rewards/accuracies": 0.328125,
|
|
"rewards/chosen": -0.00408935546875,
|
|
"rewards/margins": -0.0218505859375,
|
|
"rewards/rejected": 0.017822265625,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.006134969325153374,
|
|
"grad_norm": 307.019899938903,
|
|
"learning_rate": 1.6e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.6906,
|
|
"rewards/accuracies": 0.3515625,
|
|
"rewards/chosen": 0.01171875,
|
|
"rewards/margins": 0.01251220703125,
|
|
"rewards/rejected": -0.00078582763671875,
|
|
"step": 5
|
|
},
|
|
{
|
|
"epoch": 0.007361963190184049,
|
|
"grad_norm": 340.66722515261114,
|
|
"learning_rate": 1.9199999999999997e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.7151,
|
|
"rewards/accuracies": 0.25,
|
|
"rewards/chosen": -0.0228271484375,
|
|
"rewards/margins": -0.034912109375,
|
|
"rewards/rejected": 0.01220703125,
|
|
"step": 6
|
|
},
|
|
{
|
|
"epoch": 0.008588957055214725,
|
|
"grad_norm": 239.24556126143491,
|
|
"learning_rate": 2.2400000000000002e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.7032,
|
|
"rewards/accuracies": 0.2734375,
|
|
"rewards/chosen": -0.0003414154052734375,
|
|
"rewards/margins": -0.01373291015625,
|
|
"rewards/rejected": 0.01336669921875,
|
|
"step": 7
|
|
},
|
|
{
|
|
"epoch": 0.0098159509202454,
|
|
"grad_norm": 139.49365426176607,
|
|
"learning_rate": 2.56e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.7012,
|
|
"rewards/accuracies": 0.3671875,
|
|
"rewards/chosen": 0.0284423828125,
|
|
"rewards/margins": -0.00927734375,
|
|
"rewards/rejected": 0.037841796875,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.011042944785276074,
|
|
"grad_norm": 122.82829055069632,
|
|
"learning_rate": 2.88e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.6871,
|
|
"rewards/accuracies": 0.375,
|
|
"rewards/chosen": 0.037353515625,
|
|
"rewards/margins": 0.0244140625,
|
|
"rewards/rejected": 0.01300048828125,
|
|
"step": 9
|
|
},
|
|
{
|
|
"epoch": 0.012269938650306749,
|
|
"grad_norm": 137.4984961934261,
|
|
"learning_rate": 3.2e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.6605,
|
|
"rewards/accuracies": 0.46875,
|
|
"rewards/chosen": 0.057861328125,
|
|
"rewards/margins": 0.07275390625,
|
|
"rewards/rejected": -0.01495361328125,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.013496932515337423,
|
|
"grad_norm": 154.04829614927058,
|
|
"learning_rate": 3.52e-07,
|
|
"logits/chosen": -0.388671875,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.6512,
|
|
"rewards/accuracies": 0.4765625,
|
|
"rewards/chosen": 0.0625,
|
|
"rewards/margins": 0.09912109375,
|
|
"rewards/rejected": -0.036376953125,
|
|
"step": 11
|
|
},
|
|
{
|
|
"epoch": 0.014723926380368098,
|
|
"grad_norm": 292.49015304309825,
|
|
"learning_rate": 3.8399999999999994e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.6321,
|
|
"rewards/accuracies": 0.53125,
|
|
"rewards/chosen": 0.080078125,
|
|
"rewards/margins": 0.1435546875,
|
|
"rewards/rejected": -0.06298828125,
|
|
"step": 12
|
|
},
|
|
{
|
|
"epoch": 0.015950920245398775,
|
|
"grad_norm": 179.53722097804638,
|
|
"learning_rate": 4.16e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -249.0,
|
|
"loss": 0.6194,
|
|
"rewards/accuracies": 0.6171875,
|
|
"rewards/chosen": 0.12109375,
|
|
"rewards/margins": 0.1845703125,
|
|
"rewards/rejected": -0.0634765625,
|
|
"step": 13
|
|
},
|
|
{
|
|
"epoch": 0.01717791411042945,
|
|
"grad_norm": 108.27195029021392,
|
|
"learning_rate": 4.4800000000000004e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -472.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.5477,
|
|
"rewards/accuracies": 0.6875,
|
|
"rewards/chosen": 0.310546875,
|
|
"rewards/margins": 0.408203125,
|
|
"rewards/rejected": -0.09814453125,
|
|
"step": 14
|
|
},
|
|
{
|
|
"epoch": 0.018404907975460124,
|
|
"grad_norm": 71.36850002174661,
|
|
"learning_rate": 4.8e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.5893,
|
|
"rewards/accuracies": 0.59375,
|
|
"rewards/chosen": 0.2412109375,
|
|
"rewards/margins": 0.337890625,
|
|
"rewards/rejected": -0.09619140625,
|
|
"step": 15
|
|
},
|
|
{
|
|
"epoch": 0.0196319018404908,
|
|
"grad_norm": 80.69275734688873,
|
|
"learning_rate": 5.12e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.5357,
|
|
"rewards/accuracies": 0.6484375,
|
|
"rewards/chosen": 0.318359375,
|
|
"rewards/margins": 0.48828125,
|
|
"rewards/rejected": -0.1689453125,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.020858895705521473,
|
|
"grad_norm": 103.00148200369824,
|
|
"learning_rate": 5.44e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -255.0,
|
|
"loss": 0.5279,
|
|
"rewards/accuracies": 0.6953125,
|
|
"rewards/chosen": 0.369140625,
|
|
"rewards/margins": 0.54296875,
|
|
"rewards/rejected": -0.1728515625,
|
|
"step": 17
|
|
},
|
|
{
|
|
"epoch": 0.022085889570552148,
|
|
"grad_norm": 90.51865254706846,
|
|
"learning_rate": 5.76e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.5097,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 0.419921875,
|
|
"rewards/margins": 0.66015625,
|
|
"rewards/rejected": -0.23828125,
|
|
"step": 18
|
|
},
|
|
{
|
|
"epoch": 0.023312883435582823,
|
|
"grad_norm": 63.171857373704086,
|
|
"learning_rate": 6.079999999999999e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.5237,
|
|
"rewards/accuracies": 0.6953125,
|
|
"rewards/chosen": 0.4375,
|
|
"rewards/margins": 0.80078125,
|
|
"rewards/rejected": -0.36328125,
|
|
"step": 19
|
|
},
|
|
{
|
|
"epoch": 0.024539877300613498,
|
|
"grad_norm": 44.73691218363087,
|
|
"learning_rate": 6.4e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -462.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.4345,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 1.3984375,
|
|
"rewards/rejected": -0.5703125,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.025766871165644172,
|
|
"grad_norm": 60.19744308670884,
|
|
"learning_rate": 6.72e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.4418,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.62890625,
|
|
"rewards/margins": 1.1328125,
|
|
"rewards/rejected": -0.50390625,
|
|
"step": 21
|
|
},
|
|
{
|
|
"epoch": 0.026993865030674847,
|
|
"grad_norm": 60.06635545400663,
|
|
"learning_rate": 7.04e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.4638,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 1.25,
|
|
"rewards/rejected": -0.61328125,
|
|
"step": 22
|
|
},
|
|
{
|
|
"epoch": 0.02822085889570552,
|
|
"grad_norm": 81.66202093067754,
|
|
"learning_rate": 7.36e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.5093,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 1.421875,
|
|
"rewards/rejected": -0.6015625,
|
|
"step": 23
|
|
},
|
|
{
|
|
"epoch": 0.029447852760736196,
|
|
"grad_norm": 83.23689816485307,
|
|
"learning_rate": 7.679999999999999e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.511,
|
|
"rewards/accuracies": 0.7265625,
|
|
"rewards/chosen": 0.47265625,
|
|
"rewards/margins": 1.03125,
|
|
"rewards/rejected": -0.5546875,
|
|
"step": 24
|
|
},
|
|
{
|
|
"epoch": 0.03067484662576687,
|
|
"grad_norm": 38.90986069898983,
|
|
"learning_rate": 8e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.4765625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.4067,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 1.7578125,
|
|
"rewards/rejected": -0.90625,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.03190184049079755,
|
|
"grad_norm": 47.94846803093583,
|
|
"learning_rate": 7.996694214876033e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.4388,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 1.84375,
|
|
"rewards/rejected": -0.8203125,
|
|
"step": 26
|
|
},
|
|
{
|
|
"epoch": 0.033128834355828224,
|
|
"grad_norm": 44.30203970039667,
|
|
"learning_rate": 7.993388429752066e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.3991,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 2.234375,
|
|
"rewards/rejected": -1.0,
|
|
"step": 27
|
|
},
|
|
{
|
|
"epoch": 0.0343558282208589,
|
|
"grad_norm": 38.710283535206564,
|
|
"learning_rate": 7.990082644628099e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.1767578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.4113,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 2.015625,
|
|
"rewards/rejected": -0.74609375,
|
|
"step": 28
|
|
},
|
|
{
|
|
"epoch": 0.03558282208588957,
|
|
"grad_norm": 52.910805586215126,
|
|
"learning_rate": 7.986776859504132e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.392,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 2.546875,
|
|
"rewards/rejected": -1.1015625,
|
|
"step": 29
|
|
},
|
|
{
|
|
"epoch": 0.03680981595092025,
|
|
"grad_norm": 45.21313329781316,
|
|
"learning_rate": 7.983471074380164e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.4322,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 2.03125,
|
|
"rewards/rejected": -0.93359375,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.03803680981595092,
|
|
"grad_norm": 41.258011933772664,
|
|
"learning_rate": 7.980165289256198e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.4115,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 2.765625,
|
|
"rewards/rejected": -1.1171875,
|
|
"step": 31
|
|
},
|
|
{
|
|
"epoch": 0.0392638036809816,
|
|
"grad_norm": 39.16160748489701,
|
|
"learning_rate": 7.976859504132231e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3836,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 2.71875,
|
|
"rewards/rejected": -1.1328125,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.04049079754601227,
|
|
"grad_norm": 38.84939653302124,
|
|
"learning_rate": 7.973553719008264e-07,
|
|
"logits/chosen": -0.0869140625,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -234.0,
|
|
"loss": 0.3966,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 2.234375,
|
|
"rewards/rejected": -1.0859375,
|
|
"step": 33
|
|
},
|
|
{
|
|
"epoch": 0.04171779141104295,
|
|
"grad_norm": 44.221302785049176,
|
|
"learning_rate": 7.970247933884297e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.4038,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 2.859375,
|
|
"rewards/rejected": -1.46875,
|
|
"step": 34
|
|
},
|
|
{
|
|
"epoch": 0.04294478527607362,
|
|
"grad_norm": 46.59951208776938,
|
|
"learning_rate": 7.96694214876033e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -264.0,
|
|
"loss": 0.3758,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 2.71875,
|
|
"rewards/rejected": -1.6484375,
|
|
"step": 35
|
|
},
|
|
{
|
|
"epoch": 0.044171779141104296,
|
|
"grad_norm": 41.317139252172254,
|
|
"learning_rate": 7.963636363636364e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3557,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 3.234375,
|
|
"rewards/rejected": -1.7421875,
|
|
"step": 36
|
|
},
|
|
{
|
|
"epoch": 0.04539877300613497,
|
|
"grad_norm": 39.795000825479654,
|
|
"learning_rate": 7.960330578512397e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3698,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 3.328125,
|
|
"rewards/rejected": -1.53125,
|
|
"step": 37
|
|
},
|
|
{
|
|
"epoch": 0.046625766871165646,
|
|
"grad_norm": 35.75368356475873,
|
|
"learning_rate": 7.95702479338843e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3582,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 3.6875,
|
|
"rewards/rejected": -1.671875,
|
|
"step": 38
|
|
},
|
|
{
|
|
"epoch": 0.04785276073619632,
|
|
"grad_norm": 46.53573926538421,
|
|
"learning_rate": 7.953719008264462e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3835,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 3.671875,
|
|
"rewards/rejected": -1.984375,
|
|
"step": 39
|
|
},
|
|
{
|
|
"epoch": 0.049079754601226995,
|
|
"grad_norm": 33.56284839867345,
|
|
"learning_rate": 7.950413223140495e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3042,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 3.765625,
|
|
"rewards/rejected": -2.109375,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 0.05030674846625767,
|
|
"grad_norm": 39.46235435607716,
|
|
"learning_rate": 7.947107438016528e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3708,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 3.546875,
|
|
"rewards/rejected": -2.0,
|
|
"step": 41
|
|
},
|
|
{
|
|
"epoch": 0.051533742331288344,
|
|
"grad_norm": 37.85600478708204,
|
|
"learning_rate": 7.943801652892562e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3236,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 3.59375,
|
|
"rewards/rejected": -2.15625,
|
|
"step": 42
|
|
},
|
|
{
|
|
"epoch": 0.05276073619631902,
|
|
"grad_norm": 43.531972474165585,
|
|
"learning_rate": 7.940495867768595e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3117,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 43
|
|
},
|
|
{
|
|
"epoch": 0.053987730061349694,
|
|
"grad_norm": 63.54372274745974,
|
|
"learning_rate": 7.937190082644628e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.4265,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 3.625,
|
|
"rewards/rejected": -2.53125,
|
|
"step": 44
|
|
},
|
|
{
|
|
"epoch": 0.05521472392638037,
|
|
"grad_norm": 42.36111410389095,
|
|
"learning_rate": 7.933884297520661e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3513,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 3.875,
|
|
"rewards/rejected": -2.640625,
|
|
"step": 45
|
|
},
|
|
{
|
|
"epoch": 0.05644171779141104,
|
|
"grad_norm": 44.51420354088873,
|
|
"learning_rate": 7.930578512396694e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3109,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -2.609375,
|
|
"step": 46
|
|
},
|
|
{
|
|
"epoch": 0.05766871165644172,
|
|
"grad_norm": 61.22394416060924,
|
|
"learning_rate": 7.927272727272727e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.4432,
|
|
"rewards/accuracies": 0.7265625,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 3.5,
|
|
"rewards/rejected": -2.28125,
|
|
"step": 47
|
|
},
|
|
{
|
|
"epoch": 0.05889570552147239,
|
|
"grad_norm": 35.37901763029379,
|
|
"learning_rate": 7.92396694214876e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3719,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 3.203125,
|
|
"rewards/rejected": -2.09375,
|
|
"step": 48
|
|
},
|
|
{
|
|
"epoch": 0.06012269938650307,
|
|
"grad_norm": 31.606470718574474,
|
|
"learning_rate": 7.920661157024793e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2642,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.90625,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.75,
|
|
"step": 49
|
|
},
|
|
{
|
|
"epoch": 0.06134969325153374,
|
|
"grad_norm": 48.38174059914219,
|
|
"learning_rate": 7.917355371900826e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3848,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 3.8125,
|
|
"rewards/rejected": -2.359375,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.06257668711656442,
|
|
"grad_norm": 41.76003291409928,
|
|
"learning_rate": 7.914049586776859e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2936,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.640625,
|
|
"step": 51
|
|
},
|
|
{
|
|
"epoch": 0.0638036809815951,
|
|
"grad_norm": 36.258140724486836,
|
|
"learning_rate": 7.910743801652892e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3238,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 4.21875,
|
|
"rewards/rejected": -2.75,
|
|
"step": 52
|
|
},
|
|
{
|
|
"epoch": 0.06503067484662577,
|
|
"grad_norm": 38.82725433032811,
|
|
"learning_rate": 7.907438016528925e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3104,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 4.03125,
|
|
"rewards/rejected": -2.28125,
|
|
"step": 53
|
|
},
|
|
{
|
|
"epoch": 0.06625766871165645,
|
|
"grad_norm": 34.79397739539489,
|
|
"learning_rate": 7.904132231404959e-07,
|
|
"logits/chosen": -0.0284423828125,
|
|
"logits/rejected": -0.208984375,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -258.0,
|
|
"loss": 0.2634,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 3.78125,
|
|
"rewards/rejected": -2.546875,
|
|
"step": 54
|
|
},
|
|
{
|
|
"epoch": 0.06748466257668712,
|
|
"grad_norm": 41.093192913712485,
|
|
"learning_rate": 7.900826446280992e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3081,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.515625,
|
|
"step": 55
|
|
},
|
|
{
|
|
"epoch": 0.0687116564417178,
|
|
"grad_norm": 45.49364963721349,
|
|
"learning_rate": 7.897520661157025e-07,
|
|
"logits/chosen": -0.119140625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3753,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.265625,
|
|
"step": 56
|
|
},
|
|
{
|
|
"epoch": 0.06993865030674846,
|
|
"grad_norm": 50.666379541833365,
|
|
"learning_rate": 7.894214876033057e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.203125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3742,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 3.953125,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 57
|
|
},
|
|
{
|
|
"epoch": 0.07116564417177915,
|
|
"grad_norm": 47.64145593969582,
|
|
"learning_rate": 7.89090909090909e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.2177734375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3598,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 58
|
|
},
|
|
{
|
|
"epoch": 0.07239263803680981,
|
|
"grad_norm": 51.11345473280857,
|
|
"learning_rate": 7.887603305785123e-07,
|
|
"logits/chosen": -0.049560546875,
|
|
"logits/rejected": -0.140625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.4133,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 3.28125,
|
|
"rewards/rejected": -1.9921875,
|
|
"step": 59
|
|
},
|
|
{
|
|
"epoch": 0.0736196319018405,
|
|
"grad_norm": 39.0647396184868,
|
|
"learning_rate": 7.884297520661157e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3578,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 3.46875,
|
|
"rewards/rejected": -2.28125,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 0.07484662576687116,
|
|
"grad_norm": 39.12713457741509,
|
|
"learning_rate": 7.88099173553719e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3224,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.578125,
|
|
"step": 61
|
|
},
|
|
{
|
|
"epoch": 0.07607361963190185,
|
|
"grad_norm": 41.544347269900555,
|
|
"learning_rate": 7.877685950413223e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3433,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 2.328125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 62
|
|
},
|
|
{
|
|
"epoch": 0.07730061349693251,
|
|
"grad_norm": 52.28530938487766,
|
|
"learning_rate": 7.874380165289256e-07,
|
|
"logits/chosen": -0.0257568359375,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.331,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 3.9375,
|
|
"rewards/rejected": -2.40625,
|
|
"step": 63
|
|
},
|
|
{
|
|
"epoch": 0.0785276073619632,
|
|
"grad_norm": 44.343811483698694,
|
|
"learning_rate": 7.871074380165289e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3628,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -2.5,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.07975460122699386,
|
|
"grad_norm": 40.38168553746162,
|
|
"learning_rate": 7.867768595041322e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3193,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.453125,
|
|
"step": 65
|
|
},
|
|
{
|
|
"epoch": 0.08098159509202454,
|
|
"grad_norm": 35.90018196646264,
|
|
"learning_rate": 7.864462809917355e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.2892,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.0,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 66
|
|
},
|
|
{
|
|
"epoch": 0.08220858895705521,
|
|
"grad_norm": 39.83708858843296,
|
|
"learning_rate": 7.861157024793388e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.344,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 3.90625,
|
|
"rewards/rejected": -2.671875,
|
|
"step": 67
|
|
},
|
|
{
|
|
"epoch": 0.0834355828220859,
|
|
"grad_norm": 41.25508984701749,
|
|
"learning_rate": 7.857851239669421e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3597,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 3.953125,
|
|
"rewards/rejected": -2.46875,
|
|
"step": 68
|
|
},
|
|
{
|
|
"epoch": 0.08466257668711656,
|
|
"grad_norm": 35.61462182457114,
|
|
"learning_rate": 7.854545454545454e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2797,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.3125,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 69
|
|
},
|
|
{
|
|
"epoch": 0.08588957055214724,
|
|
"grad_norm": 43.949799113654095,
|
|
"learning_rate": 7.851239669421487e-07,
|
|
"logits/chosen": -0.12109375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3016,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.328125,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.34375,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 0.08711656441717791,
|
|
"grad_norm": 40.385403541291616,
|
|
"learning_rate": 7.84793388429752e-07,
|
|
"logits/chosen": -0.068359375,
|
|
"logits/rejected": -0.1494140625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3925,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 3.625,
|
|
"rewards/rejected": -2.265625,
|
|
"step": 71
|
|
},
|
|
{
|
|
"epoch": 0.08834355828220859,
|
|
"grad_norm": 42.66471807743761,
|
|
"learning_rate": 7.844628099173554e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2955,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -2.28125,
|
|
"step": 72
|
|
},
|
|
{
|
|
"epoch": 0.08957055214723926,
|
|
"grad_norm": 45.57465316917817,
|
|
"learning_rate": 7.841322314049587e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3924,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 73
|
|
},
|
|
{
|
|
"epoch": 0.09079754601226994,
|
|
"grad_norm": 41.41678309244041,
|
|
"learning_rate": 7.838016528925619e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2911,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 3.9375,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 74
|
|
},
|
|
{
|
|
"epoch": 0.09202453987730061,
|
|
"grad_norm": 36.36740264063768,
|
|
"learning_rate": 7.834710743801652e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.271,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.09325153374233129,
|
|
"grad_norm": 51.152036733559505,
|
|
"learning_rate": 7.831404958677685e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3262,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.53125,
|
|
"step": 76
|
|
},
|
|
{
|
|
"epoch": 0.09447852760736196,
|
|
"grad_norm": 34.55182641721203,
|
|
"learning_rate": 7.828099173553718e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2774,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.875,
|
|
"step": 77
|
|
},
|
|
{
|
|
"epoch": 0.09570552147239264,
|
|
"grad_norm": 40.758007342891354,
|
|
"learning_rate": 7.824793388429752e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2953,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -2.859375,
|
|
"step": 78
|
|
},
|
|
{
|
|
"epoch": 0.09693251533742331,
|
|
"grad_norm": 40.243771187807226,
|
|
"learning_rate": 7.821487603305785e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.324,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.796875,
|
|
"step": 79
|
|
},
|
|
{
|
|
"epoch": 0.09815950920245399,
|
|
"grad_norm": 37.86169713828253,
|
|
"learning_rate": 7.818181818181818e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2795,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 0.09938650306748466,
|
|
"grad_norm": 36.69542578194377,
|
|
"learning_rate": 7.814876033057851e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2988,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 81
|
|
},
|
|
{
|
|
"epoch": 0.10061349693251534,
|
|
"grad_norm": 52.318975452049024,
|
|
"learning_rate": 7.811570247933884e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3562,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 82
|
|
},
|
|
{
|
|
"epoch": 0.10184049079754601,
|
|
"grad_norm": 48.36250064527517,
|
|
"learning_rate": 7.808264462809916e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3065,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 83
|
|
},
|
|
{
|
|
"epoch": 0.10306748466257669,
|
|
"grad_norm": 56.4532578365936,
|
|
"learning_rate": 7.80495867768595e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2942,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.53125,
|
|
"step": 84
|
|
},
|
|
{
|
|
"epoch": 0.10429447852760736,
|
|
"grad_norm": 40.84421229525563,
|
|
"learning_rate": 7.801652892561983e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3021,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.484375,
|
|
"step": 85
|
|
},
|
|
{
|
|
"epoch": 0.10552147239263804,
|
|
"grad_norm": 42.8824866796569,
|
|
"learning_rate": 7.798347107438016e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3346,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 86
|
|
},
|
|
{
|
|
"epoch": 0.1067484662576687,
|
|
"grad_norm": 32.177789360920215,
|
|
"learning_rate": 7.795041322314049e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.2835,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.77734375,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 87
|
|
},
|
|
{
|
|
"epoch": 0.10797546012269939,
|
|
"grad_norm": 34.28909140331474,
|
|
"learning_rate": 7.791735537190082e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.265,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 88
|
|
},
|
|
{
|
|
"epoch": 0.10920245398773006,
|
|
"grad_norm": 42.40518360379319,
|
|
"learning_rate": 7.788429752066116e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3464,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 89
|
|
},
|
|
{
|
|
"epoch": 0.11042944785276074,
|
|
"grad_norm": 49.53932371851528,
|
|
"learning_rate": 7.785123966942149e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3253,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 0.1116564417177914,
|
|
"grad_norm": 40.48862678893552,
|
|
"learning_rate": 7.781818181818182e-07,
|
|
"logits/chosen": -0.039794921875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2845,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -3.203125,
|
|
"step": 91
|
|
},
|
|
{
|
|
"epoch": 0.11288343558282209,
|
|
"grad_norm": 44.87681376045074,
|
|
"learning_rate": 7.778512396694214e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2738,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.5,
|
|
"step": 92
|
|
},
|
|
{
|
|
"epoch": 0.11411042944785275,
|
|
"grad_norm": 36.61224632474363,
|
|
"learning_rate": 7.775206611570247e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.282,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 93
|
|
},
|
|
{
|
|
"epoch": 0.11533742331288344,
|
|
"grad_norm": 42.24461574764279,
|
|
"learning_rate": 7.77190082644628e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2813,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.234375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 94
|
|
},
|
|
{
|
|
"epoch": 0.1165644171779141,
|
|
"grad_norm": 42.13498046583409,
|
|
"learning_rate": 7.768595041322314e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2988,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 95
|
|
},
|
|
{
|
|
"epoch": 0.11779141104294479,
|
|
"grad_norm": 39.13305361039425,
|
|
"learning_rate": 7.765289256198347e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.253,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 96
|
|
},
|
|
{
|
|
"epoch": 0.11901840490797547,
|
|
"grad_norm": 46.1275059887527,
|
|
"learning_rate": 7.76198347107438e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.357,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 97
|
|
},
|
|
{
|
|
"epoch": 0.12024539877300613,
|
|
"grad_norm": 38.580559184520915,
|
|
"learning_rate": 7.758677685950413e-07,
|
|
"logits/chosen": -0.041015625,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3015,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.828125,
|
|
"step": 98
|
|
},
|
|
{
|
|
"epoch": 0.12147239263803682,
|
|
"grad_norm": 38.975117366609354,
|
|
"learning_rate": 7.755371900826446e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -244.0,
|
|
"loss": 0.2586,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.84375,
|
|
"step": 99
|
|
},
|
|
{
|
|
"epoch": 0.12269938650306748,
|
|
"grad_norm": 38.2314125667712,
|
|
"learning_rate": 7.752066115702479e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3048,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.12392638036809817,
|
|
"grad_norm": 47.72088602593496,
|
|
"learning_rate": 7.748760330578512e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3549,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.921875,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 101
|
|
},
|
|
{
|
|
"epoch": 0.12515337423312883,
|
|
"grad_norm": 39.20874470259191,
|
|
"learning_rate": 7.745454545454545e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3073,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.71875,
|
|
"step": 102
|
|
},
|
|
{
|
|
"epoch": 0.1263803680981595,
|
|
"grad_norm": 36.963577318972284,
|
|
"learning_rate": 7.742148760330578e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2878,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.75,
|
|
"step": 103
|
|
},
|
|
{
|
|
"epoch": 0.1276073619631902,
|
|
"grad_norm": 50.89003123288155,
|
|
"learning_rate": 7.738842975206611e-07,
|
|
"logits/chosen": -0.107421875,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3679,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -2.71875,
|
|
"step": 104
|
|
},
|
|
{
|
|
"epoch": 0.12883435582822086,
|
|
"grad_norm": 34.19259318441508,
|
|
"learning_rate": 7.735537190082644e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2586,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.984375,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 105
|
|
},
|
|
{
|
|
"epoch": 0.13006134969325153,
|
|
"grad_norm": 38.48374397973381,
|
|
"learning_rate": 7.732231404958677e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3111,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -2.828125,
|
|
"step": 106
|
|
},
|
|
{
|
|
"epoch": 0.1312883435582822,
|
|
"grad_norm": 41.75637341421824,
|
|
"learning_rate": 7.728925619834711e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2869,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.84375,
|
|
"step": 107
|
|
},
|
|
{
|
|
"epoch": 0.1325153374233129,
|
|
"grad_norm": 33.954750583294235,
|
|
"learning_rate": 7.725619834710744e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2625,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 108
|
|
},
|
|
{
|
|
"epoch": 0.13374233128834356,
|
|
"grad_norm": 36.44619320062165,
|
|
"learning_rate": 7.722314049586777e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2958,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.6875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 109
|
|
},
|
|
{
|
|
"epoch": 0.13496932515337423,
|
|
"grad_norm": 41.465536696868014,
|
|
"learning_rate": 7.719008264462809e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3213,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 2.234375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.015625,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 0.1361963190184049,
|
|
"grad_norm": 34.97820982991418,
|
|
"learning_rate": 7.715702479338842e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.2777,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -2.75,
|
|
"step": 111
|
|
},
|
|
{
|
|
"epoch": 0.1374233128834356,
|
|
"grad_norm": 37.33233370442031,
|
|
"learning_rate": 7.712396694214875e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.2729,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.125,
|
|
"step": 112
|
|
},
|
|
{
|
|
"epoch": 0.13865030674846626,
|
|
"grad_norm": 39.32831836749558,
|
|
"learning_rate": 7.709090909090909e-07,
|
|
"logits/chosen": -0.1044921875,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3238,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 113
|
|
},
|
|
{
|
|
"epoch": 0.13987730061349693,
|
|
"grad_norm": 43.77160490529073,
|
|
"learning_rate": 7.705785123966942e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3356,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 114
|
|
},
|
|
{
|
|
"epoch": 0.1411042944785276,
|
|
"grad_norm": 45.86173030028754,
|
|
"learning_rate": 7.702479338842975e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3666,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 4.0,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 115
|
|
},
|
|
{
|
|
"epoch": 0.1423312883435583,
|
|
"grad_norm": 31.833834890400677,
|
|
"learning_rate": 7.699173553719008e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2561,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 116
|
|
},
|
|
{
|
|
"epoch": 0.14355828220858896,
|
|
"grad_norm": 34.65342816309822,
|
|
"learning_rate": 7.695867768595041e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.2802,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.453125,
|
|
"step": 117
|
|
},
|
|
{
|
|
"epoch": 0.14478527607361963,
|
|
"grad_norm": 40.64680750199124,
|
|
"learning_rate": 7.692561983471075e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3232,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 118
|
|
},
|
|
{
|
|
"epoch": 0.1460122699386503,
|
|
"grad_norm": 39.06542225750342,
|
|
"learning_rate": 7.689256198347107e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3009,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.34375,
|
|
"step": 119
|
|
},
|
|
{
|
|
"epoch": 0.147239263803681,
|
|
"grad_norm": 45.793161618342445,
|
|
"learning_rate": 7.68595041322314e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2967,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 2.421875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 0.14846625766871166,
|
|
"grad_norm": 36.35506762175528,
|
|
"learning_rate": 7.682644628099173e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2827,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.265625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 121
|
|
},
|
|
{
|
|
"epoch": 0.14969325153374233,
|
|
"grad_norm": 43.61172830028863,
|
|
"learning_rate": 7.679338842975206e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2954,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.015625,
|
|
"step": 122
|
|
},
|
|
{
|
|
"epoch": 0.150920245398773,
|
|
"grad_norm": 44.76510892093205,
|
|
"learning_rate": 7.676033057851239e-07,
|
|
"logits/chosen": -0.054443359375,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3266,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 123
|
|
},
|
|
{
|
|
"epoch": 0.1521472392638037,
|
|
"grad_norm": 34.74558330134135,
|
|
"learning_rate": 7.672727272727273e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.26,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 124
|
|
},
|
|
{
|
|
"epoch": 0.15337423312883436,
|
|
"grad_norm": 42.40382525839625,
|
|
"learning_rate": 7.669421487603306e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2816,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.15460122699386503,
|
|
"grad_norm": 54.09847924655591,
|
|
"learning_rate": 7.666115702479339e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3503,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.34375,
|
|
"step": 126
|
|
},
|
|
{
|
|
"epoch": 0.1558282208588957,
|
|
"grad_norm": 42.126209906017394,
|
|
"learning_rate": 7.662809917355372e-07,
|
|
"logits/chosen": -0.09375,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3524,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.96875,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.15625,
|
|
"step": 127
|
|
},
|
|
{
|
|
"epoch": 0.1570552147239264,
|
|
"grad_norm": 41.454801423663085,
|
|
"learning_rate": 7.659504132231404e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2841,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.15828220858895706,
|
|
"grad_norm": 37.48031578456685,
|
|
"learning_rate": 7.656198347107437e-07,
|
|
"logits/chosen": -0.0478515625,
|
|
"logits/rejected": -0.1796875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3245,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.0,
|
|
"step": 129
|
|
},
|
|
{
|
|
"epoch": 0.15950920245398773,
|
|
"grad_norm": 38.14450880210714,
|
|
"learning_rate": 7.65289256198347e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3273,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.421875,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 0.1607361963190184,
|
|
"grad_norm": 35.19502262660919,
|
|
"learning_rate": 7.649586776859504e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2106,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 131
|
|
},
|
|
{
|
|
"epoch": 0.1619631901840491,
|
|
"grad_norm": 42.07184086801236,
|
|
"learning_rate": 7.646280991735537e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3252,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 132
|
|
},
|
|
{
|
|
"epoch": 0.16319018404907976,
|
|
"grad_norm": 43.059635351413924,
|
|
"learning_rate": 7.64297520661157e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2872,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 133
|
|
},
|
|
{
|
|
"epoch": 0.16441717791411042,
|
|
"grad_norm": 48.09563853008669,
|
|
"learning_rate": 7.639669421487603e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2946,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 134
|
|
},
|
|
{
|
|
"epoch": 0.1656441717791411,
|
|
"grad_norm": 46.36508764131004,
|
|
"learning_rate": 7.636363636363636e-07,
|
|
"logits/chosen": -0.05712890625,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3654,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 135
|
|
},
|
|
{
|
|
"epoch": 0.1668711656441718,
|
|
"grad_norm": 39.349987951574626,
|
|
"learning_rate": 7.63305785123967e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3195,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 136
|
|
},
|
|
{
|
|
"epoch": 0.16809815950920245,
|
|
"grad_norm": 46.78307320347251,
|
|
"learning_rate": 7.629752066115702e-07,
|
|
"logits/chosen": -0.061279296875,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3621,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -3.078125,
|
|
"step": 137
|
|
},
|
|
{
|
|
"epoch": 0.16932515337423312,
|
|
"grad_norm": 37.18351460769713,
|
|
"learning_rate": 7.626446280991735e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3022,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 138
|
|
},
|
|
{
|
|
"epoch": 0.1705521472392638,
|
|
"grad_norm": 42.545002180919454,
|
|
"learning_rate": 7.623140495867768e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.352,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 139
|
|
},
|
|
{
|
|
"epoch": 0.17177914110429449,
|
|
"grad_norm": 47.45496980791103,
|
|
"learning_rate": 7.619834710743801e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3166,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 0.17300613496932515,
|
|
"grad_norm": 34.622965144928166,
|
|
"learning_rate": 7.616528925619834e-07,
|
|
"logits/chosen": -0.0849609375,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.251,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 141
|
|
},
|
|
{
|
|
"epoch": 0.17423312883435582,
|
|
"grad_norm": 40.37071693879945,
|
|
"learning_rate": 7.613223140495868e-07,
|
|
"logits/chosen": -0.12060546875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3233,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 142
|
|
},
|
|
{
|
|
"epoch": 0.1754601226993865,
|
|
"grad_norm": 41.48963188375289,
|
|
"learning_rate": 7.609917355371901e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2941,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 143
|
|
},
|
|
{
|
|
"epoch": 0.17668711656441718,
|
|
"grad_norm": 40.697834722668915,
|
|
"learning_rate": 7.606611570247934e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2905,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.96875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 144
|
|
},
|
|
{
|
|
"epoch": 0.17791411042944785,
|
|
"grad_norm": 39.56968971060715,
|
|
"learning_rate": 7.603305785123967e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2715,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 145
|
|
},
|
|
{
|
|
"epoch": 0.17914110429447852,
|
|
"grad_norm": 34.52865223755899,
|
|
"learning_rate": 7.599999999999999e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2436,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 146
|
|
},
|
|
{
|
|
"epoch": 0.18036809815950922,
|
|
"grad_norm": 35.86987627703754,
|
|
"learning_rate": 7.596694214876032e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2741,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 147
|
|
},
|
|
{
|
|
"epoch": 0.18159509202453988,
|
|
"grad_norm": 49.681572441729514,
|
|
"learning_rate": 7.593388429752066e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3761,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.625,
|
|
"step": 148
|
|
},
|
|
{
|
|
"epoch": 0.18282208588957055,
|
|
"grad_norm": 43.932202256712266,
|
|
"learning_rate": 7.590082644628099e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2534,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.984375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 149
|
|
},
|
|
{
|
|
"epoch": 0.18404907975460122,
|
|
"grad_norm": 39.402366627439726,
|
|
"learning_rate": 7.586776859504132e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3158,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.18527607361963191,
|
|
"grad_norm": 41.79252221918161,
|
|
"learning_rate": 7.583471074380165e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3211,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 151
|
|
},
|
|
{
|
|
"epoch": 0.18650306748466258,
|
|
"grad_norm": 35.715508495832665,
|
|
"learning_rate": 7.580165289256198e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2622,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 152
|
|
},
|
|
{
|
|
"epoch": 0.18773006134969325,
|
|
"grad_norm": 36.417969122212845,
|
|
"learning_rate": 7.576859504132231e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2871,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 153
|
|
},
|
|
{
|
|
"epoch": 0.18895705521472392,
|
|
"grad_norm": 45.428337431051204,
|
|
"learning_rate": 7.573553719008265e-07,
|
|
"logits/chosen": -0.09716796875,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3165,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 154
|
|
},
|
|
{
|
|
"epoch": 0.1901840490797546,
|
|
"grad_norm": 28.34310543484037,
|
|
"learning_rate": 7.570247933884297e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.1864,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 2.5625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 155
|
|
},
|
|
{
|
|
"epoch": 0.19141104294478528,
|
|
"grad_norm": 43.05423041815277,
|
|
"learning_rate": 7.56694214876033e-07,
|
|
"logits/chosen": -0.11865234375,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2995,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 156
|
|
},
|
|
{
|
|
"epoch": 0.19263803680981595,
|
|
"grad_norm": 35.31331641855709,
|
|
"learning_rate": 7.563636363636363e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 157
|
|
},
|
|
{
|
|
"epoch": 0.19386503067484662,
|
|
"grad_norm": 32.52885612180771,
|
|
"learning_rate": 7.560330578512396e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2615,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 158
|
|
},
|
|
{
|
|
"epoch": 0.1950920245398773,
|
|
"grad_norm": 40.74250145862389,
|
|
"learning_rate": 7.55702479338843e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.312,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.25,
|
|
"step": 159
|
|
},
|
|
{
|
|
"epoch": 0.19631901840490798,
|
|
"grad_norm": 32.03909927740134,
|
|
"learning_rate": 7.553719008264463e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2149,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.765625,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 0.19754601226993865,
|
|
"grad_norm": 33.48229976640365,
|
|
"learning_rate": 7.550413223140496e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2313,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.8984375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 161
|
|
},
|
|
{
|
|
"epoch": 0.19877300613496932,
|
|
"grad_norm": 44.9989690142436,
|
|
"learning_rate": 7.547107438016529e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3547,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.390625,
|
|
"step": 162
|
|
},
|
|
{
|
|
"epoch": 0.2,
|
|
"grad_norm": 45.392211600179095,
|
|
"learning_rate": 7.543801652892561e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3203,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.4375,
|
|
"step": 163
|
|
},
|
|
{
|
|
"epoch": 0.20122699386503068,
|
|
"grad_norm": 30.10144812491185,
|
|
"learning_rate": 7.540495867768594e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2205,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 164
|
|
},
|
|
{
|
|
"epoch": 0.20245398773006135,
|
|
"grad_norm": 41.70055038557836,
|
|
"learning_rate": 7.537190082644627e-07,
|
|
"logits/chosen": -0.078125,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2684,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.484375,
|
|
"step": 165
|
|
},
|
|
{
|
|
"epoch": 0.20368098159509201,
|
|
"grad_norm": 40.19134249676154,
|
|
"learning_rate": 7.533884297520661e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3382,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 166
|
|
},
|
|
{
|
|
"epoch": 0.2049079754601227,
|
|
"grad_norm": 40.73221481198693,
|
|
"learning_rate": 7.530578512396694e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3475,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 2.21875,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 167
|
|
},
|
|
{
|
|
"epoch": 0.20613496932515338,
|
|
"grad_norm": 36.56704570459498,
|
|
"learning_rate": 7.527272727272727e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.2594,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.0,
|
|
"step": 168
|
|
},
|
|
{
|
|
"epoch": 0.20736196319018405,
|
|
"grad_norm": 36.86789794879476,
|
|
"learning_rate": 7.52396694214876e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2793,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.75,
|
|
"step": 169
|
|
},
|
|
{
|
|
"epoch": 0.2085889570552147,
|
|
"grad_norm": 43.700977925857806,
|
|
"learning_rate": 7.520661157024793e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2994,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.765625,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 0.2098159509202454,
|
|
"grad_norm": 33.49786387415663,
|
|
"learning_rate": 7.517355371900827e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2804,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 171
|
|
},
|
|
{
|
|
"epoch": 0.21104294478527608,
|
|
"grad_norm": 36.775839118353844,
|
|
"learning_rate": 7.514049586776859e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2514,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 172
|
|
},
|
|
{
|
|
"epoch": 0.21226993865030674,
|
|
"grad_norm": 29.010747243935523,
|
|
"learning_rate": 7.510743801652892e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2075,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.296875,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 173
|
|
},
|
|
{
|
|
"epoch": 0.2134969325153374,
|
|
"grad_norm": 42.10282379400345,
|
|
"learning_rate": 7.507438016528925e-07,
|
|
"logits/chosen": -0.035400390625,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3371,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 174
|
|
},
|
|
{
|
|
"epoch": 0.2147239263803681,
|
|
"grad_norm": 44.4053310320228,
|
|
"learning_rate": 7.504132231404958e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3132,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 0.21595092024539878,
|
|
"grad_norm": 38.79125305740806,
|
|
"learning_rate": 7.500826446280991e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.315,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -3.875,
|
|
"step": 176
|
|
},
|
|
{
|
|
"epoch": 0.21717791411042944,
|
|
"grad_norm": 29.08349894872895,
|
|
"learning_rate": 7.497520661157025e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2249,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.453125,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 177
|
|
},
|
|
{
|
|
"epoch": 0.2184049079754601,
|
|
"grad_norm": 39.84454997153339,
|
|
"learning_rate": 7.494214876033058e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3147,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 178
|
|
},
|
|
{
|
|
"epoch": 0.2196319018404908,
|
|
"grad_norm": 32.49552053062871,
|
|
"learning_rate": 7.490909090909091e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2585,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 179
|
|
},
|
|
{
|
|
"epoch": 0.22085889570552147,
|
|
"grad_norm": 41.53796098392134,
|
|
"learning_rate": 7.487603305785124e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3355,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.75,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 0.22208588957055214,
|
|
"grad_norm": 43.27113784190138,
|
|
"learning_rate": 7.484297520661156e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.3558,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 181
|
|
},
|
|
{
|
|
"epoch": 0.2233128834355828,
|
|
"grad_norm": 40.060092182372635,
|
|
"learning_rate": 7.480991735537189e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3329,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 182
|
|
},
|
|
{
|
|
"epoch": 0.2245398773006135,
|
|
"grad_norm": 42.893284410160845,
|
|
"learning_rate": 7.477685950413223e-07,
|
|
"logits/chosen": -0.0888671875,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.3303,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 183
|
|
},
|
|
{
|
|
"epoch": 0.22576687116564417,
|
|
"grad_norm": 44.81738203832481,
|
|
"learning_rate": 7.474380165289256e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3382,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.77734375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 184
|
|
},
|
|
{
|
|
"epoch": 0.22699386503067484,
|
|
"grad_norm": 49.18179746490089,
|
|
"learning_rate": 7.471074380165289e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3425,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 185
|
|
},
|
|
{
|
|
"epoch": 0.2282208588957055,
|
|
"grad_norm": 38.62632364463619,
|
|
"learning_rate": 7.467768595041322e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3061,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 186
|
|
},
|
|
{
|
|
"epoch": 0.2294478527607362,
|
|
"grad_norm": 37.48238102625356,
|
|
"learning_rate": 7.464462809917355e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2719,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 187
|
|
},
|
|
{
|
|
"epoch": 0.23067484662576687,
|
|
"grad_norm": 33.9965710800535,
|
|
"learning_rate": 7.461157024793388e-07,
|
|
"logits/chosen": -0.0712890625,
|
|
"logits/rejected": -0.1689453125,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2908,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 188
|
|
},
|
|
{
|
|
"epoch": 0.23190184049079754,
|
|
"grad_norm": 32.6251624460159,
|
|
"learning_rate": 7.457851239669422e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2196,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 189
|
|
},
|
|
{
|
|
"epoch": 0.2331288343558282,
|
|
"grad_norm": 33.86427029616883,
|
|
"learning_rate": 7.454545454545454e-07,
|
|
"logits/chosen": -0.05810546875,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2859,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.875,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 0.2343558282208589,
|
|
"grad_norm": 48.95754374693823,
|
|
"learning_rate": 7.451239669421487e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3495,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 191
|
|
},
|
|
{
|
|
"epoch": 0.23558282208588957,
|
|
"grad_norm": 44.694192367882906,
|
|
"learning_rate": 7.44793388429752e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3014,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.578125,
|
|
"step": 192
|
|
},
|
|
{
|
|
"epoch": 0.23680981595092024,
|
|
"grad_norm": 34.25684662561766,
|
|
"learning_rate": 7.444628099173553e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2177,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 193
|
|
},
|
|
{
|
|
"epoch": 0.23803680981595093,
|
|
"grad_norm": 39.119509435238996,
|
|
"learning_rate": 7.441322314049586e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2953,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 194
|
|
},
|
|
{
|
|
"epoch": 0.2392638036809816,
|
|
"grad_norm": 34.817335787332766,
|
|
"learning_rate": 7.43801652892562e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2301,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 195
|
|
},
|
|
{
|
|
"epoch": 0.24049079754601227,
|
|
"grad_norm": 40.102802913255154,
|
|
"learning_rate": 7.434710743801653e-07,
|
|
"logits/chosen": -0.09130859375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.297,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 196
|
|
},
|
|
{
|
|
"epoch": 0.24171779141104294,
|
|
"grad_norm": 34.798967486094014,
|
|
"learning_rate": 7.431404958677686e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2803,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 197
|
|
},
|
|
{
|
|
"epoch": 0.24294478527607363,
|
|
"grad_norm": 39.54376141056963,
|
|
"learning_rate": 7.428099173553719e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -258.0,
|
|
"loss": 0.3347,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.671875,
|
|
"step": 198
|
|
},
|
|
{
|
|
"epoch": 0.2441717791411043,
|
|
"grad_norm": 29.961664099131877,
|
|
"learning_rate": 7.424793388429751e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.1903,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 2.359375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 199
|
|
},
|
|
{
|
|
"epoch": 0.24539877300613497,
|
|
"grad_norm": 41.86791512110987,
|
|
"learning_rate": 7.421487603305784e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3056,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.24662576687116564,
|
|
"grad_norm": 32.0817049971399,
|
|
"learning_rate": 7.418181818181818e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2166,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.0625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 201
|
|
},
|
|
{
|
|
"epoch": 0.24785276073619633,
|
|
"grad_norm": 35.7245529395789,
|
|
"learning_rate": 7.414876033057851e-07,
|
|
"logits/chosen": -0.12060546875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2641,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 202
|
|
},
|
|
{
|
|
"epoch": 0.249079754601227,
|
|
"grad_norm": 42.27175209208064,
|
|
"learning_rate": 7.411570247933884e-07,
|
|
"logits/chosen": -0.08935546875,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4055,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8828125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 203
|
|
},
|
|
{
|
|
"epoch": 0.25030674846625767,
|
|
"grad_norm": 39.50994905620242,
|
|
"learning_rate": 7.408264462809917e-07,
|
|
"logits/chosen": -0.0771484375,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.3395,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -3.671875,
|
|
"step": 204
|
|
},
|
|
{
|
|
"epoch": 0.25153374233128833,
|
|
"grad_norm": 37.49332179023069,
|
|
"learning_rate": 7.40495867768595e-07,
|
|
"logits/chosen": -0.02880859375,
|
|
"logits/rejected": -0.173828125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3209,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.265625,
|
|
"step": 205
|
|
},
|
|
{
|
|
"epoch": 0.252760736196319,
|
|
"grad_norm": 27.399492756017406,
|
|
"learning_rate": 7.401652892561984e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2051,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 206
|
|
},
|
|
{
|
|
"epoch": 0.25398773006134967,
|
|
"grad_norm": 43.41671912422527,
|
|
"learning_rate": 7.398347107438017e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3584,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 207
|
|
},
|
|
{
|
|
"epoch": 0.2552147239263804,
|
|
"grad_norm": 38.06892744188352,
|
|
"learning_rate": 7.395041322314049e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3179,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 208
|
|
},
|
|
{
|
|
"epoch": 0.25644171779141106,
|
|
"grad_norm": 38.21598347000612,
|
|
"learning_rate": 7.391735537190082e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.306,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 209
|
|
},
|
|
{
|
|
"epoch": 0.25766871165644173,
|
|
"grad_norm": 45.46016539899385,
|
|
"learning_rate": 7.388429752066115e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3476,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 0.2588957055214724,
|
|
"grad_norm": 38.3264065355804,
|
|
"learning_rate": 7.385123966942148e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3035,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 211
|
|
},
|
|
{
|
|
"epoch": 0.26012269938650306,
|
|
"grad_norm": 37.36895578645757,
|
|
"learning_rate": 7.381818181818182e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2805,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 212
|
|
},
|
|
{
|
|
"epoch": 0.26134969325153373,
|
|
"grad_norm": 33.79330367194955,
|
|
"learning_rate": 7.378512396694215e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2588,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 213
|
|
},
|
|
{
|
|
"epoch": 0.2625766871165644,
|
|
"grad_norm": 41.91194027673872,
|
|
"learning_rate": 7.375206611570248e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3085,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 214
|
|
},
|
|
{
|
|
"epoch": 0.26380368098159507,
|
|
"grad_norm": 42.67216156371333,
|
|
"learning_rate": 7.371900826446281e-07,
|
|
"logits/chosen": -0.0927734375,
|
|
"logits/rejected": -0.1953125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3066,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 215
|
|
},
|
|
{
|
|
"epoch": 0.2650306748466258,
|
|
"grad_norm": 41.6917127385005,
|
|
"learning_rate": 7.368595041322314e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.93359375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 216
|
|
},
|
|
{
|
|
"epoch": 0.26625766871165646,
|
|
"grad_norm": 41.999916191643734,
|
|
"learning_rate": 7.365289256198346e-07,
|
|
"logits/chosen": -0.05322265625,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3621,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.421875,
|
|
"step": 217
|
|
},
|
|
{
|
|
"epoch": 0.2674846625766871,
|
|
"grad_norm": 48.35045291177634,
|
|
"learning_rate": 7.36198347107438e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.21875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3724,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 218
|
|
},
|
|
{
|
|
"epoch": 0.2687116564417178,
|
|
"grad_norm": 36.99776737120032,
|
|
"learning_rate": 7.358677685950413e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3041,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.875,
|
|
"step": 219
|
|
},
|
|
{
|
|
"epoch": 0.26993865030674846,
|
|
"grad_norm": 49.4734461720144,
|
|
"learning_rate": 7.355371900826446e-07,
|
|
"logits/chosen": -0.05078125,
|
|
"logits/rejected": -0.1318359375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3781,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -3.421875,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 0.27116564417177913,
|
|
"grad_norm": 50.14968062203286,
|
|
"learning_rate": 7.352066115702479e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3828,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.796875,
|
|
"step": 221
|
|
},
|
|
{
|
|
"epoch": 0.2723926380368098,
|
|
"grad_norm": 35.95574239692987,
|
|
"learning_rate": 7.348760330578512e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2649,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 222
|
|
},
|
|
{
|
|
"epoch": 0.27361963190184047,
|
|
"grad_norm": 40.443295860960646,
|
|
"learning_rate": 7.345454545454545e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2888,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.75,
|
|
"step": 223
|
|
},
|
|
{
|
|
"epoch": 0.2748466257668712,
|
|
"grad_norm": 42.19474227390439,
|
|
"learning_rate": 7.342148760330579e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.288,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 224
|
|
},
|
|
{
|
|
"epoch": 0.27607361963190186,
|
|
"grad_norm": 37.9965018970918,
|
|
"learning_rate": 7.338842975206612e-07,
|
|
"logits/chosen": -0.10546875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.261,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 0.2773006134969325,
|
|
"grad_norm": 38.80764018954137,
|
|
"learning_rate": 7.335537190082644e-07,
|
|
"logits/chosen": -0.119140625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3311,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 226
|
|
},
|
|
{
|
|
"epoch": 0.2785276073619632,
|
|
"grad_norm": 42.55177823975192,
|
|
"learning_rate": 7.332231404958677e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3198,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 227
|
|
},
|
|
{
|
|
"epoch": 0.27975460122699386,
|
|
"grad_norm": 47.83061041994659,
|
|
"learning_rate": 7.32892561983471e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3258,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 228
|
|
},
|
|
{
|
|
"epoch": 0.2809815950920245,
|
|
"grad_norm": 36.471929966124165,
|
|
"learning_rate": 7.325619834710743e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2558,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 229
|
|
},
|
|
{
|
|
"epoch": 0.2822085889570552,
|
|
"grad_norm": 36.373451110997884,
|
|
"learning_rate": 7.322314049586777e-07,
|
|
"logits/chosen": -0.06884765625,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2454,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 0.28343558282208586,
|
|
"grad_norm": 44.837599338102365,
|
|
"learning_rate": 7.31900826446281e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3895,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.625,
|
|
"step": 231
|
|
},
|
|
{
|
|
"epoch": 0.2846625766871166,
|
|
"grad_norm": 40.26904672106654,
|
|
"learning_rate": 7.315702479338843e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3097,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 232
|
|
},
|
|
{
|
|
"epoch": 0.28588957055214725,
|
|
"grad_norm": 37.854698103189165,
|
|
"learning_rate": 7.312396694214876e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3055,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 233
|
|
},
|
|
{
|
|
"epoch": 0.2871165644171779,
|
|
"grad_norm": 38.018318477115464,
|
|
"learning_rate": 7.309090909090909e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2977,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 234
|
|
},
|
|
{
|
|
"epoch": 0.2883435582822086,
|
|
"grad_norm": 37.47355891059506,
|
|
"learning_rate": 7.305785123966941e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.284,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 235
|
|
},
|
|
{
|
|
"epoch": 0.28957055214723926,
|
|
"grad_norm": 53.75756263750908,
|
|
"learning_rate": 7.302479338842975e-07,
|
|
"logits/chosen": -0.0238037109375,
|
|
"logits/rejected": -0.2080078125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.4235,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 236
|
|
},
|
|
{
|
|
"epoch": 0.2907975460122699,
|
|
"grad_norm": 34.35888091553625,
|
|
"learning_rate": 7.299173553719008e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2606,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 237
|
|
},
|
|
{
|
|
"epoch": 0.2920245398773006,
|
|
"grad_norm": 42.88801905010467,
|
|
"learning_rate": 7.295867768595041e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3538,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 238
|
|
},
|
|
{
|
|
"epoch": 0.29325153374233126,
|
|
"grad_norm": 42.29179859412283,
|
|
"learning_rate": 7.292561983471074e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3291,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 239
|
|
},
|
|
{
|
|
"epoch": 0.294478527607362,
|
|
"grad_norm": 37.481940316076816,
|
|
"learning_rate": 7.289256198347107e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2328,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 0.29570552147239265,
|
|
"grad_norm": 37.13709897048156,
|
|
"learning_rate": 7.28595041322314e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2698,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 241
|
|
},
|
|
{
|
|
"epoch": 0.2969325153374233,
|
|
"grad_norm": 30.985673644793227,
|
|
"learning_rate": 7.282644628099174e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.228,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 242
|
|
},
|
|
{
|
|
"epoch": 0.298159509202454,
|
|
"grad_norm": 53.9279211325065,
|
|
"learning_rate": 7.279338842975206e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3208,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 243
|
|
},
|
|
{
|
|
"epoch": 0.29938650306748466,
|
|
"grad_norm": 44.157193824232536,
|
|
"learning_rate": 7.276033057851239e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.4473,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 244
|
|
},
|
|
{
|
|
"epoch": 0.3006134969325153,
|
|
"grad_norm": 36.965116220324916,
|
|
"learning_rate": 7.272727272727272e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2506,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 245
|
|
},
|
|
{
|
|
"epoch": 0.301840490797546,
|
|
"grad_norm": 35.79905789258026,
|
|
"learning_rate": 7.269421487603305e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.292,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 246
|
|
},
|
|
{
|
|
"epoch": 0.3030674846625767,
|
|
"grad_norm": 46.030369589419166,
|
|
"learning_rate": 7.266115702479338e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3237,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 247
|
|
},
|
|
{
|
|
"epoch": 0.3042944785276074,
|
|
"grad_norm": 38.420936945371935,
|
|
"learning_rate": 7.262809917355372e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2497,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 248
|
|
},
|
|
{
|
|
"epoch": 0.30552147239263805,
|
|
"grad_norm": 42.67532255005307,
|
|
"learning_rate": 7.259504132231405e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -476.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2862,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.328125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 249
|
|
},
|
|
{
|
|
"epoch": 0.3067484662576687,
|
|
"grad_norm": 40.02525959257945,
|
|
"learning_rate": 7.256198347107438e-07,
|
|
"logits/chosen": -0.34375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3299,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.3079754601226994,
|
|
"grad_norm": 43.266430285019,
|
|
"learning_rate": 7.252892561983471e-07,
|
|
"logits/chosen": -0.041015625,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3567,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.75,
|
|
"step": 251
|
|
},
|
|
{
|
|
"epoch": 0.30920245398773005,
|
|
"grad_norm": 37.89906733852128,
|
|
"learning_rate": 7.249586776859503e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2982,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 252
|
|
},
|
|
{
|
|
"epoch": 0.3104294478527607,
|
|
"grad_norm": 33.28018552295872,
|
|
"learning_rate": 7.246280991735536e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2522,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 253
|
|
},
|
|
{
|
|
"epoch": 0.3116564417177914,
|
|
"grad_norm": 42.254563825083,
|
|
"learning_rate": 7.24297520661157e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3681,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 254
|
|
},
|
|
{
|
|
"epoch": 0.3128834355828221,
|
|
"grad_norm": 39.895653943663504,
|
|
"learning_rate": 7.239669421487603e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3481,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.8359375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 255
|
|
},
|
|
{
|
|
"epoch": 0.3141104294478528,
|
|
"grad_norm": 37.11298367671905,
|
|
"learning_rate": 7.236363636363636e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2374,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.34375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 0.31533742331288345,
|
|
"grad_norm": 44.74542880830349,
|
|
"learning_rate": 7.233057851239669e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3118,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 257
|
|
},
|
|
{
|
|
"epoch": 0.3165644171779141,
|
|
"grad_norm": 34.05504946265699,
|
|
"learning_rate": 7.229752066115702e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2467,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.8984375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 258
|
|
},
|
|
{
|
|
"epoch": 0.3177914110429448,
|
|
"grad_norm": 35.06646898219457,
|
|
"learning_rate": 7.226446280991736e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.278,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 259
|
|
},
|
|
{
|
|
"epoch": 0.31901840490797545,
|
|
"grad_norm": 48.15030024240453,
|
|
"learning_rate": 7.223140495867769e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3099,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 0.3202453987730061,
|
|
"grad_norm": 43.183782239877885,
|
|
"learning_rate": 7.219834710743801e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3207,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 261
|
|
},
|
|
{
|
|
"epoch": 0.3214723926380368,
|
|
"grad_norm": 43.81473087786423,
|
|
"learning_rate": 7.216528925619834e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.484375,
|
|
"step": 262
|
|
},
|
|
{
|
|
"epoch": 0.3226993865030675,
|
|
"grad_norm": 45.763039013247074,
|
|
"learning_rate": 7.213223140495867e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.345,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -3.265625,
|
|
"step": 263
|
|
},
|
|
{
|
|
"epoch": 0.3239263803680982,
|
|
"grad_norm": 41.54807658858699,
|
|
"learning_rate": 7.2099173553719e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3193,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 264
|
|
},
|
|
{
|
|
"epoch": 0.32515337423312884,
|
|
"grad_norm": 43.72770395576322,
|
|
"learning_rate": 7.206611570247934e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.296,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 265
|
|
},
|
|
{
|
|
"epoch": 0.3263803680981595,
|
|
"grad_norm": 49.68472762312363,
|
|
"learning_rate": 7.203305785123967e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.41,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.68359375,
|
|
"rewards/margins": 3.875,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 266
|
|
},
|
|
{
|
|
"epoch": 0.3276073619631902,
|
|
"grad_norm": 54.21686609921306,
|
|
"learning_rate": 7.2e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3765,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 267
|
|
},
|
|
{
|
|
"epoch": 0.32883435582822085,
|
|
"grad_norm": 36.013473849533085,
|
|
"learning_rate": 7.196694214876033e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2353,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 268
|
|
},
|
|
{
|
|
"epoch": 0.3300613496932515,
|
|
"grad_norm": 37.07985683341205,
|
|
"learning_rate": 7.193388429752066e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.5,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.229,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 269
|
|
},
|
|
{
|
|
"epoch": 0.3312883435582822,
|
|
"grad_norm": 43.39229519288019,
|
|
"learning_rate": 7.190082644628098e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3693,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.875,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 0.3325153374233129,
|
|
"grad_norm": 33.45853348793219,
|
|
"learning_rate": 7.186776859504132e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2511,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.34375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 271
|
|
},
|
|
{
|
|
"epoch": 0.3337423312883436,
|
|
"grad_norm": 36.86544969536763,
|
|
"learning_rate": 7.183471074380165e-07,
|
|
"logits/chosen": -0.1025390625,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2449,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 272
|
|
},
|
|
{
|
|
"epoch": 0.33496932515337424,
|
|
"grad_norm": 45.60464590137911,
|
|
"learning_rate": 7.180165289256198e-07,
|
|
"logits/chosen": -0.05029296875,
|
|
"logits/rejected": -0.1357421875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.4336,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.4375,
|
|
"step": 273
|
|
},
|
|
{
|
|
"epoch": 0.3361963190184049,
|
|
"grad_norm": 38.72003382453235,
|
|
"learning_rate": 7.176859504132231e-07,
|
|
"logits/chosen": -0.06591796875,
|
|
"logits/rejected": -0.21875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3223,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 274
|
|
},
|
|
{
|
|
"epoch": 0.3374233128834356,
|
|
"grad_norm": 41.46753745108232,
|
|
"learning_rate": 7.173553719008264e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3348,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.5,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 0.33865030674846625,
|
|
"grad_norm": 29.942801211334718,
|
|
"learning_rate": 7.170247933884297e-07,
|
|
"logits/chosen": -0.056396484375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2861,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 276
|
|
},
|
|
{
|
|
"epoch": 0.3398773006134969,
|
|
"grad_norm": 36.649252691867474,
|
|
"learning_rate": 7.166942148760331e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2238,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 277
|
|
},
|
|
{
|
|
"epoch": 0.3411042944785276,
|
|
"grad_norm": 41.64219712301306,
|
|
"learning_rate": 7.163636363636364e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2881,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 278
|
|
},
|
|
{
|
|
"epoch": 0.3423312883435583,
|
|
"grad_norm": 25.028100153022006,
|
|
"learning_rate": 7.160330578512396e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.1819,
|
|
"rewards/accuracies": 0.9453125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.875,
|
|
"step": 279
|
|
},
|
|
{
|
|
"epoch": 0.34355828220858897,
|
|
"grad_norm": 43.34524582256192,
|
|
"learning_rate": 7.157024793388429e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3811,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.625,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 0.34478527607361964,
|
|
"grad_norm": 39.57499610045821,
|
|
"learning_rate": 7.153719008264462e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3441,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 281
|
|
},
|
|
{
|
|
"epoch": 0.3460122699386503,
|
|
"grad_norm": 32.558914994483004,
|
|
"learning_rate": 7.150413223140495e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2162,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 282
|
|
},
|
|
{
|
|
"epoch": 0.347239263803681,
|
|
"grad_norm": 33.09320412279389,
|
|
"learning_rate": 7.147107438016529e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2537,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.5,
|
|
"step": 283
|
|
},
|
|
{
|
|
"epoch": 0.34846625766871164,
|
|
"grad_norm": 40.734540729970675,
|
|
"learning_rate": 7.143801652892562e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.272,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 284
|
|
},
|
|
{
|
|
"epoch": 0.3496932515337423,
|
|
"grad_norm": 30.9410258264426,
|
|
"learning_rate": 7.140495867768595e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2413,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 285
|
|
},
|
|
{
|
|
"epoch": 0.350920245398773,
|
|
"grad_norm": 37.246332837421065,
|
|
"learning_rate": 7.137190082644628e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3337,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 286
|
|
},
|
|
{
|
|
"epoch": 0.3521472392638037,
|
|
"grad_norm": 40.87355956326826,
|
|
"learning_rate": 7.133884297520661e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2981,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.68359375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 287
|
|
},
|
|
{
|
|
"epoch": 0.35337423312883437,
|
|
"grad_norm": 37.148615900618886,
|
|
"learning_rate": 7.130578512396693e-07,
|
|
"logits/chosen": -0.05908203125,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3085,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 288
|
|
},
|
|
{
|
|
"epoch": 0.35460122699386504,
|
|
"grad_norm": 29.32351813266847,
|
|
"learning_rate": 7.127272727272727e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.1987,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 289
|
|
},
|
|
{
|
|
"epoch": 0.3558282208588957,
|
|
"grad_norm": 34.03478605235303,
|
|
"learning_rate": 7.12396694214876e-07,
|
|
"logits/chosen": -0.359375,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2024,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 0.3570552147239264,
|
|
"grad_norm": 38.57928856233812,
|
|
"learning_rate": 7.120661157024793e-07,
|
|
"logits/chosen": -0.296875,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3224,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 291
|
|
},
|
|
{
|
|
"epoch": 0.35828220858895704,
|
|
"grad_norm": 36.5133266014805,
|
|
"learning_rate": 7.117355371900826e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2451,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 292
|
|
},
|
|
{
|
|
"epoch": 0.3595092024539877,
|
|
"grad_norm": 31.073871401773395,
|
|
"learning_rate": 7.114049586776859e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2162,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.70703125,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 293
|
|
},
|
|
{
|
|
"epoch": 0.36073619631901843,
|
|
"grad_norm": 31.419020907335653,
|
|
"learning_rate": 7.110743801652892e-07,
|
|
"logits/chosen": -0.427734375,
|
|
"logits/rejected": -0.578125,
|
|
"logps/chosen": -464.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2389,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 294
|
|
},
|
|
{
|
|
"epoch": 0.3619631901840491,
|
|
"grad_norm": 33.66575248055817,
|
|
"learning_rate": 7.107438016528926e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2242,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.93359375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 295
|
|
},
|
|
{
|
|
"epoch": 0.36319018404907977,
|
|
"grad_norm": 40.06405186289625,
|
|
"learning_rate": 7.104132231404959e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3209,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 296
|
|
},
|
|
{
|
|
"epoch": 0.36441717791411044,
|
|
"grad_norm": 30.371449913820346,
|
|
"learning_rate": 7.100826446280991e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2354,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 297
|
|
},
|
|
{
|
|
"epoch": 0.3656441717791411,
|
|
"grad_norm": 40.780831154790015,
|
|
"learning_rate": 7.097520661157024e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2901,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 298
|
|
},
|
|
{
|
|
"epoch": 0.36687116564417177,
|
|
"grad_norm": 48.39132901882327,
|
|
"learning_rate": 7.094214876033057e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3427,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.9453125,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 299
|
|
},
|
|
{
|
|
"epoch": 0.36809815950920244,
|
|
"grad_norm": 39.25730980280545,
|
|
"learning_rate": 7.09090909090909e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3287,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.3693251533742331,
|
|
"grad_norm": 34.87785766178715,
|
|
"learning_rate": 7.087603305785124e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2823,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 301
|
|
},
|
|
{
|
|
"epoch": 0.37055214723926383,
|
|
"grad_norm": 30.464291800676158,
|
|
"learning_rate": 7.084297520661157e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2182,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 302
|
|
},
|
|
{
|
|
"epoch": 0.3717791411042945,
|
|
"grad_norm": 45.78814856158198,
|
|
"learning_rate": 7.08099173553719e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3512,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 303
|
|
},
|
|
{
|
|
"epoch": 0.37300613496932516,
|
|
"grad_norm": 33.26834608933062,
|
|
"learning_rate": 7.077685950413223e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2448,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 304
|
|
},
|
|
{
|
|
"epoch": 0.37423312883435583,
|
|
"grad_norm": 35.609604648743236,
|
|
"learning_rate": 7.074380165289256e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2744,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.625,
|
|
"step": 305
|
|
},
|
|
{
|
|
"epoch": 0.3754601226993865,
|
|
"grad_norm": 43.837358024220855,
|
|
"learning_rate": 7.071074380165288e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.48046875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2737,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.5,
|
|
"step": 306
|
|
},
|
|
{
|
|
"epoch": 0.37668711656441717,
|
|
"grad_norm": 39.17650253748135,
|
|
"learning_rate": 7.067768595041322e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3713,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 307
|
|
},
|
|
{
|
|
"epoch": 0.37791411042944784,
|
|
"grad_norm": 36.807676017880326,
|
|
"learning_rate": 7.064462809917355e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3165,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 308
|
|
},
|
|
{
|
|
"epoch": 0.3791411042944785,
|
|
"grad_norm": 32.288551419968854,
|
|
"learning_rate": 7.061157024793388e-07,
|
|
"logits/chosen": -0.30078125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2283,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 309
|
|
},
|
|
{
|
|
"epoch": 0.3803680981595092,
|
|
"grad_norm": 47.130716517812814,
|
|
"learning_rate": 7.057851239669421e-07,
|
|
"logits/chosen": -0.0908203125,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3596,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.15234375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 0.3815950920245399,
|
|
"grad_norm": 34.12148114082454,
|
|
"learning_rate": 7.054545454545454e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2543,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.375,
|
|
"step": 311
|
|
},
|
|
{
|
|
"epoch": 0.38282208588957056,
|
|
"grad_norm": 44.35564676375722,
|
|
"learning_rate": 7.051239669421488e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.4044,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 312
|
|
},
|
|
{
|
|
"epoch": 0.38404907975460123,
|
|
"grad_norm": 33.82923751151301,
|
|
"learning_rate": 7.047933884297521e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2685,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8828125,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 313
|
|
},
|
|
{
|
|
"epoch": 0.3852760736196319,
|
|
"grad_norm": 33.17604766084792,
|
|
"learning_rate": 7.044628099173554e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -456.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.1847,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.25,
|
|
"rewards/margins": 7.96875,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 314
|
|
},
|
|
{
|
|
"epoch": 0.38650306748466257,
|
|
"grad_norm": 38.90508166705689,
|
|
"learning_rate": 7.041322314049586e-07,
|
|
"logits/chosen": -0.06396484375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.2412,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.5546875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 315
|
|
},
|
|
{
|
|
"epoch": 0.38773006134969323,
|
|
"grad_norm": 31.950078808944912,
|
|
"learning_rate": 7.038016528925619e-07,
|
|
"logits/chosen": -0.06982421875,
|
|
"logits/rejected": -0.2158203125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2551,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 316
|
|
},
|
|
{
|
|
"epoch": 0.3889570552147239,
|
|
"grad_norm": 34.528377745666106,
|
|
"learning_rate": 7.034710743801652e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2296,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.875,
|
|
"step": 317
|
|
},
|
|
{
|
|
"epoch": 0.3901840490797546,
|
|
"grad_norm": 46.9506582652498,
|
|
"learning_rate": 7.031404958677686e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3503,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 318
|
|
},
|
|
{
|
|
"epoch": 0.3914110429447853,
|
|
"grad_norm": 26.867054074267944,
|
|
"learning_rate": 7.028099173553719e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.1701,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -6.125,
|
|
"step": 319
|
|
},
|
|
{
|
|
"epoch": 0.39263803680981596,
|
|
"grad_norm": 40.540053142562215,
|
|
"learning_rate": 7.024793388429752e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -318.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.3046,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 0.39386503067484663,
|
|
"grad_norm": 46.49642012587718,
|
|
"learning_rate": 7.021487603305785e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.2961,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 321
|
|
},
|
|
{
|
|
"epoch": 0.3950920245398773,
|
|
"grad_norm": 33.72622281793196,
|
|
"learning_rate": 7.018181818181818e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2496,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 322
|
|
},
|
|
{
|
|
"epoch": 0.39631901840490796,
|
|
"grad_norm": 39.04984997032008,
|
|
"learning_rate": 7.014876033057851e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3837,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 323
|
|
},
|
|
{
|
|
"epoch": 0.39754601226993863,
|
|
"grad_norm": 33.57036215026917,
|
|
"learning_rate": 7.011570247933884e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2714,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.5,
|
|
"step": 324
|
|
},
|
|
{
|
|
"epoch": 0.3987730061349693,
|
|
"grad_norm": 45.64995516248102,
|
|
"learning_rate": 7.008264462809917e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.2177734375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.4112,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"grad_norm": 34.28943945713478,
|
|
"learning_rate": 7.00495867768595e-07,
|
|
"logits/chosen": -0.310546875,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2004,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 326
|
|
},
|
|
{
|
|
"epoch": 0.4012269938650307,
|
|
"grad_norm": 48.50197304580951,
|
|
"learning_rate": 7.001652892561983e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3949,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 327
|
|
},
|
|
{
|
|
"epoch": 0.40245398773006136,
|
|
"grad_norm": 40.72777385232644,
|
|
"learning_rate": 6.998347107438016e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.3146,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 328
|
|
},
|
|
{
|
|
"epoch": 0.403680981595092,
|
|
"grad_norm": 35.55525281765151,
|
|
"learning_rate": 6.995041322314049e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2936,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 329
|
|
},
|
|
{
|
|
"epoch": 0.4049079754601227,
|
|
"grad_norm": 42.0221302571432,
|
|
"learning_rate": 6.991735537190083e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.5234375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2435,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 0.40613496932515336,
|
|
"grad_norm": 32.84700653520424,
|
|
"learning_rate": 6.988429752066116e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2501,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 331
|
|
},
|
|
{
|
|
"epoch": 0.40736196319018403,
|
|
"grad_norm": 27.379927529722597,
|
|
"learning_rate": 6.985123966942148e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2017,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 332
|
|
},
|
|
{
|
|
"epoch": 0.4085889570552147,
|
|
"grad_norm": 36.70659629756495,
|
|
"learning_rate": 6.981818181818181e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3078,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 333
|
|
},
|
|
{
|
|
"epoch": 0.4098159509202454,
|
|
"grad_norm": 34.15971280742034,
|
|
"learning_rate": 6.978512396694214e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.273,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 334
|
|
},
|
|
{
|
|
"epoch": 0.4110429447852761,
|
|
"grad_norm": 37.2375068536928,
|
|
"learning_rate": 6.975206611570247e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2248,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 335
|
|
},
|
|
{
|
|
"epoch": 0.41226993865030676,
|
|
"grad_norm": 36.84981287485191,
|
|
"learning_rate": 6.971900826446281e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2764,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 336
|
|
},
|
|
{
|
|
"epoch": 0.4134969325153374,
|
|
"grad_norm": 46.94122451213228,
|
|
"learning_rate": 6.968595041322314e-07,
|
|
"logits/chosen": -0.09375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.389,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 337
|
|
},
|
|
{
|
|
"epoch": 0.4147239263803681,
|
|
"grad_norm": 38.247859512270246,
|
|
"learning_rate": 6.965289256198347e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3295,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.0,
|
|
"step": 338
|
|
},
|
|
{
|
|
"epoch": 0.41595092024539876,
|
|
"grad_norm": 40.84741702132196,
|
|
"learning_rate": 6.96198347107438e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.2109375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3275,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 339
|
|
},
|
|
{
|
|
"epoch": 0.4171779141104294,
|
|
"grad_norm": 43.700613878566976,
|
|
"learning_rate": 6.958677685950413e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2879,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 0.41840490797546015,
|
|
"grad_norm": 31.54531879054049,
|
|
"learning_rate": 6.955371900826445e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2805,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 341
|
|
},
|
|
{
|
|
"epoch": 0.4196319018404908,
|
|
"grad_norm": 39.0902727508844,
|
|
"learning_rate": 6.952066115702479e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3322,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 342
|
|
},
|
|
{
|
|
"epoch": 0.4208588957055215,
|
|
"grad_norm": 39.32051666516945,
|
|
"learning_rate": 6.948760330578512e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3545,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 343
|
|
},
|
|
{
|
|
"epoch": 0.42208588957055215,
|
|
"grad_norm": 42.12526028461151,
|
|
"learning_rate": 6.945454545454545e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3584,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 344
|
|
},
|
|
{
|
|
"epoch": 0.4233128834355828,
|
|
"grad_norm": 39.02371998906191,
|
|
"learning_rate": 6.942148760330578e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3288,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 345
|
|
},
|
|
{
|
|
"epoch": 0.4245398773006135,
|
|
"grad_norm": 41.19695447421548,
|
|
"learning_rate": 6.938842975206611e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3569,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.125,
|
|
"step": 346
|
|
},
|
|
{
|
|
"epoch": 0.42576687116564416,
|
|
"grad_norm": 28.472619739287644,
|
|
"learning_rate": 6.935537190082645e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2332,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.25,
|
|
"step": 347
|
|
},
|
|
{
|
|
"epoch": 0.4269938650306748,
|
|
"grad_norm": 33.19419614917591,
|
|
"learning_rate": 6.932231404958678e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2594,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.91796875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 348
|
|
},
|
|
{
|
|
"epoch": 0.42822085889570555,
|
|
"grad_norm": 38.111059722172605,
|
|
"learning_rate": 6.928925619834711e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3255,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 349
|
|
},
|
|
{
|
|
"epoch": 0.4294478527607362,
|
|
"grad_norm": 37.49784569992658,
|
|
"learning_rate": 6.925619834710743e-07,
|
|
"logits/chosen": -0.0986328125,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2857,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 0.4306748466257669,
|
|
"grad_norm": 42.647163206760126,
|
|
"learning_rate": 6.922314049586776e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3505,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.5078125,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 351
|
|
},
|
|
{
|
|
"epoch": 0.43190184049079755,
|
|
"grad_norm": 36.26120743659289,
|
|
"learning_rate": 6.919008264462809e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3453,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 352
|
|
},
|
|
{
|
|
"epoch": 0.4331288343558282,
|
|
"grad_norm": 30.933441803520967,
|
|
"learning_rate": 6.915702479338843e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2406,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 353
|
|
},
|
|
{
|
|
"epoch": 0.4343558282208589,
|
|
"grad_norm": 34.874499707263155,
|
|
"learning_rate": 6.912396694214876e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2963,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 354
|
|
},
|
|
{
|
|
"epoch": 0.43558282208588955,
|
|
"grad_norm": 39.848840136754845,
|
|
"learning_rate": 6.909090909090909e-07,
|
|
"logits/chosen": -0.296875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3581,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 355
|
|
},
|
|
{
|
|
"epoch": 0.4368098159509202,
|
|
"grad_norm": 44.03875893573232,
|
|
"learning_rate": 6.905785123966942e-07,
|
|
"logits/chosen": -0.0966796875,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3406,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 356
|
|
},
|
|
{
|
|
"epoch": 0.43803680981595094,
|
|
"grad_norm": 36.68968799262988,
|
|
"learning_rate": 6.902479338842975e-07,
|
|
"logits/chosen": -0.3515625,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2926,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 357
|
|
},
|
|
{
|
|
"epoch": 0.4392638036809816,
|
|
"grad_norm": 36.204158508103596,
|
|
"learning_rate": 6.899173553719008e-07,
|
|
"logits/chosen": -0.30078125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2673,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 358
|
|
},
|
|
{
|
|
"epoch": 0.4404907975460123,
|
|
"grad_norm": 40.16138545949422,
|
|
"learning_rate": 6.89586776859504e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2988,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 359
|
|
},
|
|
{
|
|
"epoch": 0.44171779141104295,
|
|
"grad_norm": 35.21609251200673,
|
|
"learning_rate": 6.892561983471074e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2614,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 0.4429447852760736,
|
|
"grad_norm": 33.95190421090596,
|
|
"learning_rate": 6.889256198347107e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2599,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 361
|
|
},
|
|
{
|
|
"epoch": 0.4441717791411043,
|
|
"grad_norm": 32.41395226001234,
|
|
"learning_rate": 6.88595041322314e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2625,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 362
|
|
},
|
|
{
|
|
"epoch": 0.44539877300613495,
|
|
"grad_norm": 31.06408733284358,
|
|
"learning_rate": 6.882644628099173e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2362,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.375,
|
|
"step": 363
|
|
},
|
|
{
|
|
"epoch": 0.4466257668711656,
|
|
"grad_norm": 40.778000027843554,
|
|
"learning_rate": 6.879338842975206e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2885,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 364
|
|
},
|
|
{
|
|
"epoch": 0.44785276073619634,
|
|
"grad_norm": 30.8545654380382,
|
|
"learning_rate": 6.87603305785124e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2419,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 365
|
|
},
|
|
{
|
|
"epoch": 0.449079754601227,
|
|
"grad_norm": 35.25923514975111,
|
|
"learning_rate": 6.872727272727273e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.21484375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3324,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 366
|
|
},
|
|
{
|
|
"epoch": 0.4503067484662577,
|
|
"grad_norm": 29.75727838549342,
|
|
"learning_rate": 6.869421487603306e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2265,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 367
|
|
},
|
|
{
|
|
"epoch": 0.45153374233128835,
|
|
"grad_norm": 41.34860966923763,
|
|
"learning_rate": 6.866115702479338e-07,
|
|
"logits/chosen": -0.060546875,
|
|
"logits/rejected": -0.11328125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3251,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 368
|
|
},
|
|
{
|
|
"epoch": 0.452760736196319,
|
|
"grad_norm": 39.802701196225414,
|
|
"learning_rate": 6.862809917355371e-07,
|
|
"logits/chosen": -0.08349609375,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3103,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 369
|
|
},
|
|
{
|
|
"epoch": 0.4539877300613497,
|
|
"grad_norm": 35.44626532768664,
|
|
"learning_rate": 6.859504132231404e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3001,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 0.45521472392638035,
|
|
"grad_norm": 34.081640470685045,
|
|
"learning_rate": 6.856198347107438e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -458.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.268,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 371
|
|
},
|
|
{
|
|
"epoch": 0.456441717791411,
|
|
"grad_norm": 32.10146522336104,
|
|
"learning_rate": 6.852892561983471e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2227,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 372
|
|
},
|
|
{
|
|
"epoch": 0.45766871165644174,
|
|
"grad_norm": 38.40212256087746,
|
|
"learning_rate": 6.849586776859504e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3491,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 373
|
|
},
|
|
{
|
|
"epoch": 0.4588957055214724,
|
|
"grad_norm": 33.89078532536025,
|
|
"learning_rate": 6.846280991735537e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2439,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 374
|
|
},
|
|
{
|
|
"epoch": 0.4601226993865031,
|
|
"grad_norm": 38.36603856770271,
|
|
"learning_rate": 6.84297520661157e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2743,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 0.46134969325153374,
|
|
"grad_norm": 32.73897251445023,
|
|
"learning_rate": 6.839669421487603e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.231,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 376
|
|
},
|
|
{
|
|
"epoch": 0.4625766871165644,
|
|
"grad_norm": 36.44881761827906,
|
|
"learning_rate": 6.836363636363636e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2817,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 377
|
|
},
|
|
{
|
|
"epoch": 0.4638036809815951,
|
|
"grad_norm": 36.92866834323212,
|
|
"learning_rate": 6.833057851239669e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3001,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.5,
|
|
"step": 378
|
|
},
|
|
{
|
|
"epoch": 0.46503067484662575,
|
|
"grad_norm": 32.73144242036404,
|
|
"learning_rate": 6.829752066115702e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2471,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 379
|
|
},
|
|
{
|
|
"epoch": 0.4662576687116564,
|
|
"grad_norm": 34.6763093414022,
|
|
"learning_rate": 6.826446280991735e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2659,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 0.46748466257668714,
|
|
"grad_norm": 34.98446202126656,
|
|
"learning_rate": 6.823140495867768e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3193,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 381
|
|
},
|
|
{
|
|
"epoch": 0.4687116564417178,
|
|
"grad_norm": 33.068661800772716,
|
|
"learning_rate": 6.819834710743801e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2838,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 382
|
|
},
|
|
{
|
|
"epoch": 0.4699386503067485,
|
|
"grad_norm": 37.187464031340724,
|
|
"learning_rate": 6.816528925619835e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.33,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 383
|
|
},
|
|
{
|
|
"epoch": 0.47116564417177914,
|
|
"grad_norm": 35.133860722289874,
|
|
"learning_rate": 6.813223140495868e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2466,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -6.5,
|
|
"step": 384
|
|
},
|
|
{
|
|
"epoch": 0.4723926380368098,
|
|
"grad_norm": 40.48440514462701,
|
|
"learning_rate": 6.809917355371901e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3252,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.2578125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 385
|
|
},
|
|
{
|
|
"epoch": 0.4736196319018405,
|
|
"grad_norm": 40.628340663991764,
|
|
"learning_rate": 6.806611570247933e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3025,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": -0.359375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 386
|
|
},
|
|
{
|
|
"epoch": 0.47484662576687114,
|
|
"grad_norm": 51.72741620856106,
|
|
"learning_rate": 6.803305785123966e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2838,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 387
|
|
},
|
|
{
|
|
"epoch": 0.47607361963190187,
|
|
"grad_norm": 43.50814776279542,
|
|
"learning_rate": 6.8e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3977,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.05029296875,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 388
|
|
},
|
|
{
|
|
"epoch": 0.47730061349693254,
|
|
"grad_norm": 44.9083628878753,
|
|
"learning_rate": 6.796694214876033e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3913,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.44140625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 389
|
|
},
|
|
{
|
|
"epoch": 0.4785276073619632,
|
|
"grad_norm": 42.893730022875914,
|
|
"learning_rate": 6.793388429752066e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3759,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": -0.263671875,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -6.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 0.47975460122699387,
|
|
"grad_norm": 46.56902163562806,
|
|
"learning_rate": 6.790082644628099e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4122,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.3984375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 391
|
|
},
|
|
{
|
|
"epoch": 0.48098159509202454,
|
|
"grad_norm": 32.46380915314015,
|
|
"learning_rate": 6.786776859504132e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2695,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.69921875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 392
|
|
},
|
|
{
|
|
"epoch": 0.4822085889570552,
|
|
"grad_norm": 36.411647842869584,
|
|
"learning_rate": 6.783471074380165e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2936,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 393
|
|
},
|
|
{
|
|
"epoch": 0.4834355828220859,
|
|
"grad_norm": 36.43693693488016,
|
|
"learning_rate": 6.780165289256199e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2504,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 394
|
|
},
|
|
{
|
|
"epoch": 0.48466257668711654,
|
|
"grad_norm": 36.92302249692751,
|
|
"learning_rate": 6.776859504132231e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3275,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 395
|
|
},
|
|
{
|
|
"epoch": 0.48588957055214727,
|
|
"grad_norm": 31.266490118113346,
|
|
"learning_rate": 6.773553719008264e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.486328125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2215,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.462890625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 396
|
|
},
|
|
{
|
|
"epoch": 0.48711656441717793,
|
|
"grad_norm": 34.937182799458625,
|
|
"learning_rate": 6.770247933884297e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2752,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 397
|
|
},
|
|
{
|
|
"epoch": 0.4883435582822086,
|
|
"grad_norm": 38.419527775162116,
|
|
"learning_rate": 6.76694214876033e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3231,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 398
|
|
},
|
|
{
|
|
"epoch": 0.48957055214723927,
|
|
"grad_norm": 35.54458075122045,
|
|
"learning_rate": 6.763636363636363e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.306,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 399
|
|
},
|
|
{
|
|
"epoch": 0.49079754601226994,
|
|
"grad_norm": 47.751486198572856,
|
|
"learning_rate": 6.760330578512397e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3896,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.4920245398773006,
|
|
"grad_norm": 31.519656830052252,
|
|
"learning_rate": 6.75702479338843e-07,
|
|
"logits/chosen": -0.328125,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.2337,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 401
|
|
},
|
|
{
|
|
"epoch": 0.49325153374233127,
|
|
"grad_norm": 36.11809592072752,
|
|
"learning_rate": 6.753719008264463e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3095,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 402
|
|
},
|
|
{
|
|
"epoch": 0.49447852760736194,
|
|
"grad_norm": 38.21909292164752,
|
|
"learning_rate": 6.750413223140496e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3169,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.875,
|
|
"step": 403
|
|
},
|
|
{
|
|
"epoch": 0.49570552147239266,
|
|
"grad_norm": 32.22630836566114,
|
|
"learning_rate": 6.747107438016528e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2283,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.0,
|
|
"step": 404
|
|
},
|
|
{
|
|
"epoch": 0.49693251533742333,
|
|
"grad_norm": 37.593793152840426,
|
|
"learning_rate": 6.743801652892561e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3305,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 405
|
|
},
|
|
{
|
|
"epoch": 0.498159509202454,
|
|
"grad_norm": 41.53333095908596,
|
|
"learning_rate": 6.740495867768595e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.2109375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3932,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 406
|
|
},
|
|
{
|
|
"epoch": 0.49938650306748467,
|
|
"grad_norm": 42.42876322706536,
|
|
"learning_rate": 6.737190082644628e-07,
|
|
"logits/chosen": -0.0986328125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3513,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 407
|
|
},
|
|
{
|
|
"epoch": 0.5006134969325153,
|
|
"grad_norm": 48.89324534838805,
|
|
"learning_rate": 6.733884297520661e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.4164,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.80078125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 408
|
|
},
|
|
{
|
|
"epoch": 0.501840490797546,
|
|
"grad_norm": 45.10686444580896,
|
|
"learning_rate": 6.730578512396694e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3079,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 409
|
|
},
|
|
{
|
|
"epoch": 0.5030674846625767,
|
|
"grad_norm": 38.62243714595701,
|
|
"learning_rate": 6.727272727272727e-07,
|
|
"logits/chosen": -0.119140625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2966,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 0.5042944785276073,
|
|
"grad_norm": 37.41136253841519,
|
|
"learning_rate": 6.72396694214876e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.284,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 411
|
|
},
|
|
{
|
|
"epoch": 0.505521472392638,
|
|
"grad_norm": 47.69531455944101,
|
|
"learning_rate": 6.720661157024794e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3619,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 412
|
|
},
|
|
{
|
|
"epoch": 0.5067484662576687,
|
|
"grad_norm": 39.571364824526874,
|
|
"learning_rate": 6.717355371900826e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3069,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 413
|
|
},
|
|
{
|
|
"epoch": 0.5079754601226993,
|
|
"grad_norm": 34.32054957685153,
|
|
"learning_rate": 6.714049586776859e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2497,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 414
|
|
},
|
|
{
|
|
"epoch": 0.50920245398773,
|
|
"grad_norm": 35.2489180238823,
|
|
"learning_rate": 6.710743801652892e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2788,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 415
|
|
},
|
|
{
|
|
"epoch": 0.5104294478527608,
|
|
"grad_norm": 28.41248357906051,
|
|
"learning_rate": 6.707438016528925e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.186,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 416
|
|
},
|
|
{
|
|
"epoch": 0.5116564417177915,
|
|
"grad_norm": 35.52909474934163,
|
|
"learning_rate": 6.704132231404958e-07,
|
|
"logits/chosen": -0.328125,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2995,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 417
|
|
},
|
|
{
|
|
"epoch": 0.5128834355828221,
|
|
"grad_norm": 36.29519060781132,
|
|
"learning_rate": 6.700826446280992e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2828,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 418
|
|
},
|
|
{
|
|
"epoch": 0.5141104294478528,
|
|
"grad_norm": 31.87680208130441,
|
|
"learning_rate": 6.697520661157025e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2486,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 419
|
|
},
|
|
{
|
|
"epoch": 0.5153374233128835,
|
|
"grad_norm": 40.881697279170936,
|
|
"learning_rate": 6.694214876033058e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3244,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 0.5165644171779141,
|
|
"grad_norm": 34.11082321167536,
|
|
"learning_rate": 6.69090909090909e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.339,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 421
|
|
},
|
|
{
|
|
"epoch": 0.5177914110429448,
|
|
"grad_norm": 35.18821029472038,
|
|
"learning_rate": 6.687603305785123e-07,
|
|
"logits/chosen": -0.0810546875,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3166,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.69921875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 422
|
|
},
|
|
{
|
|
"epoch": 0.5190184049079755,
|
|
"grad_norm": 41.06855790218669,
|
|
"learning_rate": 6.684297520661156e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3606,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 423
|
|
},
|
|
{
|
|
"epoch": 0.5202453987730061,
|
|
"grad_norm": 36.724024878608596,
|
|
"learning_rate": 6.68099173553719e-07,
|
|
"logits/chosen": -0.123046875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3438,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 424
|
|
},
|
|
{
|
|
"epoch": 0.5214723926380368,
|
|
"grad_norm": 41.01181618114964,
|
|
"learning_rate": 6.677685950413223e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.4073,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 0.5226993865030675,
|
|
"grad_norm": 44.64404869891064,
|
|
"learning_rate": 6.674380165289256e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3035,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 426
|
|
},
|
|
{
|
|
"epoch": 0.5239263803680981,
|
|
"grad_norm": 91.8646628453638,
|
|
"learning_rate": 6.671074380165289e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.299,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 427
|
|
},
|
|
{
|
|
"epoch": 0.5251533742331288,
|
|
"grad_norm": 36.99313104379447,
|
|
"learning_rate": 6.667768595041322e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3507,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 428
|
|
},
|
|
{
|
|
"epoch": 0.5263803680981595,
|
|
"grad_norm": 33.0190569729192,
|
|
"learning_rate": 6.664462809917356e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.246,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.443359375,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -6.46875,
|
|
"step": 429
|
|
},
|
|
{
|
|
"epoch": 0.5276073619631901,
|
|
"grad_norm": 30.408983999620382,
|
|
"learning_rate": 6.661157024793388e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2248,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.9921875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 0.5288343558282208,
|
|
"grad_norm": 47.87075432548269,
|
|
"learning_rate": 6.657851239669421e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.4545,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.45703125,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.25,
|
|
"step": 431
|
|
},
|
|
{
|
|
"epoch": 0.5300613496932516,
|
|
"grad_norm": 43.65339869428465,
|
|
"learning_rate": 6.654545454545454e-07,
|
|
"logits/chosen": -0.0927734375,
|
|
"logits/rejected": -0.1435546875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3724,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.875,
|
|
"step": 432
|
|
},
|
|
{
|
|
"epoch": 0.5312883435582823,
|
|
"grad_norm": 45.133784704459096,
|
|
"learning_rate": 6.651239669421487e-07,
|
|
"logits/chosen": -0.078125,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3416,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.39453125,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 433
|
|
},
|
|
{
|
|
"epoch": 0.5325153374233129,
|
|
"grad_norm": 40.47930492236747,
|
|
"learning_rate": 6.64793388429752e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.254,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.25,
|
|
"step": 434
|
|
},
|
|
{
|
|
"epoch": 0.5337423312883436,
|
|
"grad_norm": 41.262979059220854,
|
|
"learning_rate": 6.644628099173554e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 435
|
|
},
|
|
{
|
|
"epoch": 0.5349693251533743,
|
|
"grad_norm": 38.67538014065472,
|
|
"learning_rate": 6.641322314049587e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2467,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 436
|
|
},
|
|
{
|
|
"epoch": 0.5361963190184049,
|
|
"grad_norm": 31.32861125197491,
|
|
"learning_rate": 6.63801652892562e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.248,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 437
|
|
},
|
|
{
|
|
"epoch": 0.5374233128834356,
|
|
"grad_norm": 42.810824724719026,
|
|
"learning_rate": 6.634710743801653e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3969,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.80078125,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 438
|
|
},
|
|
{
|
|
"epoch": 0.5386503067484663,
|
|
"grad_norm": 36.79651022084742,
|
|
"learning_rate": 6.631404958677685e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2838,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 439
|
|
},
|
|
{
|
|
"epoch": 0.5398773006134969,
|
|
"grad_norm": 49.84881532208057,
|
|
"learning_rate": 6.628099173553718e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3608,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.384765625,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 0.5411042944785276,
|
|
"grad_norm": 37.976634967176224,
|
|
"learning_rate": 6.624793388429752e-07,
|
|
"logits/chosen": -0.06640625,
|
|
"logits/rejected": -0.1826171875,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.328,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 441
|
|
},
|
|
{
|
|
"epoch": 0.5423312883435583,
|
|
"grad_norm": 37.80188895923485,
|
|
"learning_rate": 6.621487603305785e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3304,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 442
|
|
},
|
|
{
|
|
"epoch": 0.5435582822085889,
|
|
"grad_norm": 40.24381216440989,
|
|
"learning_rate": 6.618181818181818e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3159,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 443
|
|
},
|
|
{
|
|
"epoch": 0.5447852760736196,
|
|
"grad_norm": 37.922945446281,
|
|
"learning_rate": 6.614876033057851e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2446,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 7.25,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 444
|
|
},
|
|
{
|
|
"epoch": 0.5460122699386503,
|
|
"grad_norm": 38.45325276499507,
|
|
"learning_rate": 6.611570247933884e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.389,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 445
|
|
},
|
|
{
|
|
"epoch": 0.5472392638036809,
|
|
"grad_norm": 33.02180393783636,
|
|
"learning_rate": 6.608264462809917e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2348,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 446
|
|
},
|
|
{
|
|
"epoch": 0.5484662576687117,
|
|
"grad_norm": 31.670625439038492,
|
|
"learning_rate": 6.604958677685951e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2551,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 447
|
|
},
|
|
{
|
|
"epoch": 0.5496932515337424,
|
|
"grad_norm": 37.605046960174505,
|
|
"learning_rate": 6.601652892561983e-07,
|
|
"logits/chosen": -0.10888671875,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3209,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.25,
|
|
"step": 448
|
|
},
|
|
{
|
|
"epoch": 0.550920245398773,
|
|
"grad_norm": 34.080535896094865,
|
|
"learning_rate": 6.598347107438016e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2712,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 449
|
|
},
|
|
{
|
|
"epoch": 0.5521472392638037,
|
|
"grad_norm": 35.069217343013015,
|
|
"learning_rate": 6.595041322314049e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.285,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 0.5533742331288344,
|
|
"grad_norm": 33.87966228532685,
|
|
"learning_rate": 6.591735537190082e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.3043,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.5,
|
|
"step": 451
|
|
},
|
|
{
|
|
"epoch": 0.554601226993865,
|
|
"grad_norm": 35.08833886262592,
|
|
"learning_rate": 6.588429752066115e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.258,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.421875,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 452
|
|
},
|
|
{
|
|
"epoch": 0.5558282208588957,
|
|
"grad_norm": 30.509116856340853,
|
|
"learning_rate": 6.585123966942149e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.48828125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.214,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 2.578125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 453
|
|
},
|
|
{
|
|
"epoch": 0.5570552147239264,
|
|
"grad_norm": 36.06945734305074,
|
|
"learning_rate": 6.581818181818182e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.19921875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2945,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.625,
|
|
"step": 454
|
|
},
|
|
{
|
|
"epoch": 0.558282208588957,
|
|
"grad_norm": 35.814869576892626,
|
|
"learning_rate": 6.578512396694215e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2399,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 455
|
|
},
|
|
{
|
|
"epoch": 0.5595092024539877,
|
|
"grad_norm": 38.450506141939925,
|
|
"learning_rate": 6.575206611570248e-07,
|
|
"logits/chosen": -0.12060546875,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -312.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2952,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.125,
|
|
"step": 456
|
|
},
|
|
{
|
|
"epoch": 0.5607361963190184,
|
|
"grad_norm": 37.93391723235656,
|
|
"learning_rate": 6.57190082644628e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3025,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 457
|
|
},
|
|
{
|
|
"epoch": 0.561963190184049,
|
|
"grad_norm": 33.44669035850576,
|
|
"learning_rate": 6.568595041322313e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2297,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.125,
|
|
"step": 458
|
|
},
|
|
{
|
|
"epoch": 0.5631901840490797,
|
|
"grad_norm": 34.06851635286411,
|
|
"learning_rate": 6.565289256198347e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2786,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 459
|
|
},
|
|
{
|
|
"epoch": 0.5644171779141104,
|
|
"grad_norm": 34.68751260016757,
|
|
"learning_rate": 6.56198347107438e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.26,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 0.5656441717791411,
|
|
"grad_norm": 38.71078681224242,
|
|
"learning_rate": 6.558677685950413e-07,
|
|
"logits/chosen": -0.3515625,
|
|
"logits/rejected": -0.48046875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2741,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.80078125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 461
|
|
},
|
|
{
|
|
"epoch": 0.5668711656441717,
|
|
"grad_norm": 39.327400074833584,
|
|
"learning_rate": 6.555371900826446e-07,
|
|
"logits/chosen": -0.11279296875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3248,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.875,
|
|
"step": 462
|
|
},
|
|
{
|
|
"epoch": 0.5680981595092025,
|
|
"grad_norm": 37.637164385131356,
|
|
"learning_rate": 6.552066115702479e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2848,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.98046875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 463
|
|
},
|
|
{
|
|
"epoch": 0.5693251533742332,
|
|
"grad_norm": 41.25059009443085,
|
|
"learning_rate": 6.548760330578512e-07,
|
|
"logits/chosen": -0.09619140625,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2951,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 464
|
|
},
|
|
{
|
|
"epoch": 0.5705521472392638,
|
|
"grad_norm": 27.065991136479923,
|
|
"learning_rate": 6.545454545454546e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.1759,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 465
|
|
},
|
|
{
|
|
"epoch": 0.5717791411042945,
|
|
"grad_norm": 43.40449571309543,
|
|
"learning_rate": 6.542148760330578e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2899,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 466
|
|
},
|
|
{
|
|
"epoch": 0.5730061349693252,
|
|
"grad_norm": 40.392254031751044,
|
|
"learning_rate": 6.538842975206611e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3087,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 467
|
|
},
|
|
{
|
|
"epoch": 0.5742331288343558,
|
|
"grad_norm": 41.32661857185299,
|
|
"learning_rate": 6.535537190082644e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3022,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 468
|
|
},
|
|
{
|
|
"epoch": 0.5754601226993865,
|
|
"grad_norm": 29.07874137818143,
|
|
"learning_rate": 6.532231404958677e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.1838,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 469
|
|
},
|
|
{
|
|
"epoch": 0.5766871165644172,
|
|
"grad_norm": 35.40917214666752,
|
|
"learning_rate": 6.52892561983471e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.301,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 470
|
|
},
|
|
{
|
|
"epoch": 0.5779141104294478,
|
|
"grad_norm": 38.07545863951338,
|
|
"learning_rate": 6.525619834710744e-07,
|
|
"logits/chosen": -0.03369140625,
|
|
"logits/rejected": -0.203125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3263,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 471
|
|
},
|
|
{
|
|
"epoch": 0.5791411042944785,
|
|
"grad_norm": 40.21918996431336,
|
|
"learning_rate": 6.522314049586777e-07,
|
|
"logits/chosen": -0.08447265625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3173,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 472
|
|
},
|
|
{
|
|
"epoch": 0.5803680981595092,
|
|
"grad_norm": 37.03158250739783,
|
|
"learning_rate": 6.51900826446281e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.3328,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 473
|
|
},
|
|
{
|
|
"epoch": 0.5815950920245399,
|
|
"grad_norm": 38.12266289452875,
|
|
"learning_rate": 6.515702479338843e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2648,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 474
|
|
},
|
|
{
|
|
"epoch": 0.5828220858895705,
|
|
"grad_norm": 31.08159067757818,
|
|
"learning_rate": 6.512396694214875e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2361,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 475
|
|
},
|
|
{
|
|
"epoch": 0.5840490797546012,
|
|
"grad_norm": 31.0829148956811,
|
|
"learning_rate": 6.509090909090908e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2813,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.185546875,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 476
|
|
},
|
|
{
|
|
"epoch": 0.5852760736196319,
|
|
"grad_norm": 36.7408304979899,
|
|
"learning_rate": 6.505785123966942e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.257,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.75,
|
|
"step": 477
|
|
},
|
|
{
|
|
"epoch": 0.5865030674846625,
|
|
"grad_norm": 34.872893059723594,
|
|
"learning_rate": 6.502479338842975e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2988,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 478
|
|
},
|
|
{
|
|
"epoch": 0.5877300613496933,
|
|
"grad_norm": 35.26754811760802,
|
|
"learning_rate": 6.499173553719008e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2784,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.240234375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 479
|
|
},
|
|
{
|
|
"epoch": 0.588957055214724,
|
|
"grad_norm": 33.02148763705059,
|
|
"learning_rate": 6.495867768595041e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2364,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.4765625,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -6.125,
|
|
"step": 480
|
|
},
|
|
{
|
|
"epoch": 0.5901840490797546,
|
|
"grad_norm": 39.97994667673503,
|
|
"learning_rate": 6.492561983471074e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.3732,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 481
|
|
},
|
|
{
|
|
"epoch": 0.5914110429447853,
|
|
"grad_norm": 33.72401411710678,
|
|
"learning_rate": 6.489256198347108e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2739,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 482
|
|
},
|
|
{
|
|
"epoch": 0.592638036809816,
|
|
"grad_norm": 34.19291475160813,
|
|
"learning_rate": 6.485950413223141e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2917,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.69921875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 483
|
|
},
|
|
{
|
|
"epoch": 0.5938650306748466,
|
|
"grad_norm": 35.398942571150606,
|
|
"learning_rate": 6.482644628099173e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3155,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.875,
|
|
"step": 484
|
|
},
|
|
{
|
|
"epoch": 0.5950920245398773,
|
|
"grad_norm": 30.26872146503766,
|
|
"learning_rate": 6.479338842975206e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.26,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.77734375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 485
|
|
},
|
|
{
|
|
"epoch": 0.596319018404908,
|
|
"grad_norm": 35.38748335338782,
|
|
"learning_rate": 6.476033057851239e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2658,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 486
|
|
},
|
|
{
|
|
"epoch": 0.5975460122699386,
|
|
"grad_norm": 30.285435869887902,
|
|
"learning_rate": 6.472727272727272e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.233,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 487
|
|
},
|
|
{
|
|
"epoch": 0.5987730061349693,
|
|
"grad_norm": 37.89908939404138,
|
|
"learning_rate": 6.469421487603306e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2436,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 488
|
|
},
|
|
{
|
|
"epoch": 0.6,
|
|
"grad_norm": 30.515429632482274,
|
|
"learning_rate": 6.466115702479339e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.272,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -5.5,
|
|
"step": 489
|
|
},
|
|
{
|
|
"epoch": 0.6012269938650306,
|
|
"grad_norm": 40.37971138033072,
|
|
"learning_rate": 6.462809917355372e-07,
|
|
"logits/chosen": -0.09765625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3004,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 490
|
|
},
|
|
{
|
|
"epoch": 0.6024539877300613,
|
|
"grad_norm": 34.45143017996783,
|
|
"learning_rate": 6.459504132231405e-07,
|
|
"logits/chosen": -0.08203125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2549,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 491
|
|
},
|
|
{
|
|
"epoch": 0.603680981595092,
|
|
"grad_norm": 36.58265725949472,
|
|
"learning_rate": 6.456198347107438e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -378.0,
|
|
"loss": 0.2737,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 492
|
|
},
|
|
{
|
|
"epoch": 0.6049079754601226,
|
|
"grad_norm": 42.46481332966053,
|
|
"learning_rate": 6.45289256198347e-07,
|
|
"logits/chosen": -0.024169921875,
|
|
"logits/rejected": -0.134765625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3917,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 493
|
|
},
|
|
{
|
|
"epoch": 0.6061349693251534,
|
|
"grad_norm": 33.59404853921049,
|
|
"learning_rate": 6.449586776859504e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2643,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 494
|
|
},
|
|
{
|
|
"epoch": 0.6073619631901841,
|
|
"grad_norm": 37.484469870256234,
|
|
"learning_rate": 6.446280991735537e-07,
|
|
"logits/chosen": -0.1220703125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3066,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 495
|
|
},
|
|
{
|
|
"epoch": 0.6085889570552148,
|
|
"grad_norm": 40.63137683640751,
|
|
"learning_rate": 6.44297520661157e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3177,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 496
|
|
},
|
|
{
|
|
"epoch": 0.6098159509202454,
|
|
"grad_norm": 41.064298986990785,
|
|
"learning_rate": 6.439669421487603e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3865,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.5,
|
|
"step": 497
|
|
},
|
|
{
|
|
"epoch": 0.6110429447852761,
|
|
"grad_norm": 34.6148352569408,
|
|
"learning_rate": 6.436363636363636e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2817,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 498
|
|
},
|
|
{
|
|
"epoch": 0.6122699386503068,
|
|
"grad_norm": 36.60546803390198,
|
|
"learning_rate": 6.433057851239669e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2842,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 499
|
|
},
|
|
{
|
|
"epoch": 0.6134969325153374,
|
|
"grad_norm": 33.21675862161022,
|
|
"learning_rate": 6.429752066115703e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2718,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.6147239263803681,
|
|
"grad_norm": 32.544310987499145,
|
|
"learning_rate": 6.426446280991735e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.26,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 501
|
|
},
|
|
{
|
|
"epoch": 0.6159509202453988,
|
|
"grad_norm": 32.09170254227315,
|
|
"learning_rate": 6.423140495867768e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2639,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 502
|
|
},
|
|
{
|
|
"epoch": 0.6171779141104294,
|
|
"grad_norm": 35.53103605940553,
|
|
"learning_rate": 6.419834710743801e-07,
|
|
"logits/chosen": 0.01153564453125,
|
|
"logits/rejected": -0.1904296875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2795,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.25,
|
|
"step": 503
|
|
},
|
|
{
|
|
"epoch": 0.6184049079754601,
|
|
"grad_norm": 30.299429390734634,
|
|
"learning_rate": 6.416528925619834e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2658,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 504
|
|
},
|
|
{
|
|
"epoch": 0.6196319018404908,
|
|
"grad_norm": 44.689012762744575,
|
|
"learning_rate": 6.413223140495867e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -394.0,
|
|
"loss": 0.3776,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.369140625,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 505
|
|
},
|
|
{
|
|
"epoch": 0.6208588957055214,
|
|
"grad_norm": 32.33588195685234,
|
|
"learning_rate": 6.409917355371901e-07,
|
|
"logits/chosen": -0.08349609375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2588,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.875,
|
|
"step": 506
|
|
},
|
|
{
|
|
"epoch": 0.6220858895705521,
|
|
"grad_norm": 29.030680879470466,
|
|
"learning_rate": 6.406611570247934e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.1953,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 507
|
|
},
|
|
{
|
|
"epoch": 0.6233128834355828,
|
|
"grad_norm": 31.4626526695897,
|
|
"learning_rate": 6.403305785123967e-07,
|
|
"logits/chosen": -0.0201416015625,
|
|
"logits/rejected": -0.2021484375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.24,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 508
|
|
},
|
|
{
|
|
"epoch": 0.6245398773006134,
|
|
"grad_norm": 42.61520515559976,
|
|
"learning_rate": 6.4e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -466.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.3973,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.8984375,
|
|
"rewards/margins": 7.96875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 509
|
|
},
|
|
{
|
|
"epoch": 0.6257668711656442,
|
|
"grad_norm": 30.165060750168916,
|
|
"learning_rate": 6.396694214876032e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2226,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.306640625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.625,
|
|
"step": 510
|
|
},
|
|
{
|
|
"epoch": 0.6269938650306749,
|
|
"grad_norm": 40.15670022385489,
|
|
"learning_rate": 6.393388429752065e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3348,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.1875,
|
|
"step": 511
|
|
},
|
|
{
|
|
"epoch": 0.6282208588957056,
|
|
"grad_norm": 44.290408519253646,
|
|
"learning_rate": 6.390082644628099e-07,
|
|
"logits/chosen": -0.09130859375,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.3621,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.35546875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -5.0,
|
|
"step": 512
|
|
},
|
|
{
|
|
"epoch": 0.6294478527607362,
|
|
"grad_norm": 37.44250565144745,
|
|
"learning_rate": 6.386776859504132e-07,
|
|
"logits/chosen": -0.09521484375,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.337,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.375,
|
|
"step": 513
|
|
},
|
|
{
|
|
"epoch": 0.6306748466257669,
|
|
"grad_norm": 29.97245642529846,
|
|
"learning_rate": 6.383471074380165e-07,
|
|
"logits/chosen": -0.0830078125,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2436,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": -0.051513671875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 514
|
|
},
|
|
{
|
|
"epoch": 0.6319018404907976,
|
|
"grad_norm": 36.989875573277935,
|
|
"learning_rate": 6.380165289256198e-07,
|
|
"logits/chosen": -0.09912109375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2882,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.041748046875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 515
|
|
},
|
|
{
|
|
"epoch": 0.6331288343558282,
|
|
"grad_norm": 31.306859511596812,
|
|
"learning_rate": 6.376859504132231e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2432,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.1142578125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 516
|
|
},
|
|
{
|
|
"epoch": 0.6343558282208589,
|
|
"grad_norm": 43.67170294973452,
|
|
"learning_rate": 6.373553719008264e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3884,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.53125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 517
|
|
},
|
|
{
|
|
"epoch": 0.6355828220858896,
|
|
"grad_norm": 29.354877414703584,
|
|
"learning_rate": 6.370247933884298e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.1963,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 518
|
|
},
|
|
{
|
|
"epoch": 0.6368098159509202,
|
|
"grad_norm": 40.61854272265548,
|
|
"learning_rate": 6.36694214876033e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.299,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.75390625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.5,
|
|
"step": 519
|
|
},
|
|
{
|
|
"epoch": 0.6380368098159509,
|
|
"grad_norm": 38.052325639181014,
|
|
"learning_rate": 6.363636363636363e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3436,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 520
|
|
},
|
|
{
|
|
"epoch": 0.6392638036809816,
|
|
"grad_norm": 32.37002227141111,
|
|
"learning_rate": 6.360330578512396e-07,
|
|
"logits/chosen": -0.03271484375,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2525,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.5,
|
|
"step": 521
|
|
},
|
|
{
|
|
"epoch": 0.6404907975460122,
|
|
"grad_norm": 40.41274539784558,
|
|
"learning_rate": 6.357024793388429e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3152,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 522
|
|
},
|
|
{
|
|
"epoch": 0.6417177914110429,
|
|
"grad_norm": 42.30362208986331,
|
|
"learning_rate": 6.353719008264462e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3605,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.6171875,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 523
|
|
},
|
|
{
|
|
"epoch": 0.6429447852760736,
|
|
"grad_norm": 46.893532525519284,
|
|
"learning_rate": 6.350413223140496e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3361,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 524
|
|
},
|
|
{
|
|
"epoch": 0.6441717791411042,
|
|
"grad_norm": 49.422180518954264,
|
|
"learning_rate": 6.347107438016529e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.185546875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3927,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.498046875,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 525
|
|
},
|
|
{
|
|
"epoch": 0.645398773006135,
|
|
"grad_norm": 38.5530601675623,
|
|
"learning_rate": 6.343801652892562e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2504,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 526
|
|
},
|
|
{
|
|
"epoch": 0.6466257668711657,
|
|
"grad_norm": 35.94650372227145,
|
|
"learning_rate": 6.340495867768595e-07,
|
|
"logits/chosen": -0.0859375,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2888,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 527
|
|
},
|
|
{
|
|
"epoch": 0.6478527607361964,
|
|
"grad_norm": 33.57585022481078,
|
|
"learning_rate": 6.337190082644627e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2986,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 528
|
|
},
|
|
{
|
|
"epoch": 0.649079754601227,
|
|
"grad_norm": 35.863933883867006,
|
|
"learning_rate": 6.33388429752066e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3657,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 529
|
|
},
|
|
{
|
|
"epoch": 0.6503067484662577,
|
|
"grad_norm": 41.86945425639466,
|
|
"learning_rate": 6.330578512396694e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3197,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 530
|
|
},
|
|
{
|
|
"epoch": 0.6515337423312884,
|
|
"grad_norm": 33.3934976525138,
|
|
"learning_rate": 6.327272727272727e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2751,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 531
|
|
},
|
|
{
|
|
"epoch": 0.652760736196319,
|
|
"grad_norm": 35.72821914002383,
|
|
"learning_rate": 6.32396694214876e-07,
|
|
"logits/chosen": -0.07958984375,
|
|
"logits/rejected": -0.2197265625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2782,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.875,
|
|
"step": 532
|
|
},
|
|
{
|
|
"epoch": 0.6539877300613497,
|
|
"grad_norm": 36.05732236551947,
|
|
"learning_rate": 6.320661157024793e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2611,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 533
|
|
},
|
|
{
|
|
"epoch": 0.6552147239263804,
|
|
"grad_norm": 43.004226294889776,
|
|
"learning_rate": 6.317355371900826e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2813,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 534
|
|
},
|
|
{
|
|
"epoch": 0.656441717791411,
|
|
"grad_norm": 39.5682500491772,
|
|
"learning_rate": 6.31404958677686e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2961,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.380859375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 535
|
|
},
|
|
{
|
|
"epoch": 0.6576687116564417,
|
|
"grad_norm": 36.55745143954986,
|
|
"learning_rate": 6.310743801652893e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2275,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 536
|
|
},
|
|
{
|
|
"epoch": 0.6588957055214724,
|
|
"grad_norm": 45.10266990819029,
|
|
"learning_rate": 6.307438016528925e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3939,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.265625,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 537
|
|
},
|
|
{
|
|
"epoch": 0.660122699386503,
|
|
"grad_norm": 44.09458580737532,
|
|
"learning_rate": 6.304132231404958e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3783,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 538
|
|
},
|
|
{
|
|
"epoch": 0.6613496932515337,
|
|
"grad_norm": 36.16672414551654,
|
|
"learning_rate": 6.300826446280991e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.309,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 539
|
|
},
|
|
{
|
|
"epoch": 0.6625766871165644,
|
|
"grad_norm": 39.077710779231424,
|
|
"learning_rate": 6.297520661157024e-07,
|
|
"logits/chosen": -0.0908203125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3217,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 540
|
|
},
|
|
{
|
|
"epoch": 0.6638036809815951,
|
|
"grad_norm": 35.971169171679556,
|
|
"learning_rate": 6.294214876033058e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.2080078125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2834,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.75390625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 541
|
|
},
|
|
{
|
|
"epoch": 0.6650306748466258,
|
|
"grad_norm": 34.168657597082216,
|
|
"learning_rate": 6.290909090909091e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2812,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 542
|
|
},
|
|
{
|
|
"epoch": 0.6662576687116565,
|
|
"grad_norm": 36.23729440291612,
|
|
"learning_rate": 6.287603305785124e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2622,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 543
|
|
},
|
|
{
|
|
"epoch": 0.6674846625766871,
|
|
"grad_norm": 45.238447000820585,
|
|
"learning_rate": 6.284297520661157e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3695,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.25,
|
|
"step": 544
|
|
},
|
|
{
|
|
"epoch": 0.6687116564417178,
|
|
"grad_norm": 43.061817496062964,
|
|
"learning_rate": 6.28099173553719e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.4047,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 545
|
|
},
|
|
{
|
|
"epoch": 0.6699386503067485,
|
|
"grad_norm": 40.8862609592416,
|
|
"learning_rate": 6.277685950413222e-07,
|
|
"logits/chosen": -0.10205078125,
|
|
"logits/rejected": -0.20703125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3402,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 546
|
|
},
|
|
{
|
|
"epoch": 0.6711656441717792,
|
|
"grad_norm": 32.99712535949366,
|
|
"learning_rate": 6.274380165289256e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2883,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 547
|
|
},
|
|
{
|
|
"epoch": 0.6723926380368098,
|
|
"grad_norm": 40.33054062017028,
|
|
"learning_rate": 6.271074380165289e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2583,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 548
|
|
},
|
|
{
|
|
"epoch": 0.6736196319018405,
|
|
"grad_norm": 37.72507019090557,
|
|
"learning_rate": 6.267768595041322e-07,
|
|
"logits/chosen": -0.11474609375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 549
|
|
},
|
|
{
|
|
"epoch": 0.6748466257668712,
|
|
"grad_norm": 37.30173963897267,
|
|
"learning_rate": 6.264462809917355e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3323,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 0.6760736196319018,
|
|
"grad_norm": 33.242769441390706,
|
|
"learning_rate": 6.261157024793388e-07,
|
|
"logits/chosen": -0.0986328125,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2814,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9296875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 551
|
|
},
|
|
{
|
|
"epoch": 0.6773006134969325,
|
|
"grad_norm": 35.53857403218536,
|
|
"learning_rate": 6.257851239669421e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3055,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 552
|
|
},
|
|
{
|
|
"epoch": 0.6785276073619632,
|
|
"grad_norm": 39.8186178694702,
|
|
"learning_rate": 6.254545454545455e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3556,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 553
|
|
},
|
|
{
|
|
"epoch": 0.6797546012269938,
|
|
"grad_norm": 34.396570445757696,
|
|
"learning_rate": 6.251239669421488e-07,
|
|
"logits/chosen": -0.08349609375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.261,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 554
|
|
},
|
|
{
|
|
"epoch": 0.6809815950920245,
|
|
"grad_norm": 37.46750845748253,
|
|
"learning_rate": 6.24793388429752e-07,
|
|
"logits/chosen": 0.0196533203125,
|
|
"logits/rejected": -0.107421875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3763,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 555
|
|
},
|
|
{
|
|
"epoch": 0.6822085889570552,
|
|
"grad_norm": 28.75095679008507,
|
|
"learning_rate": 6.244628099173553e-07,
|
|
"logits/chosen": -0.10693359375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2294,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 556
|
|
},
|
|
{
|
|
"epoch": 0.6834355828220859,
|
|
"grad_norm": 32.12117772558378,
|
|
"learning_rate": 6.241322314049586e-07,
|
|
"logits/chosen": -0.11669921875,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2775,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 557
|
|
},
|
|
{
|
|
"epoch": 0.6846625766871166,
|
|
"grad_norm": 30.45940478111348,
|
|
"learning_rate": 6.238016528925619e-07,
|
|
"logits/chosen": -0.07861328125,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2442,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.8359375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.875,
|
|
"step": 558
|
|
},
|
|
{
|
|
"epoch": 0.6858895705521473,
|
|
"grad_norm": 38.28759745643993,
|
|
"learning_rate": 6.234710743801653e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3234,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.25,
|
|
"step": 559
|
|
},
|
|
{
|
|
"epoch": 0.6871165644171779,
|
|
"grad_norm": 25.47174077723704,
|
|
"learning_rate": 6.231404958677686e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2016,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 560
|
|
},
|
|
{
|
|
"epoch": 0.6883435582822086,
|
|
"grad_norm": 28.61531386213562,
|
|
"learning_rate": 6.228099173553719e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2021,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 561
|
|
},
|
|
{
|
|
"epoch": 0.6895705521472393,
|
|
"grad_norm": 30.528229373220828,
|
|
"learning_rate": 6.224793388429752e-07,
|
|
"logits/chosen": -0.032470703125,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2462,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.5,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 562
|
|
},
|
|
{
|
|
"epoch": 0.69079754601227,
|
|
"grad_norm": 34.031991776353316,
|
|
"learning_rate": 6.221487603305785e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2904,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 563
|
|
},
|
|
{
|
|
"epoch": 0.6920245398773006,
|
|
"grad_norm": 24.836619890425215,
|
|
"learning_rate": 6.218181818181817e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.224,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.7109375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 564
|
|
},
|
|
{
|
|
"epoch": 0.6932515337423313,
|
|
"grad_norm": 35.09286389818353,
|
|
"learning_rate": 6.214876033057851e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3639,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 565
|
|
},
|
|
{
|
|
"epoch": 0.694478527607362,
|
|
"grad_norm": 33.71563020260373,
|
|
"learning_rate": 6.211570247933884e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2462,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 566
|
|
},
|
|
{
|
|
"epoch": 0.6957055214723926,
|
|
"grad_norm": 25.84647564741135,
|
|
"learning_rate": 6.208264462809917e-07,
|
|
"logits/chosen": -0.053955078125,
|
|
"logits/rejected": -0.1875,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2061,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.359375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 567
|
|
},
|
|
{
|
|
"epoch": 0.6969325153374233,
|
|
"grad_norm": 37.567177147049684,
|
|
"learning_rate": 6.20495867768595e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3064,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 568
|
|
},
|
|
{
|
|
"epoch": 0.698159509202454,
|
|
"grad_norm": 32.35117698678835,
|
|
"learning_rate": 6.201652892561983e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2508,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 569
|
|
},
|
|
{
|
|
"epoch": 0.6993865030674846,
|
|
"grad_norm": 25.32334987337483,
|
|
"learning_rate": 6.198347107438017e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.1649,
|
|
"rewards/accuracies": 0.9453125,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 570
|
|
},
|
|
{
|
|
"epoch": 0.7006134969325153,
|
|
"grad_norm": 31.066523168854104,
|
|
"learning_rate": 6.19504132231405e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2262,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.6015625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 571
|
|
},
|
|
{
|
|
"epoch": 0.701840490797546,
|
|
"grad_norm": 32.881821449548816,
|
|
"learning_rate": 6.191735537190083e-07,
|
|
"logits/chosen": -0.1142578125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2291,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 572
|
|
},
|
|
{
|
|
"epoch": 0.7030674846625767,
|
|
"grad_norm": 40.68529761937124,
|
|
"learning_rate": 6.188429752066115e-07,
|
|
"logits/chosen": -0.07177734375,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4034,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 573
|
|
},
|
|
{
|
|
"epoch": 0.7042944785276074,
|
|
"grad_norm": 38.820291986225,
|
|
"learning_rate": 6.185123966942148e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2792,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.25,
|
|
"step": 574
|
|
},
|
|
{
|
|
"epoch": 0.7055214723926381,
|
|
"grad_norm": 33.59717634578736,
|
|
"learning_rate": 6.181818181818181e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2284,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 575
|
|
},
|
|
{
|
|
"epoch": 0.7067484662576687,
|
|
"grad_norm": 28.04916589516562,
|
|
"learning_rate": 6.178512396694215e-07,
|
|
"logits/chosen": -0.10302734375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2001,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -5.875,
|
|
"step": 576
|
|
},
|
|
{
|
|
"epoch": 0.7079754601226994,
|
|
"grad_norm": 36.15869255614372,
|
|
"learning_rate": 6.175206611570248e-07,
|
|
"logits/chosen": -0.03955078125,
|
|
"logits/rejected": -0.1767578125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2929,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 577
|
|
},
|
|
{
|
|
"epoch": 0.7092024539877301,
|
|
"grad_norm": 36.89994352340565,
|
|
"learning_rate": 6.171900826446281e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.322,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 7.40625,
|
|
"rewards/rejected": -5.75,
|
|
"step": 578
|
|
},
|
|
{
|
|
"epoch": 0.7104294478527607,
|
|
"grad_norm": 42.55492037495249,
|
|
"learning_rate": 6.168595041322314e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3864,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.451171875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 579
|
|
},
|
|
{
|
|
"epoch": 0.7116564417177914,
|
|
"grad_norm": 35.02868013813214,
|
|
"learning_rate": 6.165289256198347e-07,
|
|
"logits/chosen": -0.07275390625,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.229,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.73046875,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -6.25,
|
|
"step": 580
|
|
},
|
|
{
|
|
"epoch": 0.7128834355828221,
|
|
"grad_norm": 29.748844282456343,
|
|
"learning_rate": 6.16198347107438e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2071,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 581
|
|
},
|
|
{
|
|
"epoch": 0.7141104294478527,
|
|
"grad_norm": 41.261123421442086,
|
|
"learning_rate": 6.158677685950413e-07,
|
|
"logits/chosen": 0.01953125,
|
|
"logits/rejected": -0.197265625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3132,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.9453125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 582
|
|
},
|
|
{
|
|
"epoch": 0.7153374233128834,
|
|
"grad_norm": 37.14443132697755,
|
|
"learning_rate": 6.155371900826446e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3036,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.875,
|
|
"step": 583
|
|
},
|
|
{
|
|
"epoch": 0.7165644171779141,
|
|
"grad_norm": 42.810461239623855,
|
|
"learning_rate": 6.152066115702479e-07,
|
|
"logits/chosen": -0.0693359375,
|
|
"logits/rejected": -0.14453125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3929,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 584
|
|
},
|
|
{
|
|
"epoch": 0.7177914110429447,
|
|
"grad_norm": 44.860424166016166,
|
|
"learning_rate": 6.148760330578512e-07,
|
|
"logits/chosen": 0.0206298828125,
|
|
"logits/rejected": -0.099609375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.4253,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 585
|
|
},
|
|
{
|
|
"epoch": 0.7190184049079754,
|
|
"grad_norm": 36.65848945690881,
|
|
"learning_rate": 6.145454545454545e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.2158203125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3226,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 586
|
|
},
|
|
{
|
|
"epoch": 0.7202453987730061,
|
|
"grad_norm": 34.66555869797275,
|
|
"learning_rate": 6.142148760330578e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.319,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.5,
|
|
"step": 587
|
|
},
|
|
{
|
|
"epoch": 0.7214723926380369,
|
|
"grad_norm": 25.81313251350675,
|
|
"learning_rate": 6.138842975206612e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2328,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 588
|
|
},
|
|
{
|
|
"epoch": 0.7226993865030675,
|
|
"grad_norm": 39.45475953492192,
|
|
"learning_rate": 6.135537190082645e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3428,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.7734375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 589
|
|
},
|
|
{
|
|
"epoch": 0.7239263803680982,
|
|
"grad_norm": 28.739570926494295,
|
|
"learning_rate": 6.132231404958677e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2629,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 590
|
|
},
|
|
{
|
|
"epoch": 0.7251533742331289,
|
|
"grad_norm": 39.542558184509424,
|
|
"learning_rate": 6.12892561983471e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2455,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 591
|
|
},
|
|
{
|
|
"epoch": 0.7263803680981595,
|
|
"grad_norm": 31.700565699717185,
|
|
"learning_rate": 6.125619834710743e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2372,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 592
|
|
},
|
|
{
|
|
"epoch": 0.7276073619631902,
|
|
"grad_norm": 43.160388567583965,
|
|
"learning_rate": 6.122314049586776e-07,
|
|
"logits/chosen": -0.08251953125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3556,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 593
|
|
},
|
|
{
|
|
"epoch": 0.7288343558282209,
|
|
"grad_norm": 33.17336982955441,
|
|
"learning_rate": 6.11900826446281e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2737,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.77734375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 594
|
|
},
|
|
{
|
|
"epoch": 0.7300613496932515,
|
|
"grad_norm": 35.78921531146286,
|
|
"learning_rate": 6.115702479338843e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2188,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 595
|
|
},
|
|
{
|
|
"epoch": 0.7312883435582822,
|
|
"grad_norm": 39.132566907776244,
|
|
"learning_rate": 6.112396694214876e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3258,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 596
|
|
},
|
|
{
|
|
"epoch": 0.7325153374233129,
|
|
"grad_norm": 36.16852227530472,
|
|
"learning_rate": 6.109090909090909e-07,
|
|
"logits/chosen": -0.03515625,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3225,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 597
|
|
},
|
|
{
|
|
"epoch": 0.7337423312883435,
|
|
"grad_norm": 29.88526191086389,
|
|
"learning_rate": 6.105785123966942e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.204,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -5.125,
|
|
"step": 598
|
|
},
|
|
{
|
|
"epoch": 0.7349693251533742,
|
|
"grad_norm": 32.885582252909025,
|
|
"learning_rate": 6.102479338842974e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2847,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 599
|
|
},
|
|
{
|
|
"epoch": 0.7361963190184049,
|
|
"grad_norm": 37.722786017646094,
|
|
"learning_rate": 6.099173553719008e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.368,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.7374233128834355,
|
|
"grad_norm": 31.64488245915107,
|
|
"learning_rate": 6.095867768595041e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2698,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.1826171875,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 601
|
|
},
|
|
{
|
|
"epoch": 0.7386503067484662,
|
|
"grad_norm": 25.124092304435703,
|
|
"learning_rate": 6.092561983471074e-07,
|
|
"logits/chosen": -0.0076904296875,
|
|
"logits/rejected": -0.1708984375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.189,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 602
|
|
},
|
|
{
|
|
"epoch": 0.7398773006134969,
|
|
"grad_norm": 30.45546682074925,
|
|
"learning_rate": 6.089256198347107e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2864,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.625,
|
|
"step": 603
|
|
},
|
|
{
|
|
"epoch": 0.7411042944785277,
|
|
"grad_norm": 31.94436328392251,
|
|
"learning_rate": 6.08595041322314e-07,
|
|
"logits/chosen": -0.2197265625,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2464,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 604
|
|
},
|
|
{
|
|
"epoch": 0.7423312883435583,
|
|
"grad_norm": 33.65745564504924,
|
|
"learning_rate": 6.082644628099173e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2598,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 605
|
|
},
|
|
{
|
|
"epoch": 0.743558282208589,
|
|
"grad_norm": 38.63161562009038,
|
|
"learning_rate": 6.079338842975207e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3454,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 606
|
|
},
|
|
{
|
|
"epoch": 0.7447852760736197,
|
|
"grad_norm": 33.230355625984764,
|
|
"learning_rate": 6.07603305785124e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2611,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.48828125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -6.40625,
|
|
"step": 607
|
|
},
|
|
{
|
|
"epoch": 0.7460122699386503,
|
|
"grad_norm": 33.50647832359134,
|
|
"learning_rate": 6.072727272727272e-07,
|
|
"logits/chosen": -0.0216064453125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2913,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.099609375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 608
|
|
},
|
|
{
|
|
"epoch": 0.747239263803681,
|
|
"grad_norm": 35.459553964773974,
|
|
"learning_rate": 6.069421487603305e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2856,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.54296875,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -6.21875,
|
|
"step": 609
|
|
},
|
|
{
|
|
"epoch": 0.7484662576687117,
|
|
"grad_norm": 31.662560042647854,
|
|
"learning_rate": 6.066115702479338e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.2452,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 7.46875,
|
|
"rewards/rejected": -6.375,
|
|
"step": 610
|
|
},
|
|
{
|
|
"epoch": 0.7496932515337423,
|
|
"grad_norm": 49.54193875289406,
|
|
"learning_rate": 6.062809917355371e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.3976,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": -0.1396484375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -6.0,
|
|
"step": 611
|
|
},
|
|
{
|
|
"epoch": 0.750920245398773,
|
|
"grad_norm": 39.062720193674714,
|
|
"learning_rate": 6.059504132231405e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3777,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.546875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 612
|
|
},
|
|
{
|
|
"epoch": 0.7521472392638037,
|
|
"grad_norm": 35.9253243800041,
|
|
"learning_rate": 6.056198347107438e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.2216796875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3394,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.10107421875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 613
|
|
},
|
|
{
|
|
"epoch": 0.7533742331288343,
|
|
"grad_norm": 54.139929990418025,
|
|
"learning_rate": 6.052892561983471e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3142,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 614
|
|
},
|
|
{
|
|
"epoch": 0.754601226993865,
|
|
"grad_norm": 37.247633297004576,
|
|
"learning_rate": 6.049586776859504e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3089,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -6.0,
|
|
"step": 615
|
|
},
|
|
{
|
|
"epoch": 0.7558282208588957,
|
|
"grad_norm": 31.187596328354086,
|
|
"learning_rate": 6.046280991735537e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2528,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.10302734375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 616
|
|
},
|
|
{
|
|
"epoch": 0.7570552147239263,
|
|
"grad_norm": 39.842219715736135,
|
|
"learning_rate": 6.04297520661157e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.208984375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3542,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.48828125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 617
|
|
},
|
|
{
|
|
"epoch": 0.758282208588957,
|
|
"grad_norm": 43.96609425022703,
|
|
"learning_rate": 6.039669421487603e-07,
|
|
"logits/chosen": -0.052734375,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3723,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.51953125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 618
|
|
},
|
|
{
|
|
"epoch": 0.7595092024539877,
|
|
"grad_norm": 32.587670722826026,
|
|
"learning_rate": 6.036363636363636e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2331,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.90625,
|
|
"step": 619
|
|
},
|
|
{
|
|
"epoch": 0.7607361963190185,
|
|
"grad_norm": 44.41642388280846,
|
|
"learning_rate": 6.033057851239669e-07,
|
|
"logits/chosen": -0.0830078125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3088,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.5234375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 620
|
|
},
|
|
{
|
|
"epoch": 0.7619631901840491,
|
|
"grad_norm": 34.02108964551065,
|
|
"learning_rate": 6.029752066115702e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -478.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.2661,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 7.3125,
|
|
"rewards/rejected": -6.125,
|
|
"step": 621
|
|
},
|
|
{
|
|
"epoch": 0.7631901840490798,
|
|
"grad_norm": 31.832438088278238,
|
|
"learning_rate": 6.026446280991735e-07,
|
|
"logits/chosen": -0.0888671875,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.236,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.72265625,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 622
|
|
},
|
|
{
|
|
"epoch": 0.7644171779141105,
|
|
"grad_norm": 38.89511615880052,
|
|
"learning_rate": 6.023140495867769e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.3167,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 623
|
|
},
|
|
{
|
|
"epoch": 0.7656441717791411,
|
|
"grad_norm": 32.19915482428781,
|
|
"learning_rate": 6.019834710743802e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2307,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 7.78125,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 624
|
|
},
|
|
{
|
|
"epoch": 0.7668711656441718,
|
|
"grad_norm": 32.198949298871604,
|
|
"learning_rate": 6.016528925619835e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2191,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 625
|
|
},
|
|
{
|
|
"epoch": 0.7680981595092025,
|
|
"grad_norm": 40.391116665536885,
|
|
"learning_rate": 6.013223140495867e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3346,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 626
|
|
},
|
|
{
|
|
"epoch": 0.7693251533742331,
|
|
"grad_norm": 33.50213686948866,
|
|
"learning_rate": 6.0099173553719e-07,
|
|
"logits/chosen": -0.06396484375,
|
|
"logits/rejected": -0.154296875,
|
|
"logps/chosen": -310.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3297,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": -0.1552734375,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 627
|
|
},
|
|
{
|
|
"epoch": 0.7705521472392638,
|
|
"grad_norm": 37.25258333430096,
|
|
"learning_rate": 6.006611570247933e-07,
|
|
"logits/chosen": -0.0947265625,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2886,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 628
|
|
},
|
|
{
|
|
"epoch": 0.7717791411042945,
|
|
"grad_norm": 23.68892864954692,
|
|
"learning_rate": 6.003305785123967e-07,
|
|
"logits/chosen": -0.1220703125,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.1789,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.375,
|
|
"step": 629
|
|
},
|
|
{
|
|
"epoch": 0.7730061349693251,
|
|
"grad_norm": 33.10208379218177,
|
|
"learning_rate": 6e-07,
|
|
"logits/chosen": -0.060546875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2663,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 630
|
|
},
|
|
{
|
|
"epoch": 0.7742331288343558,
|
|
"grad_norm": 36.528138511875646,
|
|
"learning_rate": 5.996694214876033e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3321,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.875,
|
|
"step": 631
|
|
},
|
|
{
|
|
"epoch": 0.7754601226993865,
|
|
"grad_norm": 36.37506860284647,
|
|
"learning_rate": 5.993388429752066e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3056,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 632
|
|
},
|
|
{
|
|
"epoch": 0.7766871165644171,
|
|
"grad_norm": 40.1484232566858,
|
|
"learning_rate": 5.990082644628099e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.21484375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3418,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 633
|
|
},
|
|
{
|
|
"epoch": 0.7779141104294478,
|
|
"grad_norm": 32.83617762178678,
|
|
"learning_rate": 5.986776859504132e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2202,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 2.734375,
|
|
"rewards/margins": 7.9375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 634
|
|
},
|
|
{
|
|
"epoch": 0.7791411042944786,
|
|
"grad_norm": 29.099912658258905,
|
|
"learning_rate": 5.983471074380165e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2119,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 635
|
|
},
|
|
{
|
|
"epoch": 0.7803680981595092,
|
|
"grad_norm": 39.54581126371982,
|
|
"learning_rate": 5.980165289256198e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3099,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 636
|
|
},
|
|
{
|
|
"epoch": 0.7815950920245399,
|
|
"grad_norm": 45.109718146173684,
|
|
"learning_rate": 5.976859504132231e-07,
|
|
"logits/chosen": -0.03125,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3522,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 637
|
|
},
|
|
{
|
|
"epoch": 0.7828220858895706,
|
|
"grad_norm": 29.717665569657886,
|
|
"learning_rate": 5.973553719008264e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2213,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 638
|
|
},
|
|
{
|
|
"epoch": 0.7840490797546013,
|
|
"grad_norm": 38.75212680036792,
|
|
"learning_rate": 5.970247933884297e-07,
|
|
"logits/chosen": -0.048828125,
|
|
"logits/rejected": -0.2265625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3316,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 639
|
|
},
|
|
{
|
|
"epoch": 0.7852760736196319,
|
|
"grad_norm": 39.02043679792906,
|
|
"learning_rate": 5.96694214876033e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -382.0,
|
|
"loss": 0.2894,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 640
|
|
},
|
|
{
|
|
"epoch": 0.7865030674846626,
|
|
"grad_norm": 44.494628054471185,
|
|
"learning_rate": 5.963636363636364e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.4553,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 641
|
|
},
|
|
{
|
|
"epoch": 0.7877300613496933,
|
|
"grad_norm": 27.200186618621586,
|
|
"learning_rate": 5.960330578512397e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2487,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 642
|
|
},
|
|
{
|
|
"epoch": 0.7889570552147239,
|
|
"grad_norm": 36.983988342062,
|
|
"learning_rate": 5.95702479338843e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.2631,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 643
|
|
},
|
|
{
|
|
"epoch": 0.7901840490797546,
|
|
"grad_norm": 39.431613095622076,
|
|
"learning_rate": 5.953719008264462e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2856,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 644
|
|
},
|
|
{
|
|
"epoch": 0.7914110429447853,
|
|
"grad_norm": 33.34429063414653,
|
|
"learning_rate": 5.950413223140495e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2445,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 645
|
|
},
|
|
{
|
|
"epoch": 0.7926380368098159,
|
|
"grad_norm": 34.57872089054598,
|
|
"learning_rate": 5.947107438016528e-07,
|
|
"logits/chosen": -0.0859375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2881,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 646
|
|
},
|
|
{
|
|
"epoch": 0.7938650306748466,
|
|
"grad_norm": 34.439204602754096,
|
|
"learning_rate": 5.943801652892562e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2819,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 647
|
|
},
|
|
{
|
|
"epoch": 0.7950920245398773,
|
|
"grad_norm": 41.50373217977521,
|
|
"learning_rate": 5.940495867768595e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3682,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 648
|
|
},
|
|
{
|
|
"epoch": 0.7963190184049079,
|
|
"grad_norm": 43.045318488956525,
|
|
"learning_rate": 5.937190082644628e-07,
|
|
"logits/chosen": -0.0654296875,
|
|
"logits/rejected": -0.1650390625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.491,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 649
|
|
},
|
|
{
|
|
"epoch": 0.7975460122699386,
|
|
"grad_norm": 41.49765653672477,
|
|
"learning_rate": 5.933884297520661e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3786,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 0.7987730061349694,
|
|
"grad_norm": 38.92619113627064,
|
|
"learning_rate": 5.930578512396694e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2788,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 651
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"grad_norm": 40.314158512080596,
|
|
"learning_rate": 5.927272727272728e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3015,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 652
|
|
},
|
|
{
|
|
"epoch": 0.8012269938650307,
|
|
"grad_norm": 41.91917001542575,
|
|
"learning_rate": 5.92396694214876e-07,
|
|
"logits/chosen": -0.103515625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2786,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 653
|
|
},
|
|
{
|
|
"epoch": 0.8024539877300614,
|
|
"grad_norm": 32.50826629622854,
|
|
"learning_rate": 5.920661157024793e-07,
|
|
"logits/chosen": -0.083984375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2325,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 654
|
|
},
|
|
{
|
|
"epoch": 0.803680981595092,
|
|
"grad_norm": 36.66838919465664,
|
|
"learning_rate": 5.917355371900826e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2743,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 655
|
|
},
|
|
{
|
|
"epoch": 0.8049079754601227,
|
|
"grad_norm": 40.353892524687666,
|
|
"learning_rate": 5.914049586776859e-07,
|
|
"logits/chosen": -0.080078125,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3542,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 656
|
|
},
|
|
{
|
|
"epoch": 0.8061349693251534,
|
|
"grad_norm": 32.71653938190321,
|
|
"learning_rate": 5.910743801652892e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3238,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 657
|
|
},
|
|
{
|
|
"epoch": 0.807361963190184,
|
|
"grad_norm": 35.63490892300594,
|
|
"learning_rate": 5.907438016528926e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2672,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 658
|
|
},
|
|
{
|
|
"epoch": 0.8085889570552147,
|
|
"grad_norm": 36.34676950346043,
|
|
"learning_rate": 5.904132231404959e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3323,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.2177734375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 659
|
|
},
|
|
{
|
|
"epoch": 0.8098159509202454,
|
|
"grad_norm": 31.25262331885713,
|
|
"learning_rate": 5.900826446280992e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2661,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 660
|
|
},
|
|
{
|
|
"epoch": 0.811042944785276,
|
|
"grad_norm": 44.04724360514248,
|
|
"learning_rate": 5.897520661157025e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3716,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.3125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 661
|
|
},
|
|
{
|
|
"epoch": 0.8122699386503067,
|
|
"grad_norm": 28.306859195207647,
|
|
"learning_rate": 5.894214876033057e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2117,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 662
|
|
},
|
|
{
|
|
"epoch": 0.8134969325153374,
|
|
"grad_norm": 28.755676702838585,
|
|
"learning_rate": 5.89090909090909e-07,
|
|
"logits/chosen": -0.06689453125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2615,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 663
|
|
},
|
|
{
|
|
"epoch": 0.8147239263803681,
|
|
"grad_norm": 30.20251070806522,
|
|
"learning_rate": 5.887603305785124e-07,
|
|
"logits/chosen": -0.349609375,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.1821,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 7.84375,
|
|
"rewards/rejected": -6.625,
|
|
"step": 664
|
|
},
|
|
{
|
|
"epoch": 0.8159509202453987,
|
|
"grad_norm": 43.59493072844384,
|
|
"learning_rate": 5.884297520661157e-07,
|
|
"logits/chosen": -0.05029296875,
|
|
"logits/rejected": -0.2197265625,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2979,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.5234375,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 665
|
|
},
|
|
{
|
|
"epoch": 0.8171779141104294,
|
|
"grad_norm": 42.64325196007558,
|
|
"learning_rate": 5.88099173553719e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.368,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 666
|
|
},
|
|
{
|
|
"epoch": 0.8184049079754602,
|
|
"grad_norm": 34.27649786728366,
|
|
"learning_rate": 5.877685950413223e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2601,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.75,
|
|
"step": 667
|
|
},
|
|
{
|
|
"epoch": 0.8196319018404908,
|
|
"grad_norm": 33.27456675278737,
|
|
"learning_rate": 5.874380165289256e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2841,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 668
|
|
},
|
|
{
|
|
"epoch": 0.8208588957055215,
|
|
"grad_norm": 41.4043511264187,
|
|
"learning_rate": 5.871074380165289e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3017,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.375,
|
|
"step": 669
|
|
},
|
|
{
|
|
"epoch": 0.8220858895705522,
|
|
"grad_norm": 31.799223551777082,
|
|
"learning_rate": 5.867768595041323e-07,
|
|
"logits/chosen": -0.0191650390625,
|
|
"logits/rejected": -0.1650390625,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2541,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.125,
|
|
"step": 670
|
|
},
|
|
{
|
|
"epoch": 0.8233128834355828,
|
|
"grad_norm": 31.913109572694793,
|
|
"learning_rate": 5.864462809917355e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2407,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 671
|
|
},
|
|
{
|
|
"epoch": 0.8245398773006135,
|
|
"grad_norm": 32.62700830171627,
|
|
"learning_rate": 5.861157024793388e-07,
|
|
"logits/chosen": -0.109375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2681,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 7.25,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 672
|
|
},
|
|
{
|
|
"epoch": 0.8257668711656442,
|
|
"grad_norm": 36.66995682138012,
|
|
"learning_rate": 5.857851239669421e-07,
|
|
"logits/chosen": -0.0498046875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2592,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 673
|
|
},
|
|
{
|
|
"epoch": 0.8269938650306748,
|
|
"grad_norm": 34.606200421728126,
|
|
"learning_rate": 5.854545454545454e-07,
|
|
"logits/chosen": -0.01300048828125,
|
|
"logits/rejected": -0.177734375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2822,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 674
|
|
},
|
|
{
|
|
"epoch": 0.8282208588957055,
|
|
"grad_norm": 35.038586845163366,
|
|
"learning_rate": 5.851239669421487e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3168,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 675
|
|
},
|
|
{
|
|
"epoch": 0.8294478527607362,
|
|
"grad_norm": 35.465018732105385,
|
|
"learning_rate": 5.847933884297521e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2534,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.71875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 676
|
|
},
|
|
{
|
|
"epoch": 0.8306748466257668,
|
|
"grad_norm": 36.537944400776084,
|
|
"learning_rate": 5.844628099173554e-07,
|
|
"logits/chosen": -0.06982421875,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2916,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.52734375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 677
|
|
},
|
|
{
|
|
"epoch": 0.8319018404907975,
|
|
"grad_norm": 33.168213995988694,
|
|
"learning_rate": 5.841322314049587e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2768,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.25,
|
|
"step": 678
|
|
},
|
|
{
|
|
"epoch": 0.8331288343558282,
|
|
"grad_norm": 37.74716831610831,
|
|
"learning_rate": 5.838016528925619e-07,
|
|
"logits/chosen": -0.01220703125,
|
|
"logits/rejected": -0.1259765625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3364,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 679
|
|
},
|
|
{
|
|
"epoch": 0.8343558282208589,
|
|
"grad_norm": 34.79631984317766,
|
|
"learning_rate": 5.834710743801652e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2698,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 680
|
|
},
|
|
{
|
|
"epoch": 0.8355828220858895,
|
|
"grad_norm": 27.30848566312445,
|
|
"learning_rate": 5.831404958677685e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.1854,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.25,
|
|
"step": 681
|
|
},
|
|
{
|
|
"epoch": 0.8368098159509203,
|
|
"grad_norm": 36.85289178579729,
|
|
"learning_rate": 5.828099173553719e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3333,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 682
|
|
},
|
|
{
|
|
"epoch": 0.838036809815951,
|
|
"grad_norm": 33.68839599065418,
|
|
"learning_rate": 5.824793388429752e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2767,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 683
|
|
},
|
|
{
|
|
"epoch": 0.8392638036809816,
|
|
"grad_norm": 35.217179760941754,
|
|
"learning_rate": 5.821487603305785e-07,
|
|
"logits/chosen": 0.000904083251953125,
|
|
"logits/rejected": -0.1181640625,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3304,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": -0.0341796875,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 684
|
|
},
|
|
{
|
|
"epoch": 0.8404907975460123,
|
|
"grad_norm": 34.3907812005455,
|
|
"learning_rate": 5.818181818181818e-07,
|
|
"logits/chosen": -0.11865234375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2858,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -6.59375,
|
|
"step": 685
|
|
},
|
|
{
|
|
"epoch": 0.841717791411043,
|
|
"grad_norm": 29.480838482306872,
|
|
"learning_rate": 5.814876033057851e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2598,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 686
|
|
},
|
|
{
|
|
"epoch": 0.8429447852760736,
|
|
"grad_norm": 34.28684664510481,
|
|
"learning_rate": 5.811570247933884e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2709,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 687
|
|
},
|
|
{
|
|
"epoch": 0.8441717791411043,
|
|
"grad_norm": 35.903679070126266,
|
|
"learning_rate": 5.808264462809917e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2748,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.07958984375,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -6.125,
|
|
"step": 688
|
|
},
|
|
{
|
|
"epoch": 0.845398773006135,
|
|
"grad_norm": 37.11450186775036,
|
|
"learning_rate": 5.80495867768595e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.2197265625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3107,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.375,
|
|
"step": 689
|
|
},
|
|
{
|
|
"epoch": 0.8466257668711656,
|
|
"grad_norm": 49.86231404884124,
|
|
"learning_rate": 5.801652892561983e-07,
|
|
"logits/chosen": 0.004058837890625,
|
|
"logits/rejected": -0.1943359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.4457,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 690
|
|
},
|
|
{
|
|
"epoch": 0.8478527607361963,
|
|
"grad_norm": 30.320356705626047,
|
|
"learning_rate": 5.798347107438016e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2435,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 691
|
|
},
|
|
{
|
|
"epoch": 0.849079754601227,
|
|
"grad_norm": 43.95989415122109,
|
|
"learning_rate": 5.795041322314049e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.282,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 692
|
|
},
|
|
{
|
|
"epoch": 0.8503067484662576,
|
|
"grad_norm": 33.0652784939274,
|
|
"learning_rate": 5.791735537190082e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2934,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 7.8125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 693
|
|
},
|
|
{
|
|
"epoch": 0.8515337423312883,
|
|
"grad_norm": 46.044690743816275,
|
|
"learning_rate": 5.788429752066116e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.4005,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 694
|
|
},
|
|
{
|
|
"epoch": 0.852760736196319,
|
|
"grad_norm": 47.292538608776496,
|
|
"learning_rate": 5.785123966942149e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.208984375,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.4681,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 695
|
|
},
|
|
{
|
|
"epoch": 0.8539877300613496,
|
|
"grad_norm": 38.43478974924762,
|
|
"learning_rate": 5.781818181818182e-07,
|
|
"logits/chosen": -0.07568359375,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3104,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 696
|
|
},
|
|
{
|
|
"epoch": 0.8552147239263803,
|
|
"grad_norm": 58.82691398412282,
|
|
"learning_rate": 5.778512396694214e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 697
|
|
},
|
|
{
|
|
"epoch": 0.8564417177914111,
|
|
"grad_norm": 37.608292873182215,
|
|
"learning_rate": 5.775206611570247e-07,
|
|
"logits/chosen": -0.080078125,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2946,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 698
|
|
},
|
|
{
|
|
"epoch": 0.8576687116564418,
|
|
"grad_norm": 41.848029972370156,
|
|
"learning_rate": 5.77190082644628e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.2421875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2843,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 699
|
|
},
|
|
{
|
|
"epoch": 0.8588957055214724,
|
|
"grad_norm": 39.920137771888676,
|
|
"learning_rate": 5.768595041322314e-07,
|
|
"logits/chosen": -0.054931640625,
|
|
"logits/rejected": -0.1640625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3259,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.8601226993865031,
|
|
"grad_norm": 41.52681124295287,
|
|
"learning_rate": 5.765289256198347e-07,
|
|
"logits/chosen": -0.025146484375,
|
|
"logits/rejected": -0.138671875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.385,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 701
|
|
},
|
|
{
|
|
"epoch": 0.8613496932515338,
|
|
"grad_norm": 28.252991079387233,
|
|
"learning_rate": 5.76198347107438e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.1786,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.125,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 702
|
|
},
|
|
{
|
|
"epoch": 0.8625766871165644,
|
|
"grad_norm": 31.56795802217109,
|
|
"learning_rate": 5.758677685950413e-07,
|
|
"logits/chosen": -0.046142578125,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2909,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 703
|
|
},
|
|
{
|
|
"epoch": 0.8638036809815951,
|
|
"grad_norm": 34.11635728276886,
|
|
"learning_rate": 5.755371900826446e-07,
|
|
"logits/chosen": -0.1201171875,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -320.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3105,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 704
|
|
},
|
|
{
|
|
"epoch": 0.8650306748466258,
|
|
"grad_norm": 38.95077133588588,
|
|
"learning_rate": 5.75206611570248e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2767,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.265625,
|
|
"rewards/margins": 7.15625,
|
|
"rewards/rejected": -4.875,
|
|
"step": 705
|
|
},
|
|
{
|
|
"epoch": 0.8662576687116564,
|
|
"grad_norm": 35.267737415428456,
|
|
"learning_rate": 5.748760330578512e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3068,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 706
|
|
},
|
|
{
|
|
"epoch": 0.8674846625766871,
|
|
"grad_norm": 34.32550780952695,
|
|
"learning_rate": 5.745454545454545e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2683,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 707
|
|
},
|
|
{
|
|
"epoch": 0.8687116564417178,
|
|
"grad_norm": 31.650191661242673,
|
|
"learning_rate": 5.742148760330578e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2379,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.875,
|
|
"step": 708
|
|
},
|
|
{
|
|
"epoch": 0.8699386503067484,
|
|
"grad_norm": 25.32301991794396,
|
|
"learning_rate": 5.738842975206611e-07,
|
|
"logits/chosen": -0.06640625,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.185,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 709
|
|
},
|
|
{
|
|
"epoch": 0.8711656441717791,
|
|
"grad_norm": 33.22472192244502,
|
|
"learning_rate": 5.735537190082644e-07,
|
|
"logits/chosen": -0.017333984375,
|
|
"logits/rejected": -0.15234375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.299,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.625,
|
|
"step": 710
|
|
},
|
|
{
|
|
"epoch": 0.8723926380368098,
|
|
"grad_norm": 29.02267964751041,
|
|
"learning_rate": 5.732231404958678e-07,
|
|
"logits/chosen": -0.042724609375,
|
|
"logits/rejected": -0.236328125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2551,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 711
|
|
},
|
|
{
|
|
"epoch": 0.8736196319018404,
|
|
"grad_norm": 40.668984053435686,
|
|
"learning_rate": 5.728925619834711e-07,
|
|
"logits/chosen": -0.03857421875,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3587,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 712
|
|
},
|
|
{
|
|
"epoch": 0.8748466257668711,
|
|
"grad_norm": 33.08084767770349,
|
|
"learning_rate": 5.725619834710744e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2665,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 713
|
|
},
|
|
{
|
|
"epoch": 0.8760736196319019,
|
|
"grad_norm": 36.45936464644564,
|
|
"learning_rate": 5.722314049586777e-07,
|
|
"logits/chosen": -0.0159912109375,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2938,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.75,
|
|
"step": 714
|
|
},
|
|
{
|
|
"epoch": 0.8773006134969326,
|
|
"grad_norm": 30.913206903477224,
|
|
"learning_rate": 5.719008264462809e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2684,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 715
|
|
},
|
|
{
|
|
"epoch": 0.8785276073619632,
|
|
"grad_norm": 41.616080953336535,
|
|
"learning_rate": 5.715702479338842e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3142,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.625,
|
|
"step": 716
|
|
},
|
|
{
|
|
"epoch": 0.8797546012269939,
|
|
"grad_norm": 34.18654539125182,
|
|
"learning_rate": 5.712396694214876e-07,
|
|
"logits/chosen": -0.09423828125,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.2911,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.9296875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 717
|
|
},
|
|
{
|
|
"epoch": 0.8809815950920246,
|
|
"grad_norm": 28.09115304120503,
|
|
"learning_rate": 5.709090909090909e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2254,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 718
|
|
},
|
|
{
|
|
"epoch": 0.8822085889570552,
|
|
"grad_norm": 41.08426714668054,
|
|
"learning_rate": 5.705785123966942e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3011,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.25,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 719
|
|
},
|
|
{
|
|
"epoch": 0.8834355828220859,
|
|
"grad_norm": 28.518925312083375,
|
|
"learning_rate": 5.702479338842975e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2223,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.265625,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 720
|
|
},
|
|
{
|
|
"epoch": 0.8846625766871166,
|
|
"grad_norm": 26.149570167864507,
|
|
"learning_rate": 5.699173553719008e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2073,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 7.96875,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 721
|
|
},
|
|
{
|
|
"epoch": 0.8858895705521472,
|
|
"grad_norm": 30.929567326750046,
|
|
"learning_rate": 5.695867768595041e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2229,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.125,
|
|
"step": 722
|
|
},
|
|
{
|
|
"epoch": 0.8871165644171779,
|
|
"grad_norm": 34.37293415398814,
|
|
"learning_rate": 5.692561983471075e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2808,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 723
|
|
},
|
|
{
|
|
"epoch": 0.8883435582822086,
|
|
"grad_norm": 33.90677929130864,
|
|
"learning_rate": 5.689256198347107e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.254,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 724
|
|
},
|
|
{
|
|
"epoch": 0.8895705521472392,
|
|
"grad_norm": 34.97307578381243,
|
|
"learning_rate": 5.68595041322314e-07,
|
|
"logits/chosen": -0.055419921875,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3044,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 725
|
|
},
|
|
{
|
|
"epoch": 0.8907975460122699,
|
|
"grad_norm": 28.627835648582934,
|
|
"learning_rate": 5.682644628099173e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2135,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 726
|
|
},
|
|
{
|
|
"epoch": 0.8920245398773006,
|
|
"grad_norm": 34.81272514902735,
|
|
"learning_rate": 5.679338842975206e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.248,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 727
|
|
},
|
|
{
|
|
"epoch": 0.8932515337423312,
|
|
"grad_norm": 37.8936877345969,
|
|
"learning_rate": 5.676033057851239e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2946,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 7.65625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 728
|
|
},
|
|
{
|
|
"epoch": 0.894478527607362,
|
|
"grad_norm": 29.90567314985844,
|
|
"learning_rate": 5.672727272727273e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2088,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 7.6875,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 729
|
|
},
|
|
{
|
|
"epoch": 0.8957055214723927,
|
|
"grad_norm": 31.408249750098918,
|
|
"learning_rate": 5.669421487603306e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2204,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 7.9375,
|
|
"rewards/rejected": -5.75,
|
|
"step": 730
|
|
},
|
|
{
|
|
"epoch": 0.8969325153374234,
|
|
"grad_norm": 33.91624972638253,
|
|
"learning_rate": 5.666115702479339e-07,
|
|
"logits/chosen": -0.052734375,
|
|
"logits/rejected": -0.1845703125,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2736,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 7.1875,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 731
|
|
},
|
|
{
|
|
"epoch": 0.898159509202454,
|
|
"grad_norm": 40.7773464371931,
|
|
"learning_rate": 5.662809917355372e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3647,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 732
|
|
},
|
|
{
|
|
"epoch": 0.8993865030674847,
|
|
"grad_norm": 31.001224216873553,
|
|
"learning_rate": 5.659504132231404e-07,
|
|
"logits/chosen": -0.10791015625,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2212,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 733
|
|
},
|
|
{
|
|
"epoch": 0.9006134969325154,
|
|
"grad_norm": 37.803749225628884,
|
|
"learning_rate": 5.656198347107437e-07,
|
|
"logits/chosen": -0.08837890625,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2954,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 7.5,
|
|
"rewards/rejected": -6.25,
|
|
"step": 734
|
|
},
|
|
{
|
|
"epoch": 0.901840490797546,
|
|
"grad_norm": 30.18047116909555,
|
|
"learning_rate": 5.652892561983471e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.1824,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 8.125,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 735
|
|
},
|
|
{
|
|
"epoch": 0.9030674846625767,
|
|
"grad_norm": 38.572832638835585,
|
|
"learning_rate": 5.649586776859504e-07,
|
|
"logits/chosen": -0.025634765625,
|
|
"logits/rejected": -0.1552734375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3608,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.0908203125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -6.3125,
|
|
"step": 736
|
|
},
|
|
{
|
|
"epoch": 0.9042944785276074,
|
|
"grad_norm": 46.258369054797996,
|
|
"learning_rate": 5.646280991735537e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.4383,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.1826171875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -6.25,
|
|
"step": 737
|
|
},
|
|
{
|
|
"epoch": 0.905521472392638,
|
|
"grad_norm": 32.39664225178629,
|
|
"learning_rate": 5.64297520661157e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2588,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 7.5625,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 738
|
|
},
|
|
{
|
|
"epoch": 0.9067484662576687,
|
|
"grad_norm": 34.18126494383032,
|
|
"learning_rate": 5.639669421487603e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2836,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.2470703125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -6.15625,
|
|
"step": 739
|
|
},
|
|
{
|
|
"epoch": 0.9079754601226994,
|
|
"grad_norm": 35.15083279583394,
|
|
"learning_rate": 5.636363636363636e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3072,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.2431640625,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -6.6875,
|
|
"step": 740
|
|
},
|
|
{
|
|
"epoch": 0.90920245398773,
|
|
"grad_norm": 33.535600643328806,
|
|
"learning_rate": 5.63305785123967e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2851,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 7.9375,
|
|
"rewards/rejected": -6.8125,
|
|
"step": 741
|
|
},
|
|
{
|
|
"epoch": 0.9104294478527607,
|
|
"grad_norm": 31.52809779595991,
|
|
"learning_rate": 5.629752066115702e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2244,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 742
|
|
},
|
|
{
|
|
"epoch": 0.9116564417177914,
|
|
"grad_norm": 35.37117075886455,
|
|
"learning_rate": 5.626446280991735e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.212890625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.2769,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 743
|
|
},
|
|
{
|
|
"epoch": 0.912883435582822,
|
|
"grad_norm": 26.42035749866701,
|
|
"learning_rate": 5.623140495867768e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.1917,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -6.75,
|
|
"step": 744
|
|
},
|
|
{
|
|
"epoch": 0.9141104294478528,
|
|
"grad_norm": 35.853268232755404,
|
|
"learning_rate": 5.619834710743801e-07,
|
|
"logits/chosen": -0.08251953125,
|
|
"logits/rejected": -0.1865234375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3009,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.390625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.875,
|
|
"step": 745
|
|
},
|
|
{
|
|
"epoch": 0.9153374233128835,
|
|
"grad_norm": 43.35686834196019,
|
|
"learning_rate": 5.616528925619834e-07,
|
|
"logits/chosen": -0.0849609375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.317,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -6.0,
|
|
"step": 746
|
|
},
|
|
{
|
|
"epoch": 0.9165644171779141,
|
|
"grad_norm": 34.17444251605582,
|
|
"learning_rate": 5.613223140495868e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2811,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.41796875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 747
|
|
},
|
|
{
|
|
"epoch": 0.9177914110429448,
|
|
"grad_norm": 40.3104413419788,
|
|
"learning_rate": 5.609917355371901e-07,
|
|
"logits/chosen": -0.05224609375,
|
|
"logits/rejected": -0.169921875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3882,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.0279541015625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 748
|
|
},
|
|
{
|
|
"epoch": 0.9190184049079755,
|
|
"grad_norm": 40.57709610969512,
|
|
"learning_rate": 5.606611570247934e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3642,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 749
|
|
},
|
|
{
|
|
"epoch": 0.9202453987730062,
|
|
"grad_norm": 43.682412563973024,
|
|
"learning_rate": 5.603305785123967e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3704,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.63671875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.875,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 0.9214723926380368,
|
|
"grad_norm": 42.01608666567515,
|
|
"learning_rate": 5.599999999999999e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3006,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 7.4375,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 751
|
|
},
|
|
{
|
|
"epoch": 0.9226993865030675,
|
|
"grad_norm": 32.96496448782055,
|
|
"learning_rate": 5.596694214876033e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2435,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 7.75,
|
|
"rewards/rejected": -6.34375,
|
|
"step": 752
|
|
},
|
|
{
|
|
"epoch": 0.9239263803680982,
|
|
"grad_norm": 35.77444345136624,
|
|
"learning_rate": 5.593388429752066e-07,
|
|
"logits/chosen": -0.048095703125,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2851,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 753
|
|
},
|
|
{
|
|
"epoch": 0.9251533742331288,
|
|
"grad_norm": 39.56695959824003,
|
|
"learning_rate": 5.590082644628099e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3588,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.703125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 754
|
|
},
|
|
{
|
|
"epoch": 0.9263803680981595,
|
|
"grad_norm": 33.47777157003214,
|
|
"learning_rate": 5.586776859504132e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2421,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 755
|
|
},
|
|
{
|
|
"epoch": 0.9276073619631902,
|
|
"grad_norm": 30.73963254337421,
|
|
"learning_rate": 5.583471074380165e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.1669921875,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.248,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.322265625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.75,
|
|
"step": 756
|
|
},
|
|
{
|
|
"epoch": 0.9288343558282208,
|
|
"grad_norm": 25.863947966401952,
|
|
"learning_rate": 5.580165289256198e-07,
|
|
"logits/chosen": -0.0252685546875,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2117,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 757
|
|
},
|
|
{
|
|
"epoch": 0.9300613496932515,
|
|
"grad_norm": 40.231833773135115,
|
|
"learning_rate": 5.576859504132232e-07,
|
|
"logits/chosen": -0.08837890625,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3497,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.46875,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.59375,
|
|
"step": 758
|
|
},
|
|
{
|
|
"epoch": 0.9312883435582822,
|
|
"grad_norm": 34.14796678847348,
|
|
"learning_rate": 5.573553719008265e-07,
|
|
"logits/chosen": -0.0595703125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2511,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 759
|
|
},
|
|
{
|
|
"epoch": 0.9325153374233128,
|
|
"grad_norm": 29.867916639384045,
|
|
"learning_rate": 5.570247933884297e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2724,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 7.8125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 760
|
|
},
|
|
{
|
|
"epoch": 0.9337423312883436,
|
|
"grad_norm": 38.414301746929866,
|
|
"learning_rate": 5.56694214876033e-07,
|
|
"logits/chosen": -0.0264892578125,
|
|
"logits/rejected": -0.1669921875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2874,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.375,
|
|
"step": 761
|
|
},
|
|
{
|
|
"epoch": 0.9349693251533743,
|
|
"grad_norm": 40.447076836962,
|
|
"learning_rate": 5.563636363636363e-07,
|
|
"logits/chosen": -0.01397705078125,
|
|
"logits/rejected": -0.19921875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2888,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 762
|
|
},
|
|
{
|
|
"epoch": 0.9361963190184049,
|
|
"grad_norm": 31.879754151772115,
|
|
"learning_rate": 5.560330578512396e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.171875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2711,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 763
|
|
},
|
|
{
|
|
"epoch": 0.9374233128834356,
|
|
"grad_norm": 35.145542304507835,
|
|
"learning_rate": 5.55702479338843e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2896,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.26171875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -6.0,
|
|
"step": 764
|
|
},
|
|
{
|
|
"epoch": 0.9386503067484663,
|
|
"grad_norm": 40.67019743057879,
|
|
"learning_rate": 5.553719008264463e-07,
|
|
"logits/chosen": -0.0595703125,
|
|
"logits/rejected": -0.19140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3457,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.76171875,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 765
|
|
},
|
|
{
|
|
"epoch": 0.939877300613497,
|
|
"grad_norm": 32.24931969694431,
|
|
"learning_rate": 5.550413223140496e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.274,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 766
|
|
},
|
|
{
|
|
"epoch": 0.9411042944785276,
|
|
"grad_norm": 34.46831261260643,
|
|
"learning_rate": 5.547107438016529e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3035,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 767
|
|
},
|
|
{
|
|
"epoch": 0.9423312883435583,
|
|
"grad_norm": 36.15353177119284,
|
|
"learning_rate": 5.543801652892561e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.316,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -5.75,
|
|
"step": 768
|
|
},
|
|
{
|
|
"epoch": 0.943558282208589,
|
|
"grad_norm": 29.41748843520645,
|
|
"learning_rate": 5.540495867768594e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2432,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -6.375,
|
|
"step": 769
|
|
},
|
|
{
|
|
"epoch": 0.9447852760736196,
|
|
"grad_norm": 41.73128837710126,
|
|
"learning_rate": 5.537190082644628e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3004,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 7.09375,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 770
|
|
},
|
|
{
|
|
"epoch": 0.9460122699386503,
|
|
"grad_norm": 42.981692808778455,
|
|
"learning_rate": 5.533884297520661e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.201171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.429,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.0869140625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 771
|
|
},
|
|
{
|
|
"epoch": 0.947239263803681,
|
|
"grad_norm": 35.00554577923317,
|
|
"learning_rate": 5.530578512396694e-07,
|
|
"logits/chosen": -0.038330078125,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3114,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 772
|
|
},
|
|
{
|
|
"epoch": 0.9484662576687116,
|
|
"grad_norm": 36.88335000648106,
|
|
"learning_rate": 5.527272727272727e-07,
|
|
"logits/chosen": -0.0361328125,
|
|
"logits/rejected": -0.2255859375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3175,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.1513671875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 773
|
|
},
|
|
{
|
|
"epoch": 0.9496932515337423,
|
|
"grad_norm": 38.52772066032803,
|
|
"learning_rate": 5.52396694214876e-07,
|
|
"logits/chosen": -0.058837890625,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3665,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": -0.330078125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 774
|
|
},
|
|
{
|
|
"epoch": 0.950920245398773,
|
|
"grad_norm": 38.30272199053397,
|
|
"learning_rate": 5.520661157024793e-07,
|
|
"logits/chosen": 0.0306396484375,
|
|
"logits/rejected": -0.162109375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3245,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 775
|
|
},
|
|
{
|
|
"epoch": 0.9521472392638037,
|
|
"grad_norm": 33.16873912890908,
|
|
"learning_rate": 5.517355371900827e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3113,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -5.375,
|
|
"step": 776
|
|
},
|
|
{
|
|
"epoch": 0.9533742331288344,
|
|
"grad_norm": 35.432794054015964,
|
|
"learning_rate": 5.514049586776859e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3098,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 777
|
|
},
|
|
{
|
|
"epoch": 0.9546012269938651,
|
|
"grad_norm": 35.189286152943,
|
|
"learning_rate": 5.510743801652892e-07,
|
|
"logits/chosen": -0.1083984375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.251,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.443359375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 778
|
|
},
|
|
{
|
|
"epoch": 0.9558282208588957,
|
|
"grad_norm": 41.96194349470727,
|
|
"learning_rate": 5.507438016528925e-07,
|
|
"logits/chosen": -0.1640625,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2827,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 779
|
|
},
|
|
{
|
|
"epoch": 0.9570552147239264,
|
|
"grad_norm": 35.73940962793884,
|
|
"learning_rate": 5.504132231404958e-07,
|
|
"logits/chosen": -0.059326171875,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3415,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.2138671875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 780
|
|
},
|
|
{
|
|
"epoch": 0.9582822085889571,
|
|
"grad_norm": 26.50283260922046,
|
|
"learning_rate": 5.500826446280991e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.1906,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 8.0625,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 781
|
|
},
|
|
{
|
|
"epoch": 0.9595092024539877,
|
|
"grad_norm": 29.07388144480791,
|
|
"learning_rate": 5.497520661157025e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2204,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 782
|
|
},
|
|
{
|
|
"epoch": 0.9607361963190184,
|
|
"grad_norm": 36.281519570454215,
|
|
"learning_rate": 5.494214876033058e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2544,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 7.34375,
|
|
"rewards/rejected": -6.09375,
|
|
"step": 783
|
|
},
|
|
{
|
|
"epoch": 0.9619631901840491,
|
|
"grad_norm": 41.7797714543972,
|
|
"learning_rate": 5.490909090909091e-07,
|
|
"logits/chosen": -0.11181640625,
|
|
"logits/rejected": -0.162109375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3585,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.1767578125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 784
|
|
},
|
|
{
|
|
"epoch": 0.9631901840490797,
|
|
"grad_norm": 31.81554374340158,
|
|
"learning_rate": 5.487603305785124e-07,
|
|
"logits/chosen": -0.0751953125,
|
|
"logits/rejected": -0.24609375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2472,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -5.8125,
|
|
"step": 785
|
|
},
|
|
{
|
|
"epoch": 0.9644171779141104,
|
|
"grad_norm": 27.670018917456815,
|
|
"learning_rate": 5.484297520661156e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2222,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 0.4921875,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -6.25,
|
|
"step": 786
|
|
},
|
|
{
|
|
"epoch": 0.9656441717791411,
|
|
"grad_norm": 37.16124889650476,
|
|
"learning_rate": 5.480991735537189e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2838,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 7.125,
|
|
"rewards/rejected": -6.4375,
|
|
"step": 787
|
|
},
|
|
{
|
|
"epoch": 0.9668711656441717,
|
|
"grad_norm": 24.391962735004444,
|
|
"learning_rate": 5.477685950413223e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.1706,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 7.53125,
|
|
"rewards/rejected": -6.25,
|
|
"step": 788
|
|
},
|
|
{
|
|
"epoch": 0.9680981595092024,
|
|
"grad_norm": 32.024862187416,
|
|
"learning_rate": 5.474380165289256e-07,
|
|
"logits/chosen": -0.07958984375,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2843,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.408203125,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 789
|
|
},
|
|
{
|
|
"epoch": 0.9693251533742331,
|
|
"grad_norm": 27.51754897072937,
|
|
"learning_rate": 5.471074380165289e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.1866,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.70703125,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -6.03125,
|
|
"step": 790
|
|
},
|
|
{
|
|
"epoch": 0.9705521472392638,
|
|
"grad_norm": 37.440979455018265,
|
|
"learning_rate": 5.467768595041322e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2983,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -6.375,
|
|
"step": 791
|
|
},
|
|
{
|
|
"epoch": 0.9717791411042945,
|
|
"grad_norm": 36.704827910830204,
|
|
"learning_rate": 5.464462809917355e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3431,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 792
|
|
},
|
|
{
|
|
"epoch": 0.9730061349693252,
|
|
"grad_norm": 37.60757704339766,
|
|
"learning_rate": 5.461157024793389e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3237,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 793
|
|
},
|
|
{
|
|
"epoch": 0.9742331288343559,
|
|
"grad_norm": 37.59792190871499,
|
|
"learning_rate": 5.457851239669422e-07,
|
|
"logits/chosen": -0.103515625,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3909,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.21875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 794
|
|
},
|
|
{
|
|
"epoch": 0.9754601226993865,
|
|
"grad_norm": 28.85691430451381,
|
|
"learning_rate": 5.454545454545454e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2394,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 8.5,
|
|
"rewards/rejected": -6.90625,
|
|
"step": 795
|
|
},
|
|
{
|
|
"epoch": 0.9766871165644172,
|
|
"grad_norm": 41.11938718622486,
|
|
"learning_rate": 5.451239669421487e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2984,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": -0.0673828125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -6.375,
|
|
"step": 796
|
|
},
|
|
{
|
|
"epoch": 0.9779141104294479,
|
|
"grad_norm": 30.942308293626358,
|
|
"learning_rate": 5.44793388429752e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.254,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.6953125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.625,
|
|
"step": 797
|
|
},
|
|
{
|
|
"epoch": 0.9791411042944785,
|
|
"grad_norm": 36.786308023977305,
|
|
"learning_rate": 5.444628099173553e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2823,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -6.125,
|
|
"step": 798
|
|
},
|
|
{
|
|
"epoch": 0.9803680981595092,
|
|
"grad_norm": 40.89607573095922,
|
|
"learning_rate": 5.441322314049587e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3222,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 799
|
|
},
|
|
{
|
|
"epoch": 0.9815950920245399,
|
|
"grad_norm": 36.10519822526459,
|
|
"learning_rate": 5.43801652892562e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2658,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.61328125,
|
|
"rewards/margins": 6.625,
|
|
"rewards/rejected": -6.0,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.9828220858895705,
|
|
"grad_norm": 42.48783397210697,
|
|
"learning_rate": 5.434710743801653e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.3944,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 801
|
|
},
|
|
{
|
|
"epoch": 0.9840490797546012,
|
|
"grad_norm": 43.552763355211724,
|
|
"learning_rate": 5.431404958677686e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3899,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.453125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 802
|
|
},
|
|
{
|
|
"epoch": 0.9852760736196319,
|
|
"grad_norm": 32.09849729457811,
|
|
"learning_rate": 5.428099173553719e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2997,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.78125,
|
|
"step": 803
|
|
},
|
|
{
|
|
"epoch": 0.9865030674846625,
|
|
"grad_norm": 32.33615160143916,
|
|
"learning_rate": 5.424793388429751e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.240234375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3208,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.251953125,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 804
|
|
},
|
|
{
|
|
"epoch": 0.9877300613496932,
|
|
"grad_norm": 39.61512778302868,
|
|
"learning_rate": 5.421487603305785e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -380.0,
|
|
"loss": 0.3968,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.96875,
|
|
"step": 805
|
|
},
|
|
{
|
|
"epoch": 0.9889570552147239,
|
|
"grad_norm": 40.74080836640833,
|
|
"learning_rate": 5.418181818181818e-07,
|
|
"logits/chosen": -0.01202392578125,
|
|
"logits/rejected": -0.1474609375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3811,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 806
|
|
},
|
|
{
|
|
"epoch": 0.9901840490797545,
|
|
"grad_norm": 36.652039792723684,
|
|
"learning_rate": 5.414876033057851e-07,
|
|
"logits/chosen": -0.1005859375,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2327,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 7.21875,
|
|
"rewards/rejected": -6.125,
|
|
"step": 807
|
|
},
|
|
{
|
|
"epoch": 0.9914110429447853,
|
|
"grad_norm": 33.57961890786831,
|
|
"learning_rate": 5.411570247933884e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3091,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.11572265625,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 808
|
|
},
|
|
{
|
|
"epoch": 0.992638036809816,
|
|
"grad_norm": 35.60132202210017,
|
|
"learning_rate": 5.408264462809917e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2618,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 7.875,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 809
|
|
},
|
|
{
|
|
"epoch": 0.9938650306748467,
|
|
"grad_norm": 33.977380429457895,
|
|
"learning_rate": 5.40495867768595e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.227,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 7.28125,
|
|
"rewards/rejected": -6.0625,
|
|
"step": 810
|
|
},
|
|
{
|
|
"epoch": 0.9950920245398773,
|
|
"grad_norm": 31.635546850530186,
|
|
"learning_rate": 5.401652892561984e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2741,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.71875,
|
|
"step": 811
|
|
},
|
|
{
|
|
"epoch": 0.996319018404908,
|
|
"grad_norm": 40.80563506539147,
|
|
"learning_rate": 5.398347107438017e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2687,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.4609375,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -6.53125,
|
|
"step": 812
|
|
},
|
|
{
|
|
"epoch": 0.9975460122699387,
|
|
"grad_norm": 30.19996782105712,
|
|
"learning_rate": 5.395041322314049e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.1913,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.87109375,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -6.1875,
|
|
"step": 813
|
|
},
|
|
{
|
|
"epoch": 0.9987730061349693,
|
|
"grad_norm": 37.73036986625575,
|
|
"learning_rate": 5.391735537190082e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.201171875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.3425,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": -0.1416015625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 814
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"grad_norm": 40.870567834511405,
|
|
"learning_rate": 5.388429752066115e-07,
|
|
"logits/chosen": -0.0458984375,
|
|
"logits/rejected": -0.23828125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2125,
|
|
"rewards/accuracies": 0.8571429252624512,
|
|
"rewards/chosen": 0.59375,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -6.28125,
|
|
"step": 815
|
|
}
|
|
],
|
|
"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
|
|
}
|