9756dbb265
Model: shisa-ai/ablation-144-a128.dpo.armorm.rp.tl.5e7-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": 151.44113305401245,
|
|
"learning_rate": 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": 162.5072043033413,
|
|
"learning_rate": 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": 108.37171307643952,
|
|
"learning_rate": 6e-08,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.6896,
|
|
"rewards/accuracies": 0.328125,
|
|
"rewards/chosen": 0.005706787109375,
|
|
"rewards/margins": 0.01055908203125,
|
|
"rewards/rejected": -0.004852294921875,
|
|
"step": 3
|
|
},
|
|
{
|
|
"epoch": 0.0049079754601227,
|
|
"grad_norm": 152.5657567319493,
|
|
"learning_rate": 8e-08,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -242.0,
|
|
"loss": 0.7141,
|
|
"rewards/accuracies": 0.3359375,
|
|
"rewards/chosen": -0.009765625,
|
|
"rewards/margins": -0.025634765625,
|
|
"rewards/rejected": 0.0159912109375,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.006134969325153374,
|
|
"grad_norm": 288.1306811063231,
|
|
"learning_rate": 1e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.6984,
|
|
"rewards/accuracies": 0.421875,
|
|
"rewards/chosen": 0.0216064453125,
|
|
"rewards/margins": -0.0030670166015625,
|
|
"rewards/rejected": 0.0247802734375,
|
|
"step": 5
|
|
},
|
|
{
|
|
"epoch": 0.007361963190184049,
|
|
"grad_norm": 331.41871326801424,
|
|
"learning_rate": 1.2e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.7177,
|
|
"rewards/accuracies": 0.2578125,
|
|
"rewards/chosen": -0.01434326171875,
|
|
"rewards/margins": -0.03466796875,
|
|
"rewards/rejected": 0.0203857421875,
|
|
"step": 6
|
|
},
|
|
{
|
|
"epoch": 0.008588957055214725,
|
|
"grad_norm": 249.74586517146642,
|
|
"learning_rate": 1.4e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.7072,
|
|
"rewards/accuracies": 0.3359375,
|
|
"rewards/chosen": -0.000904083251953125,
|
|
"rewards/margins": -0.0172119140625,
|
|
"rewards/rejected": 0.016357421875,
|
|
"step": 7
|
|
},
|
|
{
|
|
"epoch": 0.0098159509202454,
|
|
"grad_norm": 128.1050509063449,
|
|
"learning_rate": 1.6e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.7045,
|
|
"rewards/accuracies": 0.2421875,
|
|
"rewards/chosen": 0.01513671875,
|
|
"rewards/margins": -0.01055908203125,
|
|
"rewards/rejected": 0.0257568359375,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.011042944785276074,
|
|
"grad_norm": 128.31082936106668,
|
|
"learning_rate": 1.8e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.7057,
|
|
"rewards/accuracies": 0.2890625,
|
|
"rewards/chosen": 0.0185546875,
|
|
"rewards/margins": -0.0194091796875,
|
|
"rewards/rejected": 0.0380859375,
|
|
"step": 9
|
|
},
|
|
{
|
|
"epoch": 0.012269938650306749,
|
|
"grad_norm": 148.13248133021244,
|
|
"learning_rate": 2e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.6879,
|
|
"rewards/accuracies": 0.359375,
|
|
"rewards/chosen": 0.03466796875,
|
|
"rewards/margins": 0.011962890625,
|
|
"rewards/rejected": 0.022705078125,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.013496932515337423,
|
|
"grad_norm": 188.4465050278061,
|
|
"learning_rate": 2.1999999999999998e-07,
|
|
"logits/chosen": -0.38671875,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.6847,
|
|
"rewards/accuracies": 0.421875,
|
|
"rewards/chosen": 0.04296875,
|
|
"rewards/margins": 0.025634765625,
|
|
"rewards/rejected": 0.017333984375,
|
|
"step": 11
|
|
},
|
|
{
|
|
"epoch": 0.014723926380368098,
|
|
"grad_norm": 283.48200245967496,
|
|
"learning_rate": 2.4e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.6763,
|
|
"rewards/accuracies": 0.4140625,
|
|
"rewards/chosen": 0.036376953125,
|
|
"rewards/margins": 0.042724609375,
|
|
"rewards/rejected": -0.00634765625,
|
|
"step": 12
|
|
},
|
|
{
|
|
"epoch": 0.015950920245398775,
|
|
"grad_norm": 157.86633246900865,
|
|
"learning_rate": 2.6e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -249.0,
|
|
"loss": 0.6399,
|
|
"rewards/accuracies": 0.59375,
|
|
"rewards/chosen": 0.068359375,
|
|
"rewards/margins": 0.1240234375,
|
|
"rewards/rejected": -0.055419921875,
|
|
"step": 13
|
|
},
|
|
{
|
|
"epoch": 0.01717791411042945,
|
|
"grad_norm": 170.28826054471034,
|
|
"learning_rate": 2.8e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -472.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.645,
|
|
"rewards/accuracies": 0.5390625,
|
|
"rewards/chosen": 0.1201171875,
|
|
"rewards/margins": 0.12109375,
|
|
"rewards/rejected": -0.000881195068359375,
|
|
"step": 14
|
|
},
|
|
{
|
|
"epoch": 0.018404907975460124,
|
|
"grad_norm": 196.84243180155542,
|
|
"learning_rate": 3e-07,
|
|
"logits/chosen": -0.12451171875,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.6433,
|
|
"rewards/accuracies": 0.5,
|
|
"rewards/chosen": 0.0849609375,
|
|
"rewards/margins": 0.1298828125,
|
|
"rewards/rejected": -0.044677734375,
|
|
"step": 15
|
|
},
|
|
{
|
|
"epoch": 0.0196319018404908,
|
|
"grad_norm": 150.44401802967914,
|
|
"learning_rate": 3.2e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.6163,
|
|
"rewards/accuracies": 0.609375,
|
|
"rewards/chosen": 0.1513671875,
|
|
"rewards/margins": 0.189453125,
|
|
"rewards/rejected": -0.03759765625,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.020858895705521473,
|
|
"grad_norm": 158.3424645259106,
|
|
"learning_rate": 3.4000000000000003e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -254.0,
|
|
"loss": 0.5994,
|
|
"rewards/accuracies": 0.6171875,
|
|
"rewards/chosen": 0.173828125,
|
|
"rewards/margins": 0.25390625,
|
|
"rewards/rejected": -0.08056640625,
|
|
"step": 17
|
|
},
|
|
{
|
|
"epoch": 0.022085889570552148,
|
|
"grad_norm": 127.76158860883122,
|
|
"learning_rate": 3.6e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.5706,
|
|
"rewards/accuracies": 0.6953125,
|
|
"rewards/chosen": 0.2294921875,
|
|
"rewards/margins": 0.3359375,
|
|
"rewards/rejected": -0.10595703125,
|
|
"step": 18
|
|
},
|
|
{
|
|
"epoch": 0.023312883435582823,
|
|
"grad_norm": 69.38468528362102,
|
|
"learning_rate": 3.7999999999999996e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.567,
|
|
"rewards/accuracies": 0.6484375,
|
|
"rewards/chosen": 0.2197265625,
|
|
"rewards/margins": 0.37890625,
|
|
"rewards/rejected": -0.158203125,
|
|
"step": 19
|
|
},
|
|
{
|
|
"epoch": 0.024539877300613498,
|
|
"grad_norm": 82.21345162463909,
|
|
"learning_rate": 4e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -466.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.5153,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 0.37890625,
|
|
"rewards/margins": 0.58984375,
|
|
"rewards/rejected": -0.2109375,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.025766871165644172,
|
|
"grad_norm": 105.40201871897163,
|
|
"learning_rate": 4.1999999999999995e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.5326,
|
|
"rewards/accuracies": 0.703125,
|
|
"rewards/chosen": 0.296875,
|
|
"rewards/margins": 0.490234375,
|
|
"rewards/rejected": -0.1923828125,
|
|
"step": 21
|
|
},
|
|
{
|
|
"epoch": 0.026993865030674847,
|
|
"grad_norm": 70.78030423389544,
|
|
"learning_rate": 4.3999999999999997e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.5334,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": 0.291015625,
|
|
"rewards/margins": 0.5390625,
|
|
"rewards/rejected": -0.2470703125,
|
|
"step": 22
|
|
},
|
|
{
|
|
"epoch": 0.02822085889570552,
|
|
"grad_norm": 60.46408711421665,
|
|
"learning_rate": 4.6e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.513,
|
|
"rewards/accuracies": 0.6953125,
|
|
"rewards/chosen": 0.498046875,
|
|
"rewards/margins": 0.75390625,
|
|
"rewards/rejected": -0.2578125,
|
|
"step": 23
|
|
},
|
|
{
|
|
"epoch": 0.029447852760736196,
|
|
"grad_norm": 54.54733596050591,
|
|
"learning_rate": 4.8e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.5458,
|
|
"rewards/accuracies": 0.625,
|
|
"rewards/chosen": 0.36328125,
|
|
"rewards/margins": 0.6015625,
|
|
"rewards/rejected": -0.2392578125,
|
|
"step": 24
|
|
},
|
|
{
|
|
"epoch": 0.03067484662576687,
|
|
"grad_norm": 46.35027564725166,
|
|
"learning_rate": 5e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.468,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": 0.65234375,
|
|
"rewards/margins": 1.0859375,
|
|
"rewards/rejected": -0.43359375,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.03190184049079755,
|
|
"grad_norm": 49.250838916767385,
|
|
"learning_rate": 4.99793388429752e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.4753,
|
|
"rewards/accuracies": 0.7265625,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 1.15625,
|
|
"rewards/rejected": -0.404296875,
|
|
"step": 26
|
|
},
|
|
{
|
|
"epoch": 0.033128834355828224,
|
|
"grad_norm": 49.243509562031704,
|
|
"learning_rate": 4.995867768595041e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.4481,
|
|
"rewards/accuracies": 0.734375,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 1.3125,
|
|
"rewards/rejected": -0.46484375,
|
|
"step": 27
|
|
},
|
|
{
|
|
"epoch": 0.0343558282208589,
|
|
"grad_norm": 43.853948458317184,
|
|
"learning_rate": 4.993801652892562e-07,
|
|
"logits/chosen": -0.08203125,
|
|
"logits/rejected": -0.1884765625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -258.0,
|
|
"loss": 0.4667,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 1.125,
|
|
"rewards/rejected": -0.33203125,
|
|
"step": 28
|
|
},
|
|
{
|
|
"epoch": 0.03558282208588957,
|
|
"grad_norm": 46.27857873377814,
|
|
"learning_rate": 4.991735537190083e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.442,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 1.46875,
|
|
"rewards/rejected": -0.5234375,
|
|
"step": 29
|
|
},
|
|
{
|
|
"epoch": 0.03680981595092025,
|
|
"grad_norm": 54.43172367072751,
|
|
"learning_rate": 4.989669421487603e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -260.0,
|
|
"loss": 0.4973,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 1.1015625,
|
|
"rewards/rejected": -0.4296875,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.03803680981595092,
|
|
"grad_norm": 49.563120237934434,
|
|
"learning_rate": 4.987603305785124e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.4585,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 1.5390625,
|
|
"rewards/rejected": -0.5703125,
|
|
"step": 31
|
|
},
|
|
{
|
|
"epoch": 0.0392638036809816,
|
|
"grad_norm": 47.082618739282914,
|
|
"learning_rate": 4.985537190082644e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.4295,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 1.578125,
|
|
"rewards/rejected": -0.61328125,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.04049079754601227,
|
|
"grad_norm": 43.28996668354881,
|
|
"learning_rate": 4.983471074380165e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -229.0,
|
|
"loss": 0.4502,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.73828125,
|
|
"rewards/margins": 1.3203125,
|
|
"rewards/rejected": -0.58203125,
|
|
"step": 33
|
|
},
|
|
{
|
|
"epoch": 0.04171779141104295,
|
|
"grad_norm": 53.08103037992315,
|
|
"learning_rate": 4.981404958677686e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.4275,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 0.9296875,
|
|
"rewards/margins": 1.78125,
|
|
"rewards/rejected": -0.84765625,
|
|
"step": 34
|
|
},
|
|
{
|
|
"epoch": 0.04294478527607362,
|
|
"grad_norm": 51.02011562118897,
|
|
"learning_rate": 4.979338842975207e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -258.0,
|
|
"loss": 0.4005,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 1.78125,
|
|
"rewards/rejected": -1.0234375,
|
|
"step": 35
|
|
},
|
|
{
|
|
"epoch": 0.044171779141104296,
|
|
"grad_norm": 39.1724548423627,
|
|
"learning_rate": 4.977272727272727e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3849,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 2.1875,
|
|
"rewards/rejected": -1.0546875,
|
|
"step": 36
|
|
},
|
|
{
|
|
"epoch": 0.04539877300613497,
|
|
"grad_norm": 41.851898602487374,
|
|
"learning_rate": 4.975206611570248e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.4066,
|
|
"rewards/accuracies": 0.7421875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 2.296875,
|
|
"rewards/rejected": -0.9765625,
|
|
"step": 37
|
|
},
|
|
{
|
|
"epoch": 0.046625766871165646,
|
|
"grad_norm": 38.15131350129523,
|
|
"learning_rate": 4.973140495867769e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3886,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 2.578125,
|
|
"rewards/rejected": -1.046875,
|
|
"step": 38
|
|
},
|
|
{
|
|
"epoch": 0.04785276073619632,
|
|
"grad_norm": 41.71923927967426,
|
|
"learning_rate": 4.971074380165288e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3805,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 2.640625,
|
|
"rewards/rejected": -1.3125,
|
|
"step": 39
|
|
},
|
|
{
|
|
"epoch": 0.049079754601226995,
|
|
"grad_norm": 38.402003702110484,
|
|
"learning_rate": 4.96900826446281e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3421,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 2.734375,
|
|
"rewards/rejected": -1.421875,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 0.05030674846625767,
|
|
"grad_norm": 34.67962672406851,
|
|
"learning_rate": 4.96694214876033e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.3787,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 2.65625,
|
|
"rewards/rejected": -1.3515625,
|
|
"step": 41
|
|
},
|
|
{
|
|
"epoch": 0.051533742331288344,
|
|
"grad_norm": 46.35208610876743,
|
|
"learning_rate": 4.964876033057851e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3611,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 2.71875,
|
|
"rewards/rejected": -1.515625,
|
|
"step": 42
|
|
},
|
|
{
|
|
"epoch": 0.05276073619631902,
|
|
"grad_norm": 48.36417348819516,
|
|
"learning_rate": 4.962809917355371e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3483,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 3.109375,
|
|
"rewards/rejected": -1.7265625,
|
|
"step": 43
|
|
},
|
|
{
|
|
"epoch": 0.053987730061349694,
|
|
"grad_norm": 58.150094006043,
|
|
"learning_rate": 4.960743801652892e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.4261,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 2.75,
|
|
"rewards/rejected": -1.8046875,
|
|
"step": 44
|
|
},
|
|
{
|
|
"epoch": 0.05521472392638037,
|
|
"grad_norm": 41.77249990347939,
|
|
"learning_rate": 4.958677685950413e-07,
|
|
"logits/chosen": -0.34765625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3831,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 2.890625,
|
|
"rewards/rejected": -1.8671875,
|
|
"step": 45
|
|
},
|
|
{
|
|
"epoch": 0.05644171779141104,
|
|
"grad_norm": 42.522353640053986,
|
|
"learning_rate": 4.956611570247934e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3376,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 3.5,
|
|
"rewards/rejected": -1.953125,
|
|
"step": 46
|
|
},
|
|
{
|
|
"epoch": 0.05766871165644172,
|
|
"grad_norm": 59.82362665796057,
|
|
"learning_rate": 4.954545454545454e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.4408,
|
|
"rewards/accuracies": 0.734375,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 2.6875,
|
|
"rewards/rejected": -1.6484375,
|
|
"step": 47
|
|
},
|
|
{
|
|
"epoch": 0.05889570552147239,
|
|
"grad_norm": 36.37462895121556,
|
|
"learning_rate": 4.952479338842975e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.3918,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 2.4375,
|
|
"rewards/rejected": -1.4296875,
|
|
"step": 48
|
|
},
|
|
{
|
|
"epoch": 0.06012269938650307,
|
|
"grad_norm": 33.380998200862365,
|
|
"learning_rate": 4.950413223140495e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2888,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 3.65625,
|
|
"rewards/rejected": -2.078125,
|
|
"step": 49
|
|
},
|
|
{
|
|
"epoch": 0.06134969325153374,
|
|
"grad_norm": 49.61344030643431,
|
|
"learning_rate": 4.948347107438016e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.4032,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 2.890625,
|
|
"rewards/rejected": -1.609375,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.06257668711656442,
|
|
"grad_norm": 44.94363880784817,
|
|
"learning_rate": 4.946280991735537e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3064,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 3.578125,
|
|
"rewards/rejected": -1.9375,
|
|
"step": 51
|
|
},
|
|
{
|
|
"epoch": 0.0638036809815951,
|
|
"grad_norm": 41.35776161230766,
|
|
"learning_rate": 4.944214876033058e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.369,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 3.34375,
|
|
"rewards/rejected": -2.15625,
|
|
"step": 52
|
|
},
|
|
{
|
|
"epoch": 0.06503067484662577,
|
|
"grad_norm": 44.728082439821826,
|
|
"learning_rate": 4.942148760330578e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3683,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 3.109375,
|
|
"rewards/rejected": -1.7578125,
|
|
"step": 53
|
|
},
|
|
{
|
|
"epoch": 0.06625766871165645,
|
|
"grad_norm": 39.009654963149515,
|
|
"learning_rate": 4.940082644628099e-07,
|
|
"logits/chosen": -0.072265625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -251.0,
|
|
"loss": 0.3097,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 2.796875,
|
|
"rewards/rejected": -1.8203125,
|
|
"step": 54
|
|
},
|
|
{
|
|
"epoch": 0.06748466257668712,
|
|
"grad_norm": 40.85012819896818,
|
|
"learning_rate": 4.93801652892562e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3452,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 3.140625,
|
|
"rewards/rejected": -1.9609375,
|
|
"step": 55
|
|
},
|
|
{
|
|
"epoch": 0.0687116564417178,
|
|
"grad_norm": 46.70951578377314,
|
|
"learning_rate": 4.935950413223141e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3916,
|
|
"rewards/accuracies": 0.7578125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 3.234375,
|
|
"rewards/rejected": -1.875,
|
|
"step": 56
|
|
},
|
|
{
|
|
"epoch": 0.06993865030674846,
|
|
"grad_norm": 51.30803055192484,
|
|
"learning_rate": 4.933884297520661e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3947,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 3.203125,
|
|
"rewards/rejected": -2.015625,
|
|
"step": 57
|
|
},
|
|
{
|
|
"epoch": 0.07116564417177915,
|
|
"grad_norm": 47.592024830773475,
|
|
"learning_rate": 4.931818181818182e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3758,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 3.328125,
|
|
"rewards/rejected": -1.9140625,
|
|
"step": 58
|
|
},
|
|
{
|
|
"epoch": 0.07239263803680981,
|
|
"grad_norm": 48.17219624892423,
|
|
"learning_rate": 4.929752066115702e-07,
|
|
"logits/chosen": -0.0947265625,
|
|
"logits/rejected": -0.1923828125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.4219,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 2.59375,
|
|
"rewards/rejected": -1.5390625,
|
|
"step": 59
|
|
},
|
|
{
|
|
"epoch": 0.0736196319018405,
|
|
"grad_norm": 39.36985463274499,
|
|
"learning_rate": 4.927685950413223e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.3667,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 2.796875,
|
|
"rewards/rejected": -1.8125,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 0.07484662576687116,
|
|
"grad_norm": 38.11628572877564,
|
|
"learning_rate": 4.925619834710743e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3265,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 3.65625,
|
|
"rewards/rejected": -2.0625,
|
|
"step": 61
|
|
},
|
|
{
|
|
"epoch": 0.07607361963190185,
|
|
"grad_norm": 40.58395817948392,
|
|
"learning_rate": 4.923553719008265e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3375,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.265625,
|
|
"step": 62
|
|
},
|
|
{
|
|
"epoch": 0.07730061349693251,
|
|
"grad_norm": 43.13092829482704,
|
|
"learning_rate": 4.921487603305785e-07,
|
|
"logits/chosen": -0.06787109375,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3277,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 3.203125,
|
|
"rewards/rejected": -1.96875,
|
|
"step": 63
|
|
},
|
|
{
|
|
"epoch": 0.0785276073619632,
|
|
"grad_norm": 57.40916665463975,
|
|
"learning_rate": 4.919421487603306e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3991,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 3.609375,
|
|
"rewards/rejected": -2.09375,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.07975460122699386,
|
|
"grad_norm": 39.641816149243105,
|
|
"learning_rate": 4.917355371900826e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3463,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 3.4375,
|
|
"rewards/rejected": -2.0,
|
|
"step": 65
|
|
},
|
|
{
|
|
"epoch": 0.08098159509202454,
|
|
"grad_norm": 49.84344547746062,
|
|
"learning_rate": 4.915289256198346e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -254.0,
|
|
"loss": 0.3217,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 3.796875,
|
|
"rewards/rejected": -2.3125,
|
|
"step": 66
|
|
},
|
|
{
|
|
"epoch": 0.08220858895705521,
|
|
"grad_norm": 37.32761829857322,
|
|
"learning_rate": 4.913223140495867e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3388,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 3.125,
|
|
"rewards/rejected": -2.1875,
|
|
"step": 67
|
|
},
|
|
{
|
|
"epoch": 0.0834355828220859,
|
|
"grad_norm": 44.340094240446255,
|
|
"learning_rate": 4.911157024793388e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3709,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 3.234375,
|
|
"rewards/rejected": -2.078125,
|
|
"step": 68
|
|
},
|
|
{
|
|
"epoch": 0.08466257668711656,
|
|
"grad_norm": 36.72487633784899,
|
|
"learning_rate": 4.909090909090909e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3115,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.25,
|
|
"step": 69
|
|
},
|
|
{
|
|
"epoch": 0.08588957055214724,
|
|
"grad_norm": 38.53295537225695,
|
|
"learning_rate": 4.907024793388429e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2897,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 3.875,
|
|
"rewards/rejected": -2.109375,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 0.08711656441717791,
|
|
"grad_norm": 41.12312548370276,
|
|
"learning_rate": 4.90495867768595e-07,
|
|
"logits/chosen": -0.095703125,
|
|
"logits/rejected": -0.1796875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.4118,
|
|
"rewards/accuracies": 0.75,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 2.921875,
|
|
"rewards/rejected": -1.9453125,
|
|
"step": 71
|
|
},
|
|
{
|
|
"epoch": 0.08834355828220859,
|
|
"grad_norm": 39.321341602611746,
|
|
"learning_rate": 4.902892561983471e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3197,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 3.53125,
|
|
"rewards/rejected": -2.03125,
|
|
"step": 72
|
|
},
|
|
{
|
|
"epoch": 0.08957055214723926,
|
|
"grad_norm": 44.53160035761988,
|
|
"learning_rate": 4.900826446280992e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3927,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 3.34375,
|
|
"rewards/rejected": -2.21875,
|
|
"step": 73
|
|
},
|
|
{
|
|
"epoch": 0.09079754601226994,
|
|
"grad_norm": 40.466626208621626,
|
|
"learning_rate": 4.898760330578512e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.303,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 3.25,
|
|
"rewards/rejected": -2.28125,
|
|
"step": 74
|
|
},
|
|
{
|
|
"epoch": 0.09202453987730061,
|
|
"grad_norm": 35.5267105162623,
|
|
"learning_rate": 4.896694214876033e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2978,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 3.8125,
|
|
"rewards/rejected": -2.359375,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.09325153374233129,
|
|
"grad_norm": 47.01910580765987,
|
|
"learning_rate": 4.894628099173553e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3205,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 3.671875,
|
|
"rewards/rejected": -2.203125,
|
|
"step": 76
|
|
},
|
|
{
|
|
"epoch": 0.09447852760736196,
|
|
"grad_norm": 37.028647946403694,
|
|
"learning_rate": 4.892561983471074e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3058,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 3.59375,
|
|
"rewards/rejected": -2.46875,
|
|
"step": 77
|
|
},
|
|
{
|
|
"epoch": 0.09570552147239264,
|
|
"grad_norm": 41.57134852043157,
|
|
"learning_rate": 4.890495867768595e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3141,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 3.75,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 78
|
|
},
|
|
{
|
|
"epoch": 0.09693251533742331,
|
|
"grad_norm": 34.691948521740784,
|
|
"learning_rate": 4.888429752066116e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3063,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 3.40625,
|
|
"rewards/rejected": -2.3125,
|
|
"step": 79
|
|
},
|
|
{
|
|
"epoch": 0.09815950920245399,
|
|
"grad_norm": 49.17092037552417,
|
|
"learning_rate": 4.886363636363636e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2984,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 0.09938650306748466,
|
|
"grad_norm": 39.19664643244926,
|
|
"learning_rate": 4.884297520661157e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3241,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -2.578125,
|
|
"step": 81
|
|
},
|
|
{
|
|
"epoch": 0.10061349693251534,
|
|
"grad_norm": 49.311491941358796,
|
|
"learning_rate": 4.882231404958677e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3395,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 3.65625,
|
|
"rewards/rejected": -2.640625,
|
|
"step": 82
|
|
},
|
|
{
|
|
"epoch": 0.10184049079754601,
|
|
"grad_norm": 49.376121428612976,
|
|
"learning_rate": 4.880165289256198e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.32,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 3.734375,
|
|
"rewards/rejected": -2.59375,
|
|
"step": 83
|
|
},
|
|
{
|
|
"epoch": 0.10306748466257669,
|
|
"grad_norm": 64.06732250692619,
|
|
"learning_rate": 4.878099173553718e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3022,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.875,
|
|
"step": 84
|
|
},
|
|
{
|
|
"epoch": 0.10429447852760736,
|
|
"grad_norm": 44.025841834176795,
|
|
"learning_rate": 4.87603305785124e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3406,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 3.71875,
|
|
"rewards/rejected": -2.75,
|
|
"step": 85
|
|
},
|
|
{
|
|
"epoch": 0.10552147239263804,
|
|
"grad_norm": 41.64304248400373,
|
|
"learning_rate": 4.87396694214876e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3372,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 3.78125,
|
|
"rewards/rejected": -2.359375,
|
|
"step": 86
|
|
},
|
|
{
|
|
"epoch": 0.1067484662576687,
|
|
"grad_norm": 33.11395742992561,
|
|
"learning_rate": 4.871900826446281e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -252.0,
|
|
"loss": 0.2974,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.7421875,
|
|
"rewards/margins": 3.484375,
|
|
"rewards/rejected": -2.75,
|
|
"step": 87
|
|
},
|
|
{
|
|
"epoch": 0.10797546012269939,
|
|
"grad_norm": 42.17124755394979,
|
|
"learning_rate": 4.869834710743801e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.2793,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 3.6875,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 88
|
|
},
|
|
{
|
|
"epoch": 0.10920245398773006,
|
|
"grad_norm": 43.24754945978333,
|
|
"learning_rate": 4.867768595041323e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3704,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 3.65625,
|
|
"rewards/rejected": -2.328125,
|
|
"step": 89
|
|
},
|
|
{
|
|
"epoch": 0.11042944785276074,
|
|
"grad_norm": 36.97477331926065,
|
|
"learning_rate": 4.865702479338843e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3415,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 3.5,
|
|
"rewards/rejected": -2.53125,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 0.1116564417177914,
|
|
"grad_norm": 43.208078047171966,
|
|
"learning_rate": 4.863636363636364e-07,
|
|
"logits/chosen": -0.04638671875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2904,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 3.78125,
|
|
"rewards/rejected": -2.53125,
|
|
"step": 91
|
|
},
|
|
{
|
|
"epoch": 0.11288343558282209,
|
|
"grad_norm": 45.84043985703993,
|
|
"learning_rate": 4.861570247933884e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2811,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 3.921875,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 92
|
|
},
|
|
{
|
|
"epoch": 0.11411042944785275,
|
|
"grad_norm": 36.66550841517969,
|
|
"learning_rate": 4.859504132231405e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3038,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.515625,
|
|
"step": 93
|
|
},
|
|
{
|
|
"epoch": 0.11533742331288344,
|
|
"grad_norm": 45.36547804836722,
|
|
"learning_rate": 4.857438016528925e-07,
|
|
"logits/chosen": -0.181640625,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.298,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -2.4375,
|
|
"step": 94
|
|
},
|
|
{
|
|
"epoch": 0.1165644171779141,
|
|
"grad_norm": 36.1886645335595,
|
|
"learning_rate": 4.855371900826447e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3254,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 95
|
|
},
|
|
{
|
|
"epoch": 0.11779141104294479,
|
|
"grad_norm": 47.78990929288722,
|
|
"learning_rate": 4.853305785123967e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2721,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -2.734375,
|
|
"step": 96
|
|
},
|
|
{
|
|
"epoch": 0.11901840490797547,
|
|
"grad_norm": 46.646580854293525,
|
|
"learning_rate": 4.851239669421487e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3608,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 4.0,
|
|
"rewards/rejected": -2.328125,
|
|
"step": 97
|
|
},
|
|
{
|
|
"epoch": 0.12024539877300613,
|
|
"grad_norm": 38.68608855483517,
|
|
"learning_rate": 4.849173553719008e-07,
|
|
"logits/chosen": -0.0419921875,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 3.71875,
|
|
"rewards/rejected": -2.21875,
|
|
"step": 98
|
|
},
|
|
{
|
|
"epoch": 0.12147239263803682,
|
|
"grad_norm": 39.562539416001016,
|
|
"learning_rate": 4.847107438016528e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -239.0,
|
|
"loss": 0.2701,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 3.46875,
|
|
"rewards/rejected": -2.34375,
|
|
"step": 99
|
|
},
|
|
{
|
|
"epoch": 0.12269938650306748,
|
|
"grad_norm": 37.50806782806318,
|
|
"learning_rate": 4.84504132231405e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3124,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -2.578125,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.12392638036809817,
|
|
"grad_norm": 46.32080898977058,
|
|
"learning_rate": 4.842975206611569e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3659,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.375,
|
|
"step": 101
|
|
},
|
|
{
|
|
"epoch": 0.12515337423312883,
|
|
"grad_norm": 38.46942420347679,
|
|
"learning_rate": 4.840909090909091e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2962,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 4.0,
|
|
"rewards/rejected": -2.328125,
|
|
"step": 102
|
|
},
|
|
{
|
|
"epoch": 0.1263803680981595,
|
|
"grad_norm": 39.79508252800174,
|
|
"learning_rate": 4.838842975206611e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3121,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 3.734375,
|
|
"rewards/rejected": -2.265625,
|
|
"step": 103
|
|
},
|
|
{
|
|
"epoch": 0.1276073619631902,
|
|
"grad_norm": 47.40646017292448,
|
|
"learning_rate": 4.836776859504132e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3411,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 3.734375,
|
|
"rewards/rejected": -2.265625,
|
|
"step": 104
|
|
},
|
|
{
|
|
"epoch": 0.12883435582822086,
|
|
"grad_norm": 38.094017856958104,
|
|
"learning_rate": 4.834710743801652e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2782,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.515625,
|
|
"step": 105
|
|
},
|
|
{
|
|
"epoch": 0.13006134969325153,
|
|
"grad_norm": 38.92036345370265,
|
|
"learning_rate": 4.832644628099174e-07,
|
|
"logits/chosen": -0.177734375,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3126,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 3.953125,
|
|
"rewards/rejected": -2.34375,
|
|
"step": 106
|
|
},
|
|
{
|
|
"epoch": 0.1312883435582822,
|
|
"grad_norm": 41.08010358460662,
|
|
"learning_rate": 4.830578512396694e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2914,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 3.8125,
|
|
"rewards/rejected": -2.375,
|
|
"step": 107
|
|
},
|
|
{
|
|
"epoch": 0.1325153374233129,
|
|
"grad_norm": 36.5889842300964,
|
|
"learning_rate": 4.828512396694215e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2899,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 3.921875,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 108
|
|
},
|
|
{
|
|
"epoch": 0.13374233128834356,
|
|
"grad_norm": 39.09556301559485,
|
|
"learning_rate": 4.826446280991735e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3113,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 109
|
|
},
|
|
{
|
|
"epoch": 0.13496932515337423,
|
|
"grad_norm": 40.98286093074122,
|
|
"learning_rate": 4.824380165289256e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3227,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 0.1361963190184049,
|
|
"grad_norm": 35.811211169912276,
|
|
"learning_rate": 4.822314049586776e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.303,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 3.546875,
|
|
"rewards/rejected": -2.15625,
|
|
"step": 111
|
|
},
|
|
{
|
|
"epoch": 0.1374233128834356,
|
|
"grad_norm": 34.57286118091264,
|
|
"learning_rate": 4.820247933884298e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.259,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 3.828125,
|
|
"rewards/rejected": -2.40625,
|
|
"step": 112
|
|
},
|
|
{
|
|
"epoch": 0.13865030674846626,
|
|
"grad_norm": 39.00609248421783,
|
|
"learning_rate": 4.818181818181818e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.216796875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3242,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 3.8125,
|
|
"rewards/rejected": -2.234375,
|
|
"step": 113
|
|
},
|
|
{
|
|
"epoch": 0.13987730061349693,
|
|
"grad_norm": 44.795935075325445,
|
|
"learning_rate": 4.816115702479339e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3214,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 3.484375,
|
|
"rewards/rejected": -2.296875,
|
|
"step": 114
|
|
},
|
|
{
|
|
"epoch": 0.1411042944785276,
|
|
"grad_norm": 43.827084866837005,
|
|
"learning_rate": 4.814049586776859e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.338,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 3.296875,
|
|
"rewards/rejected": -2.34375,
|
|
"step": 115
|
|
},
|
|
{
|
|
"epoch": 0.1423312883435583,
|
|
"grad_norm": 32.836140603929685,
|
|
"learning_rate": 4.81198347107438e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2561,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -2.546875,
|
|
"step": 116
|
|
},
|
|
{
|
|
"epoch": 0.14355828220858896,
|
|
"grad_norm": 34.86086710464441,
|
|
"learning_rate": 4.809917355371901e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -255.0,
|
|
"loss": 0.289,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 3.828125,
|
|
"rewards/rejected": -2.6875,
|
|
"step": 117
|
|
},
|
|
{
|
|
"epoch": 0.14478527607361963,
|
|
"grad_norm": 39.71090585702439,
|
|
"learning_rate": 4.807851239669422e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3177,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -2.296875,
|
|
"step": 118
|
|
},
|
|
{
|
|
"epoch": 0.1460122699386503,
|
|
"grad_norm": 39.95542336983019,
|
|
"learning_rate": 4.805785123966942e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3108,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 3.921875,
|
|
"rewards/rejected": -2.65625,
|
|
"step": 119
|
|
},
|
|
{
|
|
"epoch": 0.147239263803681,
|
|
"grad_norm": 43.1225245187085,
|
|
"learning_rate": 4.803719008264463e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2861,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -2.53125,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 0.14846625766871166,
|
|
"grad_norm": 37.11625554264979,
|
|
"learning_rate": 4.801652892561983e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2783,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.9296875,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 121
|
|
},
|
|
{
|
|
"epoch": 0.14969325153374233,
|
|
"grad_norm": 40.09291676904836,
|
|
"learning_rate": 4.799586776859503e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2905,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 3.859375,
|
|
"rewards/rejected": -2.390625,
|
|
"step": 122
|
|
},
|
|
{
|
|
"epoch": 0.150920245398773,
|
|
"grad_norm": 40.539065533878755,
|
|
"learning_rate": 4.797520661157025e-07,
|
|
"logits/chosen": -0.06982421875,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -272.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 3.734375,
|
|
"rewards/rejected": -2.59375,
|
|
"step": 123
|
|
},
|
|
{
|
|
"epoch": 0.1521472392638037,
|
|
"grad_norm": 36.86795414024157,
|
|
"learning_rate": 4.795454545454545e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2623,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.75,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -2.78125,
|
|
"step": 124
|
|
},
|
|
{
|
|
"epoch": 0.15337423312883436,
|
|
"grad_norm": 41.75540815129781,
|
|
"learning_rate": 4.793388429752066e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2841,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.15460122699386503,
|
|
"grad_norm": 80.39792909106652,
|
|
"learning_rate": 4.791322314049586e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3527,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.83203125,
|
|
"rewards/margins": 3.578125,
|
|
"rewards/rejected": -2.75,
|
|
"step": 126
|
|
},
|
|
{
|
|
"epoch": 0.1558282208588957,
|
|
"grad_norm": 40.94598816709626,
|
|
"learning_rate": 4.789256198347108e-07,
|
|
"logits/chosen": -0.1103515625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3476,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 4.21875,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 127
|
|
},
|
|
{
|
|
"epoch": 0.1570552147239264,
|
|
"grad_norm": 39.47992513478673,
|
|
"learning_rate": 4.787190082644627e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2814,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.75,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.15828220858895706,
|
|
"grad_norm": 37.925898083738666,
|
|
"learning_rate": 4.785123966942149e-07,
|
|
"logits/chosen": -0.06396484375,
|
|
"logits/rejected": -0.1953125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.3421,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 3.546875,
|
|
"rewards/rejected": -2.421875,
|
|
"step": 129
|
|
},
|
|
{
|
|
"epoch": 0.15950920245398773,
|
|
"grad_norm": 38.0262211895724,
|
|
"learning_rate": 4.783057851239669e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.3115,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 3.96875,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 0.1607361963190184,
|
|
"grad_norm": 35.33717303960686,
|
|
"learning_rate": 4.78099173553719e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2154,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.0,
|
|
"step": 131
|
|
},
|
|
{
|
|
"epoch": 0.1619631901840491,
|
|
"grad_norm": 40.955300358810696,
|
|
"learning_rate": 4.77892561983471e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.314,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -2.84375,
|
|
"step": 132
|
|
},
|
|
{
|
|
"epoch": 0.16319018404907976,
|
|
"grad_norm": 43.6440596613334,
|
|
"learning_rate": 4.776859504132231e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2844,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.078125,
|
|
"step": 133
|
|
},
|
|
{
|
|
"epoch": 0.16441717791411042,
|
|
"grad_norm": 43.873623198146426,
|
|
"learning_rate": 4.774793388429752e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2948,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 3.96875,
|
|
"rewards/rejected": -2.59375,
|
|
"step": 134
|
|
},
|
|
{
|
|
"epoch": 0.1656441717791411,
|
|
"grad_norm": 44.973502848979926,
|
|
"learning_rate": 4.772727272727273e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.1865234375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3459,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 3.6875,
|
|
"rewards/rejected": -2.828125,
|
|
"step": 135
|
|
},
|
|
{
|
|
"epoch": 0.1668711656441718,
|
|
"grad_norm": 42.51816162906932,
|
|
"learning_rate": 4.770661157024793e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3308,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.078125,
|
|
"step": 136
|
|
},
|
|
{
|
|
"epoch": 0.16809815950920245,
|
|
"grad_norm": 48.104926340758944,
|
|
"learning_rate": 4.768595041322314e-07,
|
|
"logits/chosen": -0.080078125,
|
|
"logits/rejected": -0.2236328125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3458,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 3.640625,
|
|
"rewards/rejected": -2.515625,
|
|
"step": 137
|
|
},
|
|
{
|
|
"epoch": 0.16932515337423312,
|
|
"grad_norm": 38.354545275687684,
|
|
"learning_rate": 4.766528925619834e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.304,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 3.78125,
|
|
"rewards/rejected": -2.765625,
|
|
"step": 138
|
|
},
|
|
{
|
|
"epoch": 0.1705521472392638,
|
|
"grad_norm": 43.72111370882085,
|
|
"learning_rate": 4.764462809917355e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3477,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 139
|
|
},
|
|
{
|
|
"epoch": 0.17177914110429449,
|
|
"grad_norm": 49.390666427837495,
|
|
"learning_rate": 4.7623966942148756e-07,
|
|
"logits/chosen": -0.1201171875,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3233,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.90625,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 0.17300613496932515,
|
|
"grad_norm": 37.18049684192118,
|
|
"learning_rate": 4.7603305785123966e-07,
|
|
"logits/chosen": -0.10205078125,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2587,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 141
|
|
},
|
|
{
|
|
"epoch": 0.17423312883435582,
|
|
"grad_norm": 38.15326329982988,
|
|
"learning_rate": 4.758264462809917e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3083,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -3.0,
|
|
"step": 142
|
|
},
|
|
{
|
|
"epoch": 0.1754601226993865,
|
|
"grad_norm": 42.239479521280984,
|
|
"learning_rate": 4.756198347107438e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3086,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 143
|
|
},
|
|
{
|
|
"epoch": 0.17668711656441718,
|
|
"grad_norm": 36.71090019780343,
|
|
"learning_rate": 4.7541322314049586e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2799,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 144
|
|
},
|
|
{
|
|
"epoch": 0.17791411042944785,
|
|
"grad_norm": 39.65398403637813,
|
|
"learning_rate": 4.7520661157024796e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2858,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.953125,
|
|
"step": 145
|
|
},
|
|
{
|
|
"epoch": 0.17914110429447852,
|
|
"grad_norm": 36.72907127712686,
|
|
"learning_rate": 4.7499999999999995e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2676,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -3.15625,
|
|
"step": 146
|
|
},
|
|
{
|
|
"epoch": 0.18036809815950922,
|
|
"grad_norm": 35.42441477087175,
|
|
"learning_rate": 4.7479338842975205e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.275,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.203125,
|
|
"step": 147
|
|
},
|
|
{
|
|
"epoch": 0.18159509202453988,
|
|
"grad_norm": 51.82377713556493,
|
|
"learning_rate": 4.745867768595041e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3525,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 148
|
|
},
|
|
{
|
|
"epoch": 0.18282208588957055,
|
|
"grad_norm": 41.25940325564231,
|
|
"learning_rate": 4.743801652892562e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.253,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 149
|
|
},
|
|
{
|
|
"epoch": 0.18404907975460122,
|
|
"grad_norm": 43.207544914218985,
|
|
"learning_rate": 4.7417355371900825e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3367,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 4.03125,
|
|
"rewards/rejected": -2.625,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.18527607361963191,
|
|
"grad_norm": 42.2843854545005,
|
|
"learning_rate": 4.739669421487603e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.319,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.609375,
|
|
"step": 151
|
|
},
|
|
{
|
|
"epoch": 0.18650306748466258,
|
|
"grad_norm": 36.146243015735116,
|
|
"learning_rate": 4.737603305785124e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2584,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 152
|
|
},
|
|
{
|
|
"epoch": 0.18773006134969325,
|
|
"grad_norm": 38.30453694202532,
|
|
"learning_rate": 4.7355371900826444e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2846,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.0625,
|
|
"step": 153
|
|
},
|
|
{
|
|
"epoch": 0.18895705521472392,
|
|
"grad_norm": 48.193035851495836,
|
|
"learning_rate": 4.7334710743801654e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3123,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.734375,
|
|
"step": 154
|
|
},
|
|
{
|
|
"epoch": 0.1901840490797546,
|
|
"grad_norm": 27.96292821117509,
|
|
"learning_rate": 4.7314049586776853e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.1878,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.15625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.25,
|
|
"step": 155
|
|
},
|
|
{
|
|
"epoch": 0.19141104294478528,
|
|
"grad_norm": 42.366226442840826,
|
|
"learning_rate": 4.7293388429752063e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.2470703125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2881,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 3.953125,
|
|
"rewards/rejected": -2.765625,
|
|
"step": 156
|
|
},
|
|
{
|
|
"epoch": 0.19263803680981595,
|
|
"grad_norm": 40.72999796402848,
|
|
"learning_rate": 4.727272727272727e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2888,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 157
|
|
},
|
|
{
|
|
"epoch": 0.19386503067484662,
|
|
"grad_norm": 33.577674005153725,
|
|
"learning_rate": 4.725206611570248e-07,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.51171875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2863,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -3.15625,
|
|
"step": 158
|
|
},
|
|
{
|
|
"epoch": 0.1950920245398773,
|
|
"grad_norm": 43.09895886139825,
|
|
"learning_rate": 4.7231404958677683e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2941,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -2.84375,
|
|
"step": 159
|
|
},
|
|
{
|
|
"epoch": 0.19631901840490798,
|
|
"grad_norm": 33.061685198617056,
|
|
"learning_rate": 4.7210743801652893e-07,
|
|
"logits/chosen": -0.1201171875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2365,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 0.19754601226993865,
|
|
"grad_norm": 34.262461708297515,
|
|
"learning_rate": 4.71900826446281e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2321,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.078125,
|
|
"step": 161
|
|
},
|
|
{
|
|
"epoch": 0.19877300613496932,
|
|
"grad_norm": 46.44613400198569,
|
|
"learning_rate": 4.716942148760331e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3532,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.875,
|
|
"step": 162
|
|
},
|
|
{
|
|
"epoch": 0.2,
|
|
"grad_norm": 40.81709447673029,
|
|
"learning_rate": 4.7148760330578507e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2919,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 163
|
|
},
|
|
{
|
|
"epoch": 0.20122699386503068,
|
|
"grad_norm": 30.76141719102355,
|
|
"learning_rate": 4.7128099173553717e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.5234375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2276,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.28125,
|
|
"step": 164
|
|
},
|
|
{
|
|
"epoch": 0.20245398773006135,
|
|
"grad_norm": 39.308244617085876,
|
|
"learning_rate": 4.710743801652892e-07,
|
|
"logits/chosen": -0.09228515625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2561,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -2.78125,
|
|
"step": 165
|
|
},
|
|
{
|
|
"epoch": 0.20368098159509201,
|
|
"grad_norm": 40.482102907615214,
|
|
"learning_rate": 4.708677685950413e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.335,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.640625,
|
|
"step": 166
|
|
},
|
|
{
|
|
"epoch": 0.2049079754601227,
|
|
"grad_norm": 41.24091114160285,
|
|
"learning_rate": 4.7066115702479336e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3427,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.703125,
|
|
"step": 167
|
|
},
|
|
{
|
|
"epoch": 0.20613496932515338,
|
|
"grad_norm": 36.55347289055332,
|
|
"learning_rate": 4.704545454545454e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.236328125,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -270.0,
|
|
"loss": 0.2471,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.265625,
|
|
"step": 168
|
|
},
|
|
{
|
|
"epoch": 0.20736196319018405,
|
|
"grad_norm": 42.227854487848845,
|
|
"learning_rate": 4.702479338842975e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3043,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -3.0625,
|
|
"step": 169
|
|
},
|
|
{
|
|
"epoch": 0.2085889570552147,
|
|
"grad_norm": 41.26301353968675,
|
|
"learning_rate": 4.7004132231404956e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.282,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.015625,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 0.2098159509202454,
|
|
"grad_norm": 32.40263540928059,
|
|
"learning_rate": 4.6983471074380166e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.267,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 171
|
|
},
|
|
{
|
|
"epoch": 0.21104294478527608,
|
|
"grad_norm": 39.16155801484506,
|
|
"learning_rate": 4.6962809917355365e-07,
|
|
"logits/chosen": -0.1455078125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2624,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.828125,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 172
|
|
},
|
|
{
|
|
"epoch": 0.21226993865030674,
|
|
"grad_norm": 32.6638006062681,
|
|
"learning_rate": 4.6942148760330575e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2304,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.015625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -3.28125,
|
|
"step": 173
|
|
},
|
|
{
|
|
"epoch": 0.2134969325153374,
|
|
"grad_norm": 41.117611256664055,
|
|
"learning_rate": 4.692148760330578e-07,
|
|
"logits/chosen": -0.057861328125,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.15625,
|
|
"step": 174
|
|
},
|
|
{
|
|
"epoch": 0.2147239263803681,
|
|
"grad_norm": 43.97682714628577,
|
|
"learning_rate": 4.690082644628099e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3103,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -3.140625,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 0.21595092024539878,
|
|
"grad_norm": 39.95318826911176,
|
|
"learning_rate": 4.6880165289256195e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3105,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 176
|
|
},
|
|
{
|
|
"epoch": 0.21717791411042944,
|
|
"grad_norm": 31.646657225559906,
|
|
"learning_rate": 4.6859504132231405e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.233,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 177
|
|
},
|
|
{
|
|
"epoch": 0.2184049079754601,
|
|
"grad_norm": 41.80865379666853,
|
|
"learning_rate": 4.683884297520661e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3445,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 178
|
|
},
|
|
{
|
|
"epoch": 0.2196319018404908,
|
|
"grad_norm": 33.273068108107914,
|
|
"learning_rate": 4.681818181818182e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2545,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.8828125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 179
|
|
},
|
|
{
|
|
"epoch": 0.22085889570552147,
|
|
"grad_norm": 41.89399116746023,
|
|
"learning_rate": 4.6797520661157024e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.3137,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 0.22208588957055214,
|
|
"grad_norm": 44.979423553734385,
|
|
"learning_rate": 4.677685950413223e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3496,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 181
|
|
},
|
|
{
|
|
"epoch": 0.2233128834355828,
|
|
"grad_norm": 41.75351319386413,
|
|
"learning_rate": 4.6756198347107434e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3303,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -3.15625,
|
|
"step": 182
|
|
},
|
|
{
|
|
"epoch": 0.2245398773006135,
|
|
"grad_norm": 46.40866345124908,
|
|
"learning_rate": 4.6735537190082644e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.3255,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 3.953125,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 183
|
|
},
|
|
{
|
|
"epoch": 0.22576687116564417,
|
|
"grad_norm": 46.76226224495867,
|
|
"learning_rate": 4.671487603305785e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3388,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 0.78515625,
|
|
"rewards/margins": 3.984375,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 184
|
|
},
|
|
{
|
|
"epoch": 0.22699386503067484,
|
|
"grad_norm": 48.77760268554608,
|
|
"learning_rate": 4.669421487603306e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3219,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -2.859375,
|
|
"step": 185
|
|
},
|
|
{
|
|
"epoch": 0.2282208588957055,
|
|
"grad_norm": 38.08873895580722,
|
|
"learning_rate": 4.6673553719008263e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2901,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 186
|
|
},
|
|
{
|
|
"epoch": 0.2294478527607362,
|
|
"grad_norm": 40.67564497850856,
|
|
"learning_rate": 4.665289256198347e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2813,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.359375,
|
|
"step": 187
|
|
},
|
|
{
|
|
"epoch": 0.23067484662576687,
|
|
"grad_norm": 35.35467709831456,
|
|
"learning_rate": 4.663223140495868e-07,
|
|
"logits/chosen": -0.09716796875,
|
|
"logits/rejected": -0.189453125,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.2813,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 188
|
|
},
|
|
{
|
|
"epoch": 0.23190184049079754,
|
|
"grad_norm": 32.02217527416385,
|
|
"learning_rate": 4.661157024793388e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2159,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.34375,
|
|
"step": 189
|
|
},
|
|
{
|
|
"epoch": 0.2331288343558282,
|
|
"grad_norm": 35.517601629388636,
|
|
"learning_rate": 4.6590909090909087e-07,
|
|
"logits/chosen": -0.08154296875,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2873,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 0.2343558282208589,
|
|
"grad_norm": 46.62318205281028,
|
|
"learning_rate": 4.657024793388429e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3248,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.203125,
|
|
"step": 191
|
|
},
|
|
{
|
|
"epoch": 0.23558282208588957,
|
|
"grad_norm": 47.302561931824144,
|
|
"learning_rate": 4.65495867768595e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3212,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.03125,
|
|
"rewards/rejected": -2.75,
|
|
"step": 192
|
|
},
|
|
{
|
|
"epoch": 0.23680981595092024,
|
|
"grad_norm": 38.194528570730164,
|
|
"learning_rate": 4.6528925619834707e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.5234375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2332,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 193
|
|
},
|
|
{
|
|
"epoch": 0.23803680981595093,
|
|
"grad_norm": 42.44224201595828,
|
|
"learning_rate": 4.6508264462809917e-07,
|
|
"logits/chosen": -0.32421875,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3223,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 194
|
|
},
|
|
{
|
|
"epoch": 0.2392638036809816,
|
|
"grad_norm": 34.0790216803257,
|
|
"learning_rate": 4.648760330578512e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.219,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.0,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -3.5,
|
|
"step": 195
|
|
},
|
|
{
|
|
"epoch": 0.24049079754601227,
|
|
"grad_norm": 43.004434879940106,
|
|
"learning_rate": 4.646694214876033e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.3054,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.0,
|
|
"step": 196
|
|
},
|
|
{
|
|
"epoch": 0.24171779141104294,
|
|
"grad_norm": 34.15269412729779,
|
|
"learning_rate": 4.6446280991735536e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2756,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 197
|
|
},
|
|
{
|
|
"epoch": 0.24294478527607363,
|
|
"grad_norm": 36.78925425492936,
|
|
"learning_rate": 4.6425619834710746e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.255859375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -250.0,
|
|
"loss": 0.2945,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -2.9375,
|
|
"step": 198
|
|
},
|
|
{
|
|
"epoch": 0.2441717791411043,
|
|
"grad_norm": 28.311028034096932,
|
|
"learning_rate": 4.6404958677685945e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2045,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.125,
|
|
"step": 199
|
|
},
|
|
{
|
|
"epoch": 0.24539877300613497,
|
|
"grad_norm": 44.46315305804351,
|
|
"learning_rate": 4.6384297520661155e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.303,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9921875,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.24662576687116564,
|
|
"grad_norm": 32.929953461927006,
|
|
"learning_rate": 4.636363636363636e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2219,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -2.953125,
|
|
"step": 201
|
|
},
|
|
{
|
|
"epoch": 0.24785276073619633,
|
|
"grad_norm": 38.191540513799985,
|
|
"learning_rate": 4.634297520661157e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2735,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9453125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 202
|
|
},
|
|
{
|
|
"epoch": 0.249079754601227,
|
|
"grad_norm": 43.899916034277716,
|
|
"learning_rate": 4.6322314049586775e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3715,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.46875,
|
|
"step": 203
|
|
},
|
|
{
|
|
"epoch": 0.25030674846625767,
|
|
"grad_norm": 40.42534652251951,
|
|
"learning_rate": 4.630165289256198e-07,
|
|
"logits/chosen": -0.10498046875,
|
|
"logits/rejected": -0.2490234375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.314,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -2.921875,
|
|
"step": 204
|
|
},
|
|
{
|
|
"epoch": 0.25153374233128833,
|
|
"grad_norm": 37.717470986712556,
|
|
"learning_rate": 4.628099173553719e-07,
|
|
"logits/chosen": -0.059326171875,
|
|
"logits/rejected": -0.2001953125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3111,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -2.4375,
|
|
"step": 205
|
|
},
|
|
{
|
|
"epoch": 0.252760736196319,
|
|
"grad_norm": 30.55406782868523,
|
|
"learning_rate": 4.6260330578512394e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2196,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 2.0625,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.21875,
|
|
"step": 206
|
|
},
|
|
{
|
|
"epoch": 0.25398773006134967,
|
|
"grad_norm": 44.99416041741556,
|
|
"learning_rate": 4.6239669421487604e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3683,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 207
|
|
},
|
|
{
|
|
"epoch": 0.2552147239263804,
|
|
"grad_norm": 38.298643600670886,
|
|
"learning_rate": 4.6219008264462804e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 208
|
|
},
|
|
{
|
|
"epoch": 0.25644171779141106,
|
|
"grad_norm": 42.08974951434723,
|
|
"learning_rate": 4.6198347107438014e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3252,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.84375,
|
|
"step": 209
|
|
},
|
|
{
|
|
"epoch": 0.25766871165644173,
|
|
"grad_norm": 49.252789458115785,
|
|
"learning_rate": 4.617768595041322e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3478,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -2.8125,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 0.2588957055214724,
|
|
"grad_norm": 35.19521189692808,
|
|
"learning_rate": 4.615702479338843e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2943,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.7421875,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.25,
|
|
"step": 211
|
|
},
|
|
{
|
|
"epoch": 0.26012269938650306,
|
|
"grad_norm": 38.588101334546415,
|
|
"learning_rate": 4.6136363636363633e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2772,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 212
|
|
},
|
|
{
|
|
"epoch": 0.26134969325153373,
|
|
"grad_norm": 36.19543986527335,
|
|
"learning_rate": 4.6115702479338843e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2568,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -3.015625,
|
|
"step": 213
|
|
},
|
|
{
|
|
"epoch": 0.2625766871165644,
|
|
"grad_norm": 36.76253302400916,
|
|
"learning_rate": 4.609504132231405e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2624,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6953125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.125,
|
|
"step": 214
|
|
},
|
|
{
|
|
"epoch": 0.26380368098159507,
|
|
"grad_norm": 37.194255093432446,
|
|
"learning_rate": 4.607438016528926e-07,
|
|
"logits/chosen": -0.111328125,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2725,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -3.09375,
|
|
"step": 215
|
|
},
|
|
{
|
|
"epoch": 0.2650306748466258,
|
|
"grad_norm": 37.819502778321414,
|
|
"learning_rate": 4.605371900826446e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2757,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 4.03125,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 216
|
|
},
|
|
{
|
|
"epoch": 0.26625766871165646,
|
|
"grad_norm": 38.426937338322844,
|
|
"learning_rate": 4.6033057851239667e-07,
|
|
"logits/chosen": -0.072265625,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3129,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -2.890625,
|
|
"step": 217
|
|
},
|
|
{
|
|
"epoch": 0.2674846625766871,
|
|
"grad_norm": 48.711804010557216,
|
|
"learning_rate": 4.601239669421487e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3689,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 4.0625,
|
|
"rewards/rejected": -2.9375,
|
|
"step": 218
|
|
},
|
|
{
|
|
"epoch": 0.2687116564417178,
|
|
"grad_norm": 34.859463323248626,
|
|
"learning_rate": 4.599173553719008e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2779,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.5,
|
|
"step": 219
|
|
},
|
|
{
|
|
"epoch": 0.26993865030674846,
|
|
"grad_norm": 46.80026994213379,
|
|
"learning_rate": 4.5971074380165287e-07,
|
|
"logits/chosen": -0.07568359375,
|
|
"logits/rejected": -0.1572265625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3534,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 3.765625,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 0.27116564417177913,
|
|
"grad_norm": 44.236722785199355,
|
|
"learning_rate": 4.595041322314049e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3432,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 221
|
|
},
|
|
{
|
|
"epoch": 0.2723926380368098,
|
|
"grad_norm": 37.55504945465003,
|
|
"learning_rate": 4.59297520661157e-07,
|
|
"logits/chosen": -0.33203125,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2529,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 222
|
|
},
|
|
{
|
|
"epoch": 0.27361963190184047,
|
|
"grad_norm": 51.564475699604486,
|
|
"learning_rate": 4.5909090909090906e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3058,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.953125,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 223
|
|
},
|
|
{
|
|
"epoch": 0.2748466257668712,
|
|
"grad_norm": 43.15791338185834,
|
|
"learning_rate": 4.5888429752066116e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.435546875,
|
|
"logps/chosen": -450.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2807,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 224
|
|
},
|
|
{
|
|
"epoch": 0.27607361963190186,
|
|
"grad_norm": 35.879838019408076,
|
|
"learning_rate": 4.586776859504132e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2506,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 0.2773006134969325,
|
|
"grad_norm": 44.55257679536464,
|
|
"learning_rate": 4.5847107438016525e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3177,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 226
|
|
},
|
|
{
|
|
"epoch": 0.2785276073619632,
|
|
"grad_norm": 43.069867055252956,
|
|
"learning_rate": 4.582644628099173e-07,
|
|
"logits/chosen": -0.17578125,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2933,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.875,
|
|
"step": 227
|
|
},
|
|
{
|
|
"epoch": 0.27975460122699386,
|
|
"grad_norm": 59.89729428981541,
|
|
"learning_rate": 4.580578512396694e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.324,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.62109375,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 228
|
|
},
|
|
{
|
|
"epoch": 0.2809815950920245,
|
|
"grad_norm": 41.26714714905276,
|
|
"learning_rate": 4.5785123966942145e-07,
|
|
"logits/chosen": -0.1025390625,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.27,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 229
|
|
},
|
|
{
|
|
"epoch": 0.2822085889570552,
|
|
"grad_norm": 33.72568576392679,
|
|
"learning_rate": 4.5764462809917355e-07,
|
|
"logits/chosen": -0.08935546875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.235,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.75,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 0.28343558282208586,
|
|
"grad_norm": 42.97736121941397,
|
|
"learning_rate": 4.574380165289256e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3436,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 231
|
|
},
|
|
{
|
|
"epoch": 0.2846625766871166,
|
|
"grad_norm": 42.93355591121654,
|
|
"learning_rate": 4.572314049586777e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3127,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.5546875,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -3.75,
|
|
"step": 232
|
|
},
|
|
{
|
|
"epoch": 0.28588957055214725,
|
|
"grad_norm": 37.7852023248725,
|
|
"learning_rate": 4.5702479338842974e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.2759,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 233
|
|
},
|
|
{
|
|
"epoch": 0.2871165644171779,
|
|
"grad_norm": 38.62852043438936,
|
|
"learning_rate": 4.5681818181818184e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2764,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.578125,
|
|
"step": 234
|
|
},
|
|
{
|
|
"epoch": 0.2883435582822086,
|
|
"grad_norm": 37.335415454122945,
|
|
"learning_rate": 4.5661157024793384e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2629,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 235
|
|
},
|
|
{
|
|
"epoch": 0.28957055214723926,
|
|
"grad_norm": 55.00073114295456,
|
|
"learning_rate": 4.5640495867768594e-07,
|
|
"logits/chosen": -0.04736328125,
|
|
"logits/rejected": -0.232421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.374,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.859375,
|
|
"step": 236
|
|
},
|
|
{
|
|
"epoch": 0.2907975460122699,
|
|
"grad_norm": 34.92202343855084,
|
|
"learning_rate": 4.56198347107438e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.252,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.578125,
|
|
"step": 237
|
|
},
|
|
{
|
|
"epoch": 0.2920245398773006,
|
|
"grad_norm": 42.067317323205096,
|
|
"learning_rate": 4.559917355371901e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3198,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.34375,
|
|
"step": 238
|
|
},
|
|
{
|
|
"epoch": 0.29325153374233126,
|
|
"grad_norm": 41.04109475522015,
|
|
"learning_rate": 4.5578512396694213e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3007,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 239
|
|
},
|
|
{
|
|
"epoch": 0.294478527607362,
|
|
"grad_norm": 42.47439204624692,
|
|
"learning_rate": 4.555785123966942e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2516,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 0.29570552147239265,
|
|
"grad_norm": 38.63581064698526,
|
|
"learning_rate": 4.553719008264463e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2718,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.5,
|
|
"step": 241
|
|
},
|
|
{
|
|
"epoch": 0.2969325153374233,
|
|
"grad_norm": 31.580876023766372,
|
|
"learning_rate": 4.551652892561983e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2369,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 242
|
|
},
|
|
{
|
|
"epoch": 0.298159509202454,
|
|
"grad_norm": 38.48835239358082,
|
|
"learning_rate": 4.5495867768595037e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2935,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.671875,
|
|
"step": 243
|
|
},
|
|
{
|
|
"epoch": 0.29938650306748466,
|
|
"grad_norm": 41.633020456629765,
|
|
"learning_rate": 4.547520661157024e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3582,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.5,
|
|
"step": 244
|
|
},
|
|
{
|
|
"epoch": 0.3006134969325153,
|
|
"grad_norm": 36.05397433070667,
|
|
"learning_rate": 4.545454545454545e-07,
|
|
"logits/chosen": -0.220703125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.234,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 245
|
|
},
|
|
{
|
|
"epoch": 0.301840490797546,
|
|
"grad_norm": 38.58934325410754,
|
|
"learning_rate": 4.5433884297520657e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2805,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 246
|
|
},
|
|
{
|
|
"epoch": 0.3030674846625767,
|
|
"grad_norm": 46.170807958237866,
|
|
"learning_rate": 4.5413223140495867e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3061,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 247
|
|
},
|
|
{
|
|
"epoch": 0.3042944785276074,
|
|
"grad_norm": 40.70518240157979,
|
|
"learning_rate": 4.539256198347107e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2562,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 248
|
|
},
|
|
{
|
|
"epoch": 0.30552147239263805,
|
|
"grad_norm": 43.152905506219625,
|
|
"learning_rate": 4.537190082644628e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.466796875,
|
|
"logps/chosen": -480.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.274,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.625,
|
|
"step": 249
|
|
},
|
|
{
|
|
"epoch": 0.3067484662576687,
|
|
"grad_norm": 39.51182461700564,
|
|
"learning_rate": 4.5351239669421486e-07,
|
|
"logits/chosen": -0.349609375,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2913,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.97265625,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.625,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.3079754601226994,
|
|
"grad_norm": 47.42800586657486,
|
|
"learning_rate": 4.5330578512396696e-07,
|
|
"logits/chosen": -0.055908203125,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3574,
|
|
"rewards/accuracies": 0.765625,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 4.1875,
|
|
"rewards/rejected": -3.046875,
|
|
"step": 251
|
|
},
|
|
{
|
|
"epoch": 0.30920245398773005,
|
|
"grad_norm": 38.9167492571078,
|
|
"learning_rate": 4.5309917355371896e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2806,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.234375,
|
|
"step": 252
|
|
},
|
|
{
|
|
"epoch": 0.3104294478527607,
|
|
"grad_norm": 31.88015147725951,
|
|
"learning_rate": 4.5289256198347106e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2435,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 253
|
|
},
|
|
{
|
|
"epoch": 0.3116564417177914,
|
|
"grad_norm": 41.38320327055996,
|
|
"learning_rate": 4.526859504132231e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3288,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.46875,
|
|
"step": 254
|
|
},
|
|
{
|
|
"epoch": 0.3128834355828221,
|
|
"grad_norm": 39.369898300453734,
|
|
"learning_rate": 4.524793388429752e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3362,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.046875,
|
|
"step": 255
|
|
},
|
|
{
|
|
"epoch": 0.3141104294478528,
|
|
"grad_norm": 32.76851278241994,
|
|
"learning_rate": 4.5227272727272725e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2341,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.4375,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 0.31533742331288345,
|
|
"grad_norm": 45.250746891569236,
|
|
"learning_rate": 4.520661157024793e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3181,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.7734375,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 257
|
|
},
|
|
{
|
|
"epoch": 0.3165644171779141,
|
|
"grad_norm": 35.04986410208359,
|
|
"learning_rate": 4.518595041322314e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2317,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 258
|
|
},
|
|
{
|
|
"epoch": 0.3177914110429448,
|
|
"grad_norm": 34.37988001182442,
|
|
"learning_rate": 4.5165289256198344e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2627,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 259
|
|
},
|
|
{
|
|
"epoch": 0.31901840490797545,
|
|
"grad_norm": 39.73259510340199,
|
|
"learning_rate": 4.5144628099173554e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2938,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 0.3202453987730061,
|
|
"grad_norm": 42.63140351827292,
|
|
"learning_rate": 4.5123966942148754e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2956,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4453125,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.203125,
|
|
"step": 261
|
|
},
|
|
{
|
|
"epoch": 0.3214723926380368,
|
|
"grad_norm": 48.00556286038852,
|
|
"learning_rate": 4.5103305785123964e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3166,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.0625,
|
|
"step": 262
|
|
},
|
|
{
|
|
"epoch": 0.3226993865030675,
|
|
"grad_norm": 41.40798456165999,
|
|
"learning_rate": 4.508264462809917e-07,
|
|
"logits/chosen": -0.197265625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2953,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 4.03125,
|
|
"rewards/rejected": -2.921875,
|
|
"step": 263
|
|
},
|
|
{
|
|
"epoch": 0.3239263803680982,
|
|
"grad_norm": 39.73351721310888,
|
|
"learning_rate": 4.506198347107438e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3138,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.96875,
|
|
"step": 264
|
|
},
|
|
{
|
|
"epoch": 0.32515337423312884,
|
|
"grad_norm": 39.72934408632108,
|
|
"learning_rate": 4.5041322314049583e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.2712,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -3.1875,
|
|
"step": 265
|
|
},
|
|
{
|
|
"epoch": 0.3263803680981595,
|
|
"grad_norm": 47.72732688002189,
|
|
"learning_rate": 4.5020661157024793e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3794,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.64453125,
|
|
"rewards/margins": 3.625,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 266
|
|
},
|
|
{
|
|
"epoch": 0.3276073619631902,
|
|
"grad_norm": 52.825089270201424,
|
|
"learning_rate": 4.5e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3429,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -2.734375,
|
|
"step": 267
|
|
},
|
|
{
|
|
"epoch": 0.32883435582822085,
|
|
"grad_norm": 35.92749837261359,
|
|
"learning_rate": 4.497933884297521e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2275,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.109375,
|
|
"step": 268
|
|
},
|
|
{
|
|
"epoch": 0.3300613496932515,
|
|
"grad_norm": 35.81499005195925,
|
|
"learning_rate": 4.4958677685950413e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.50390625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2385,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.25,
|
|
"step": 269
|
|
},
|
|
{
|
|
"epoch": 0.3312883435582822,
|
|
"grad_norm": 44.38655671717473,
|
|
"learning_rate": 4.493801652892562e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.345,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 3.71875,
|
|
"rewards/rejected": -2.5625,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 0.3325153374233129,
|
|
"grad_norm": 34.101391505637224,
|
|
"learning_rate": 4.491735537190082e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2456,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -2.796875,
|
|
"step": 271
|
|
},
|
|
{
|
|
"epoch": 0.3337423312883436,
|
|
"grad_norm": 36.8738947032395,
|
|
"learning_rate": 4.489669421487603e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2495,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.25,
|
|
"rewards/margins": 4.21875,
|
|
"rewards/rejected": -2.984375,
|
|
"step": 272
|
|
},
|
|
{
|
|
"epoch": 0.33496932515337424,
|
|
"grad_norm": 44.70898948794442,
|
|
"learning_rate": 4.4876033057851237e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.1611328125,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.3884,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 3.90625,
|
|
"rewards/rejected": -2.734375,
|
|
"step": 273
|
|
},
|
|
{
|
|
"epoch": 0.3361963190184049,
|
|
"grad_norm": 42.40600245146773,
|
|
"learning_rate": 4.4855371900826447e-07,
|
|
"logits/chosen": -0.0927734375,
|
|
"logits/rejected": -0.2421875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3338,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -2.484375,
|
|
"step": 274
|
|
},
|
|
{
|
|
"epoch": 0.3374233128834356,
|
|
"grad_norm": 40.8103963168886,
|
|
"learning_rate": 4.483471074380165e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2999,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 4.15625,
|
|
"rewards/rejected": -2.6875,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 0.33865030674846625,
|
|
"grad_norm": 31.049107272836952,
|
|
"learning_rate": 4.4814049586776856e-07,
|
|
"logits/chosen": -0.08154296875,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2799,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 4.3125,
|
|
"rewards/rejected": -2.75,
|
|
"step": 276
|
|
},
|
|
{
|
|
"epoch": 0.3398773006134969,
|
|
"grad_norm": 32.630832893605,
|
|
"learning_rate": 4.4793388429752066e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2319,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.0,
|
|
"step": 277
|
|
},
|
|
{
|
|
"epoch": 0.3411042944785276,
|
|
"grad_norm": 44.309610437735856,
|
|
"learning_rate": 4.477272727272727e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.2373046875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2724,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 4.46875,
|
|
"rewards/rejected": -2.828125,
|
|
"step": 278
|
|
},
|
|
{
|
|
"epoch": 0.3423312883435583,
|
|
"grad_norm": 26.353655721882397,
|
|
"learning_rate": 4.4752066115702476e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.1863,
|
|
"rewards/accuracies": 0.9453125,
|
|
"rewards/chosen": 1.7578125,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.21875,
|
|
"step": 279
|
|
},
|
|
{
|
|
"epoch": 0.34355828220858897,
|
|
"grad_norm": 44.718053283042785,
|
|
"learning_rate": 4.473140495867768e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3351,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 1.953125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.171875,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 0.34478527607361964,
|
|
"grad_norm": 42.34591075354157,
|
|
"learning_rate": 4.471074380165289e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3447,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 4.125,
|
|
"rewards/rejected": -2.6875,
|
|
"step": 281
|
|
},
|
|
{
|
|
"epoch": 0.3460122699386503,
|
|
"grad_norm": 41.75221231669429,
|
|
"learning_rate": 4.4690082644628095e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2518,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.71875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 282
|
|
},
|
|
{
|
|
"epoch": 0.347239263803681,
|
|
"grad_norm": 33.916128769423274,
|
|
"learning_rate": 4.4669421487603305e-07,
|
|
"logits/chosen": -0.3359375,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2746,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -3.046875,
|
|
"step": 283
|
|
},
|
|
{
|
|
"epoch": 0.34846625766871164,
|
|
"grad_norm": 35.84425714657199,
|
|
"learning_rate": 4.464876033057851e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2426,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 284
|
|
},
|
|
{
|
|
"epoch": 0.3496932515337423,
|
|
"grad_norm": 31.897221284753925,
|
|
"learning_rate": 4.462809917355372e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2458,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 285
|
|
},
|
|
{
|
|
"epoch": 0.350920245398773,
|
|
"grad_norm": 40.0301906925459,
|
|
"learning_rate": 4.4607438016528924e-07,
|
|
"logits/chosen": -0.1708984375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3071,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 286
|
|
},
|
|
{
|
|
"epoch": 0.3521472392638037,
|
|
"grad_norm": 50.881866049082234,
|
|
"learning_rate": 4.4586776859504135e-07,
|
|
"logits/chosen": -0.26171875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.288,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 287
|
|
},
|
|
{
|
|
"epoch": 0.35337423312883437,
|
|
"grad_norm": 37.27905667507029,
|
|
"learning_rate": 4.4566115702479334e-07,
|
|
"logits/chosen": -0.08642578125,
|
|
"logits/rejected": -0.2392578125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2887,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.99609375,
|
|
"rewards/margins": 4.25,
|
|
"rewards/rejected": -3.265625,
|
|
"step": 288
|
|
},
|
|
{
|
|
"epoch": 0.35460122699386504,
|
|
"grad_norm": 38.51669077378738,
|
|
"learning_rate": 4.4545454545454544e-07,
|
|
"logits/chosen": -0.2158203125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2274,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 289
|
|
},
|
|
{
|
|
"epoch": 0.3558282208588957,
|
|
"grad_norm": 39.101756003947045,
|
|
"learning_rate": 4.452479338842975e-07,
|
|
"logits/chosen": -0.3828125,
|
|
"logits/rejected": -0.53125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.197,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 0.3570552147239264,
|
|
"grad_norm": 35.84563757340601,
|
|
"learning_rate": 4.450413223140496e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.264,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 2.03125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 291
|
|
},
|
|
{
|
|
"epoch": 0.35828220858895704,
|
|
"grad_norm": 48.83799779484981,
|
|
"learning_rate": 4.4483471074380163e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2282,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 292
|
|
},
|
|
{
|
|
"epoch": 0.3595092024539877,
|
|
"grad_norm": 33.24417072591366,
|
|
"learning_rate": 4.446280991735537e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.236,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 293
|
|
},
|
|
{
|
|
"epoch": 0.36073619631901843,
|
|
"grad_norm": 33.17016716719111,
|
|
"learning_rate": 4.444214876033058e-07,
|
|
"logits/chosen": -0.443359375,
|
|
"logits/rejected": -0.59765625,
|
|
"logps/chosen": -464.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2358,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 294
|
|
},
|
|
{
|
|
"epoch": 0.3619631901840491,
|
|
"grad_norm": 41.08997488080215,
|
|
"learning_rate": 4.4421487603305783e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2229,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 295
|
|
},
|
|
{
|
|
"epoch": 0.36319018404907977,
|
|
"grad_norm": 36.46585175010118,
|
|
"learning_rate": 4.4400826446280993e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2845,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 296
|
|
},
|
|
{
|
|
"epoch": 0.36441717791411044,
|
|
"grad_norm": 30.507553736995185,
|
|
"learning_rate": 4.438016528925619e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2352,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 297
|
|
},
|
|
{
|
|
"epoch": 0.3656441717791411,
|
|
"grad_norm": 42.744426202574054,
|
|
"learning_rate": 4.43595041322314e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2875,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 298
|
|
},
|
|
{
|
|
"epoch": 0.36687116564417177,
|
|
"grad_norm": 57.11420288456295,
|
|
"learning_rate": 4.4338842975206607e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3303,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 299
|
|
},
|
|
{
|
|
"epoch": 0.36809815950920244,
|
|
"grad_norm": 44.0642342343628,
|
|
"learning_rate": 4.4318181818181817e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2951,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.3693251533742331,
|
|
"grad_norm": 36.50971474693751,
|
|
"learning_rate": 4.429752066115702e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 301
|
|
},
|
|
{
|
|
"epoch": 0.37055214723926383,
|
|
"grad_norm": 31.988576378738134,
|
|
"learning_rate": 4.427685950413223e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2208,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 302
|
|
},
|
|
{
|
|
"epoch": 0.3717791411042945,
|
|
"grad_norm": 46.63040649021854,
|
|
"learning_rate": 4.4256198347107436e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3215,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -4.125,
|
|
"step": 303
|
|
},
|
|
{
|
|
"epoch": 0.37300613496932516,
|
|
"grad_norm": 38.26286144297621,
|
|
"learning_rate": 4.4235537190082646e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.244,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 304
|
|
},
|
|
{
|
|
"epoch": 0.37423312883435583,
|
|
"grad_norm": 36.3552235356213,
|
|
"learning_rate": 4.421487603305785e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2568,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 305
|
|
},
|
|
{
|
|
"epoch": 0.3754601226993865,
|
|
"grad_norm": 40.279423014087726,
|
|
"learning_rate": 4.4194214876033056e-07,
|
|
"logits/chosen": -0.322265625,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2584,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 306
|
|
},
|
|
{
|
|
"epoch": 0.37668711656441717,
|
|
"grad_norm": 39.63802917296688,
|
|
"learning_rate": 4.417355371900826e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3214,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.0,
|
|
"step": 307
|
|
},
|
|
{
|
|
"epoch": 0.37791411042944784,
|
|
"grad_norm": 38.40390766672212,
|
|
"learning_rate": 4.415289256198347e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3262,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.95703125,
|
|
"rewards/margins": 4.28125,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 308
|
|
},
|
|
{
|
|
"epoch": 0.3791411042944785,
|
|
"grad_norm": 32.4379821627175,
|
|
"learning_rate": 4.4132231404958675e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2098,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 309
|
|
},
|
|
{
|
|
"epoch": 0.3803680981595092,
|
|
"grad_norm": 40.47301853181884,
|
|
"learning_rate": 4.411157024793388e-07,
|
|
"logits/chosen": -0.107421875,
|
|
"logits/rejected": -0.205078125,
|
|
"logps/chosen": -310.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3181,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.32421875,
|
|
"rewards/margins": 4.0,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 0.3815950920245399,
|
|
"grad_norm": 33.919361023276416,
|
|
"learning_rate": 4.409090909090909e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2306,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 311
|
|
},
|
|
{
|
|
"epoch": 0.38282208588957056,
|
|
"grad_norm": 48.72311046374392,
|
|
"learning_rate": 4.4070247933884295e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3689,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.40625,
|
|
"step": 312
|
|
},
|
|
{
|
|
"epoch": 0.38404907975460123,
|
|
"grad_norm": 32.93519257063822,
|
|
"learning_rate": 4.4049586776859505e-07,
|
|
"logits/chosen": -0.134765625,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2417,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 313
|
|
},
|
|
{
|
|
"epoch": 0.3852760736196319,
|
|
"grad_norm": 29.761092790659628,
|
|
"learning_rate": 4.402892561983471e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.1846,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 2.515625,
|
|
"rewards/margins": 6.71875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 314
|
|
},
|
|
{
|
|
"epoch": 0.38650306748466257,
|
|
"grad_norm": 33.15936964822026,
|
|
"learning_rate": 4.4008264462809914e-07,
|
|
"logits/chosen": -0.06982421875,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -274.0,
|
|
"loss": 0.2307,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 315
|
|
},
|
|
{
|
|
"epoch": 0.38773006134969323,
|
|
"grad_norm": 30.385645470245766,
|
|
"learning_rate": 4.398760330578512e-07,
|
|
"logits/chosen": -0.07666015625,
|
|
"logits/rejected": -0.2197265625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.2284,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 316
|
|
},
|
|
{
|
|
"epoch": 0.3889570552147239,
|
|
"grad_norm": 34.30838449187616,
|
|
"learning_rate": 4.396694214876033e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2337,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 317
|
|
},
|
|
{
|
|
"epoch": 0.3901840490797546,
|
|
"grad_norm": 43.51216041497958,
|
|
"learning_rate": 4.3946280991735533e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3148,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.625,
|
|
"step": 318
|
|
},
|
|
{
|
|
"epoch": 0.3914110429447853,
|
|
"grad_norm": 23.822105871778678,
|
|
"learning_rate": 4.3925619834710743e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.157,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 319
|
|
},
|
|
{
|
|
"epoch": 0.39263803680981596,
|
|
"grad_norm": 37.656689387782905,
|
|
"learning_rate": 4.390495867768595e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -316.0,
|
|
"logps/rejected": -262.0,
|
|
"loss": 0.2576,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.8046875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 0.39386503067484663,
|
|
"grad_norm": 47.198644192326896,
|
|
"learning_rate": 4.388429752066116e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.2837,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 321
|
|
},
|
|
{
|
|
"epoch": 0.3950920245398773,
|
|
"grad_norm": 35.22081458129445,
|
|
"learning_rate": 4.3863636363636363e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2397,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 322
|
|
},
|
|
{
|
|
"epoch": 0.39631901840490796,
|
|
"grad_norm": 40.535682371905,
|
|
"learning_rate": 4.3842975206611573e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3459,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.9453125,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 323
|
|
},
|
|
{
|
|
"epoch": 0.39754601226993863,
|
|
"grad_norm": 34.724438451994665,
|
|
"learning_rate": 4.382231404958677e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.453125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2659,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8671875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 324
|
|
},
|
|
{
|
|
"epoch": 0.3987730061349693,
|
|
"grad_norm": 47.440215471486304,
|
|
"learning_rate": 4.380165289256198e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.228515625,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.383,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"grad_norm": 36.66273937163835,
|
|
"learning_rate": 4.3780991735537187e-07,
|
|
"logits/chosen": -0.31640625,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2043,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.7265625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 326
|
|
},
|
|
{
|
|
"epoch": 0.4012269938650307,
|
|
"grad_norm": 51.442585409024396,
|
|
"learning_rate": 4.3760330578512397e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3681,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -3.4375,
|
|
"step": 327
|
|
},
|
|
{
|
|
"epoch": 0.40245398773006136,
|
|
"grad_norm": 44.606600240172085,
|
|
"learning_rate": 4.37396694214876e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.3204,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.59375,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -4.0,
|
|
"step": 328
|
|
},
|
|
{
|
|
"epoch": 0.403680981595092,
|
|
"grad_norm": 36.3592106715033,
|
|
"learning_rate": 4.3719008264462806e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -458.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2828,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 329
|
|
},
|
|
{
|
|
"epoch": 0.4049079754601227,
|
|
"grad_norm": 36.0973250750226,
|
|
"learning_rate": 4.3698347107438016e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2257,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 0.40613496932515336,
|
|
"grad_norm": 35.83835721797351,
|
|
"learning_rate": 4.367768595041322e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2402,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 331
|
|
},
|
|
{
|
|
"epoch": 0.40736196319018403,
|
|
"grad_norm": 29.00587176168557,
|
|
"learning_rate": 4.3657024793388426e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2051,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.703125,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 332
|
|
},
|
|
{
|
|
"epoch": 0.4085889570552147,
|
|
"grad_norm": 37.2793403072948,
|
|
"learning_rate": 4.363636363636363e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2787,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 333
|
|
},
|
|
{
|
|
"epoch": 0.4098159509202454,
|
|
"grad_norm": 37.233291436969324,
|
|
"learning_rate": 4.361570247933884e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.263,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 334
|
|
},
|
|
{
|
|
"epoch": 0.4110429447852761,
|
|
"grad_norm": 40.307147676871146,
|
|
"learning_rate": 4.3595041322314045e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2393,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.9375,
|
|
"step": 335
|
|
},
|
|
{
|
|
"epoch": 0.41226993865030676,
|
|
"grad_norm": 35.98003323778127,
|
|
"learning_rate": 4.3574380165289255e-07,
|
|
"logits/chosen": -0.3046875,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2573,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 336
|
|
},
|
|
{
|
|
"epoch": 0.4134969325153374,
|
|
"grad_norm": 47.71731749469287,
|
|
"learning_rate": 4.355371900826446e-07,
|
|
"logits/chosen": -0.10595703125,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3291,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 337
|
|
},
|
|
{
|
|
"epoch": 0.4147239263803681,
|
|
"grad_norm": 40.237231718518984,
|
|
"learning_rate": 4.353305785123967e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3113,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 338
|
|
},
|
|
{
|
|
"epoch": 0.41595092024539876,
|
|
"grad_norm": 43.74604259781836,
|
|
"learning_rate": 4.3512396694214875e-07,
|
|
"logits/chosen": -0.08935546875,
|
|
"logits/rejected": -0.2275390625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3036,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.796875,
|
|
"step": 339
|
|
},
|
|
{
|
|
"epoch": 0.4171779141104294,
|
|
"grad_norm": 40.57164499134951,
|
|
"learning_rate": 4.3491735537190085e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2623,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 0.41840490797546015,
|
|
"grad_norm": 35.445167805912895,
|
|
"learning_rate": 4.3471074380165284e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2979,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 341
|
|
},
|
|
{
|
|
"epoch": 0.4196319018404908,
|
|
"grad_norm": 43.283231525706334,
|
|
"learning_rate": 4.3450413223140494e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3065,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 342
|
|
},
|
|
{
|
|
"epoch": 0.4208588957055215,
|
|
"grad_norm": 41.84183041867056,
|
|
"learning_rate": 4.34297520661157e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3363,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -3.328125,
|
|
"step": 343
|
|
},
|
|
{
|
|
"epoch": 0.42208588957055215,
|
|
"grad_norm": 40.634867535224096,
|
|
"learning_rate": 4.340909090909091e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3062,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 344
|
|
},
|
|
{
|
|
"epoch": 0.4233128834355828,
|
|
"grad_norm": 44.35750118248226,
|
|
"learning_rate": 4.3388429752066114e-07,
|
|
"logits/chosen": -0.1875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3386,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -3.671875,
|
|
"step": 345
|
|
},
|
|
{
|
|
"epoch": 0.4245398773006135,
|
|
"grad_norm": 39.322612546267074,
|
|
"learning_rate": 4.336776859504132e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.3136,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.546875,
|
|
"step": 346
|
|
},
|
|
{
|
|
"epoch": 0.42576687116564416,
|
|
"grad_norm": 30.94392895882633,
|
|
"learning_rate": 4.334710743801653e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2331,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.296875,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.484375,
|
|
"step": 347
|
|
},
|
|
{
|
|
"epoch": 0.4269938650306748,
|
|
"grad_norm": 32.258435591671414,
|
|
"learning_rate": 4.3326446280991733e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.455078125,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2507,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0625,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 348
|
|
},
|
|
{
|
|
"epoch": 0.42822085889570555,
|
|
"grad_norm": 35.47879422672426,
|
|
"learning_rate": 4.3305785123966943e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2708,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 349
|
|
},
|
|
{
|
|
"epoch": 0.4294478527607362,
|
|
"grad_norm": 41.30786537567203,
|
|
"learning_rate": 4.328512396694214e-07,
|
|
"logits/chosen": -0.1103515625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2859,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 0.4306748466257669,
|
|
"grad_norm": 40.51120105706657,
|
|
"learning_rate": 4.326446280991735e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3046,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.875,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.796875,
|
|
"step": 351
|
|
},
|
|
{
|
|
"epoch": 0.43190184049079755,
|
|
"grad_norm": 41.727810614577876,
|
|
"learning_rate": 4.3243801652892557e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.8203125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.0625,
|
|
"step": 352
|
|
},
|
|
{
|
|
"epoch": 0.4331288343558282,
|
|
"grad_norm": 31.222074641479136,
|
|
"learning_rate": 4.3223140495867767e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2275,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 353
|
|
},
|
|
{
|
|
"epoch": 0.4343558282208589,
|
|
"grad_norm": 36.249821649908675,
|
|
"learning_rate": 4.320247933884297e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.281,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.03125,
|
|
"step": 354
|
|
},
|
|
{
|
|
"epoch": 0.43558282208588955,
|
|
"grad_norm": 40.95221006715783,
|
|
"learning_rate": 4.318181818181818e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3157,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.5625,
|
|
"step": 355
|
|
},
|
|
{
|
|
"epoch": 0.4368098159509202,
|
|
"grad_norm": 36.1989317653681,
|
|
"learning_rate": 4.3161157024793386e-07,
|
|
"logits/chosen": -0.11962890625,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3022,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.5,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.046875,
|
|
"step": 356
|
|
},
|
|
{
|
|
"epoch": 0.43803680981595094,
|
|
"grad_norm": 36.959818116679386,
|
|
"learning_rate": 4.3140495867768597e-07,
|
|
"logits/chosen": -0.365234375,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2585,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 357
|
|
},
|
|
{
|
|
"epoch": 0.4392638036809816,
|
|
"grad_norm": 35.616695396791314,
|
|
"learning_rate": 4.31198347107438e-07,
|
|
"logits/chosen": -0.318359375,
|
|
"logits/rejected": -0.4765625,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2431,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.59375,
|
|
"step": 358
|
|
},
|
|
{
|
|
"epoch": 0.4404907975460123,
|
|
"grad_norm": 39.323762039132475,
|
|
"learning_rate": 4.3099173553719006e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2505,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.296875,
|
|
"step": 359
|
|
},
|
|
{
|
|
"epoch": 0.44171779141104295,
|
|
"grad_norm": 39.07034175993009,
|
|
"learning_rate": 4.307851239669421e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2575,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 0.4429447852760736,
|
|
"grad_norm": 35.06710617877251,
|
|
"learning_rate": 4.305785123966942e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2472,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 361
|
|
},
|
|
{
|
|
"epoch": 0.4441717791411043,
|
|
"grad_norm": 34.7975606217305,
|
|
"learning_rate": 4.3037190082644625e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2548,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 362
|
|
},
|
|
{
|
|
"epoch": 0.44539877300613495,
|
|
"grad_norm": 35.42891956468369,
|
|
"learning_rate": 4.3016528925619835e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.237,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 363
|
|
},
|
|
{
|
|
"epoch": 0.4466257668711656,
|
|
"grad_norm": 44.20888419689405,
|
|
"learning_rate": 4.299586776859504e-07,
|
|
"logits/chosen": -0.1083984375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2487,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.625,
|
|
"step": 364
|
|
},
|
|
{
|
|
"epoch": 0.44785276073619634,
|
|
"grad_norm": 32.07969290277028,
|
|
"learning_rate": 4.2975206611570245e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2268,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 365
|
|
},
|
|
{
|
|
"epoch": 0.449079754601227,
|
|
"grad_norm": 35.77821392399441,
|
|
"learning_rate": 4.2954545454545455e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 366
|
|
},
|
|
{
|
|
"epoch": 0.4503067484662577,
|
|
"grad_norm": 33.47062276922143,
|
|
"learning_rate": 4.293388429752066e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2494,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 367
|
|
},
|
|
{
|
|
"epoch": 0.45153374233128835,
|
|
"grad_norm": 41.18940309867523,
|
|
"learning_rate": 4.2913223140495864e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.1494140625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3001,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.625,
|
|
"step": 368
|
|
},
|
|
{
|
|
"epoch": 0.452760736196319,
|
|
"grad_norm": 37.92759892609332,
|
|
"learning_rate": 4.289256198347107e-07,
|
|
"logits/chosen": -0.115234375,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.2816,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.75,
|
|
"step": 369
|
|
},
|
|
{
|
|
"epoch": 0.4539877300613497,
|
|
"grad_norm": 32.92636269991582,
|
|
"learning_rate": 4.287190082644628e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2538,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.703125,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 0.45521472392638035,
|
|
"grad_norm": 38.74700708912898,
|
|
"learning_rate": 4.2851239669421484e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -460.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2839,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 371
|
|
},
|
|
{
|
|
"epoch": 0.456441717791411,
|
|
"grad_norm": 35.56525332139474,
|
|
"learning_rate": 4.2830578512396694e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.239,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.53125,
|
|
"step": 372
|
|
},
|
|
{
|
|
"epoch": 0.45766871165644174,
|
|
"grad_norm": 38.919873513342004,
|
|
"learning_rate": 4.28099173553719e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.325,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 373
|
|
},
|
|
{
|
|
"epoch": 0.4588957055214724,
|
|
"grad_norm": 33.72191218536702,
|
|
"learning_rate": 4.278925619834711e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2315,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.6875,
|
|
"step": 374
|
|
},
|
|
{
|
|
"epoch": 0.4601226993865031,
|
|
"grad_norm": 39.982978401052485,
|
|
"learning_rate": 4.2768595041322313e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2831,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.3125,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 0.46134969325153374,
|
|
"grad_norm": 33.94023445384835,
|
|
"learning_rate": 4.2747933884297523e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2271,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.421875,
|
|
"step": 376
|
|
},
|
|
{
|
|
"epoch": 0.4625766871165644,
|
|
"grad_norm": 37.21157793471459,
|
|
"learning_rate": 4.272727272727272e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.262,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 377
|
|
},
|
|
{
|
|
"epoch": 0.4638036809815951,
|
|
"grad_norm": 38.20919614495083,
|
|
"learning_rate": 4.270661157024793e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.28,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 378
|
|
},
|
|
{
|
|
"epoch": 0.46503067484662575,
|
|
"grad_norm": 36.24612780889627,
|
|
"learning_rate": 4.2685950413223137e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.257,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 379
|
|
},
|
|
{
|
|
"epoch": 0.4662576687116564,
|
|
"grad_norm": 36.557393364699294,
|
|
"learning_rate": 4.2665289256198347e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2671,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 0.46748466257668714,
|
|
"grad_norm": 37.32259328400382,
|
|
"learning_rate": 4.264462809917355e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3045,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 381
|
|
},
|
|
{
|
|
"epoch": 0.4687116564417178,
|
|
"grad_norm": 34.81605521306601,
|
|
"learning_rate": 4.2623966942148757e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2736,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 382
|
|
},
|
|
{
|
|
"epoch": 0.4699386503067485,
|
|
"grad_norm": 38.419341258289826,
|
|
"learning_rate": 4.2603305785123967e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2858,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 383
|
|
},
|
|
{
|
|
"epoch": 0.47116564417177914,
|
|
"grad_norm": 33.49732006194629,
|
|
"learning_rate": 4.258264462809917e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.52734375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2126,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 384
|
|
},
|
|
{
|
|
"epoch": 0.4723926380368098,
|
|
"grad_norm": 38.114919863333334,
|
|
"learning_rate": 4.256198347107438e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2925,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 385
|
|
},
|
|
{
|
|
"epoch": 0.4736196319018405,
|
|
"grad_norm": 42.83706264363719,
|
|
"learning_rate": 4.254132231404958e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2814,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.439453125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 386
|
|
},
|
|
{
|
|
"epoch": 0.47484662576687114,
|
|
"grad_norm": 43.964728733564016,
|
|
"learning_rate": 4.252066115702479e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2666,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 387
|
|
},
|
|
{
|
|
"epoch": 0.47607361963190187,
|
|
"grad_norm": 44.03543501865791,
|
|
"learning_rate": 4.2499999999999995e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3549,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.6328125,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -3.875,
|
|
"step": 388
|
|
},
|
|
{
|
|
"epoch": 0.47730061349693254,
|
|
"grad_norm": 42.68974337451638,
|
|
"learning_rate": 4.2479338842975205e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3419,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 389
|
|
},
|
|
{
|
|
"epoch": 0.4785276073619632,
|
|
"grad_norm": 42.06628954709381,
|
|
"learning_rate": 4.245867768595041e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3237,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.361328125,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 0.47975460122699387,
|
|
"grad_norm": 49.18980712796409,
|
|
"learning_rate": 4.243801652892562e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3698,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 391
|
|
},
|
|
{
|
|
"epoch": 0.48098159509202454,
|
|
"grad_norm": 34.35189463552141,
|
|
"learning_rate": 4.2417355371900825e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2639,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.9296875,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 392
|
|
},
|
|
{
|
|
"epoch": 0.4822085889570552,
|
|
"grad_norm": 38.67130742947215,
|
|
"learning_rate": 4.2396694214876035e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.2890625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2885,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 393
|
|
},
|
|
{
|
|
"epoch": 0.4834355828220859,
|
|
"grad_norm": 40.57744523526432,
|
|
"learning_rate": 4.237603305785124e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2657,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 394
|
|
},
|
|
{
|
|
"epoch": 0.48466257668711654,
|
|
"grad_norm": 38.619317047033334,
|
|
"learning_rate": 4.2355371900826444e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.307,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 395
|
|
},
|
|
{
|
|
"epoch": 0.48588957055214727,
|
|
"grad_norm": 30.114712115596447,
|
|
"learning_rate": 4.233471074380165e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.215,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.439453125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 396
|
|
},
|
|
{
|
|
"epoch": 0.48711656441717793,
|
|
"grad_norm": 34.39736388154625,
|
|
"learning_rate": 4.231404958677686e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2349,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.84765625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 397
|
|
},
|
|
{
|
|
"epoch": 0.4883435582822086,
|
|
"grad_norm": 40.00891469568207,
|
|
"learning_rate": 4.2293388429752064e-07,
|
|
"logits/chosen": -0.2490234375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3096,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 398
|
|
},
|
|
{
|
|
"epoch": 0.48957055214723927,
|
|
"grad_norm": 35.41096130608413,
|
|
"learning_rate": 4.227272727272727e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2995,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 399
|
|
},
|
|
{
|
|
"epoch": 0.49079754601226994,
|
|
"grad_norm": 48.45396143006796,
|
|
"learning_rate": 4.225206611570248e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3464,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.4920245398773006,
|
|
"grad_norm": 30.00175559700857,
|
|
"learning_rate": 4.2231404958677683e-07,
|
|
"logits/chosen": -0.337890625,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -366.0,
|
|
"loss": 0.2176,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.46875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 401
|
|
},
|
|
{
|
|
"epoch": 0.49325153374233127,
|
|
"grad_norm": 37.529393439709146,
|
|
"learning_rate": 4.2210743801652893e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2923,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.25,
|
|
"step": 402
|
|
},
|
|
{
|
|
"epoch": 0.49447852760736194,
|
|
"grad_norm": 41.86969893893965,
|
|
"learning_rate": 4.21900826446281e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3168,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.828125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 403
|
|
},
|
|
{
|
|
"epoch": 0.49570552147239266,
|
|
"grad_norm": 34.20945313316955,
|
|
"learning_rate": 4.21694214876033e-07,
|
|
"logits/chosen": -0.1513671875,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2255,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 404
|
|
},
|
|
{
|
|
"epoch": 0.49693251533742333,
|
|
"grad_norm": 35.86293374874338,
|
|
"learning_rate": 4.2148760330578507e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.3106,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 405
|
|
},
|
|
{
|
|
"epoch": 0.498159509202454,
|
|
"grad_norm": 43.071679894395345,
|
|
"learning_rate": 4.2128099173553717e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3625,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.86328125,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 406
|
|
},
|
|
{
|
|
"epoch": 0.49938650306748467,
|
|
"grad_norm": 42.39188898560159,
|
|
"learning_rate": 4.210743801652892e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3105,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 407
|
|
},
|
|
{
|
|
"epoch": 0.5006134969325153,
|
|
"grad_norm": 45.27741438195794,
|
|
"learning_rate": 4.208677685950413e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3592,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.73828125,
|
|
"rewards/margins": 4.4375,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 408
|
|
},
|
|
{
|
|
"epoch": 0.501840490797546,
|
|
"grad_norm": 39.84136285513692,
|
|
"learning_rate": 4.2066115702479337e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2668,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.25,
|
|
"step": 409
|
|
},
|
|
{
|
|
"epoch": 0.5030674846625767,
|
|
"grad_norm": 40.12041017968796,
|
|
"learning_rate": 4.2045454545454547e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2892,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.91796875,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 0.5042944785276073,
|
|
"grad_norm": 37.21127075191681,
|
|
"learning_rate": 4.202479338842975e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2656,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.65625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 411
|
|
},
|
|
{
|
|
"epoch": 0.505521472392638,
|
|
"grad_norm": 41.819486450620104,
|
|
"learning_rate": 4.200413223140496e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.3095,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 412
|
|
},
|
|
{
|
|
"epoch": 0.5067484662576687,
|
|
"grad_norm": 38.33771429389002,
|
|
"learning_rate": 4.198347107438016e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2722,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3828125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 413
|
|
},
|
|
{
|
|
"epoch": 0.5079754601226993,
|
|
"grad_norm": 37.12482857003386,
|
|
"learning_rate": 4.196280991735537e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2323,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 414
|
|
},
|
|
{
|
|
"epoch": 0.50920245398773,
|
|
"grad_norm": 34.655790564533014,
|
|
"learning_rate": 4.1942148760330576e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2656,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.125,
|
|
"step": 415
|
|
},
|
|
{
|
|
"epoch": 0.5104294478527608,
|
|
"grad_norm": 32.98043530408525,
|
|
"learning_rate": 4.1921487603305786e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2198,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.62109375,
|
|
"rewards/margins": 4.75,
|
|
"rewards/rejected": -4.125,
|
|
"step": 416
|
|
},
|
|
{
|
|
"epoch": 0.5116564417177915,
|
|
"grad_norm": 36.96587452637558,
|
|
"learning_rate": 4.190082644628099e-07,
|
|
"logits/chosen": -0.3515625,
|
|
"logits/rejected": -0.53125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2649,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 417
|
|
},
|
|
{
|
|
"epoch": 0.5128834355828221,
|
|
"grad_norm": 34.4789260314332,
|
|
"learning_rate": 4.1880165289256195e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2518,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.25,
|
|
"step": 418
|
|
},
|
|
{
|
|
"epoch": 0.5141104294478528,
|
|
"grad_norm": 34.78381462111674,
|
|
"learning_rate": 4.1859504132231405e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2278,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 419
|
|
},
|
|
{
|
|
"epoch": 0.5153374233128835,
|
|
"grad_norm": 39.079403507615055,
|
|
"learning_rate": 4.183884297520661e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.3113,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 0.5165644171779141,
|
|
"grad_norm": 34.54058833309064,
|
|
"learning_rate": 4.1818181818181814e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3077,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -3.875,
|
|
"step": 421
|
|
},
|
|
{
|
|
"epoch": 0.5177914110429448,
|
|
"grad_norm": 36.04776328314745,
|
|
"learning_rate": 4.179752066115702e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2784,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 422
|
|
},
|
|
{
|
|
"epoch": 0.5190184049079755,
|
|
"grad_norm": 43.01937925233782,
|
|
"learning_rate": 4.177685950413223e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3382,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 423
|
|
},
|
|
{
|
|
"epoch": 0.5202453987730061,
|
|
"grad_norm": 36.48453900377271,
|
|
"learning_rate": 4.1756198347107434e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3126,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.375,
|
|
"step": 424
|
|
},
|
|
{
|
|
"epoch": 0.5214723926380368,
|
|
"grad_norm": 41.79473108392073,
|
|
"learning_rate": 4.1735537190082644e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.318359375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3372,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.5078125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 0.5226993865030675,
|
|
"grad_norm": 48.26366364465108,
|
|
"learning_rate": 4.171487603305785e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2573,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 426
|
|
},
|
|
{
|
|
"epoch": 0.5239263803680981,
|
|
"grad_norm": 99.7573372246345,
|
|
"learning_rate": 4.169421487603306e-07,
|
|
"logits/chosen": -0.1982421875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2668,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 427
|
|
},
|
|
{
|
|
"epoch": 0.5251533742331288,
|
|
"grad_norm": 37.06815581413299,
|
|
"learning_rate": 4.1673553719008263e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3028,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.125,
|
|
"step": 428
|
|
},
|
|
{
|
|
"epoch": 0.5263803680981595,
|
|
"grad_norm": 33.97292039326683,
|
|
"learning_rate": 4.1652892561983473e-07,
|
|
"logits/chosen": -0.224609375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.243,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 429
|
|
},
|
|
{
|
|
"epoch": 0.5276073619631901,
|
|
"grad_norm": 29.82247469458647,
|
|
"learning_rate": 4.163223140495867e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.1971,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 0.5288343558282208,
|
|
"grad_norm": 50.10617111263047,
|
|
"learning_rate": 4.161157024793388e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3808,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.7109375,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 431
|
|
},
|
|
{
|
|
"epoch": 0.5300613496932516,
|
|
"grad_norm": 46.34719786232677,
|
|
"learning_rate": 4.1590909090909087e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.185546875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3379,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 432
|
|
},
|
|
{
|
|
"epoch": 0.5312883435582823,
|
|
"grad_norm": 42.692948753426734,
|
|
"learning_rate": 4.1570247933884297e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3119,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.55078125,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 433
|
|
},
|
|
{
|
|
"epoch": 0.5325153374233129,
|
|
"grad_norm": 39.50920458351492,
|
|
"learning_rate": 4.15495867768595e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2357,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.61328125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 434
|
|
},
|
|
{
|
|
"epoch": 0.5337423312883436,
|
|
"grad_norm": 41.38331544820595,
|
|
"learning_rate": 4.1528925619834707e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2711,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 435
|
|
},
|
|
{
|
|
"epoch": 0.5349693251533743,
|
|
"grad_norm": 39.363819944524835,
|
|
"learning_rate": 4.1508264462809917e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2469,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6796875,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 436
|
|
},
|
|
{
|
|
"epoch": 0.5361963190184049,
|
|
"grad_norm": 32.4215285898728,
|
|
"learning_rate": 4.148760330578512e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2367,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 437
|
|
},
|
|
{
|
|
"epoch": 0.5374233128834356,
|
|
"grad_norm": 41.11999044241744,
|
|
"learning_rate": 4.146694214876033e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3424,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 438
|
|
},
|
|
{
|
|
"epoch": 0.5386503067484663,
|
|
"grad_norm": 36.95463031219475,
|
|
"learning_rate": 4.144628099173553e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2791,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.7734375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 439
|
|
},
|
|
{
|
|
"epoch": 0.5398773006134969,
|
|
"grad_norm": 49.69926246487377,
|
|
"learning_rate": 4.142561983471074e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3218,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.35546875,
|
|
"rewards/margins": 4.40625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 0.5411042944785276,
|
|
"grad_norm": 39.707681073149594,
|
|
"learning_rate": 4.1404958677685946e-07,
|
|
"logits/chosen": -0.107421875,
|
|
"logits/rejected": -0.224609375,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2837,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 4.78125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 441
|
|
},
|
|
{
|
|
"epoch": 0.5423312883435583,
|
|
"grad_norm": 38.960801334852974,
|
|
"learning_rate": 4.1384297520661156e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3027,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.0234375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 442
|
|
},
|
|
{
|
|
"epoch": 0.5435582822085889,
|
|
"grad_norm": 42.815669493727604,
|
|
"learning_rate": 4.136363636363636e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.303,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.0,
|
|
"step": 443
|
|
},
|
|
{
|
|
"epoch": 0.5447852760736196,
|
|
"grad_norm": 35.80955378025667,
|
|
"learning_rate": 4.134297520661157e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2204,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 444
|
|
},
|
|
{
|
|
"epoch": 0.5460122699386503,
|
|
"grad_norm": 38.126966452680875,
|
|
"learning_rate": 4.1322314049586775e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3507,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 445
|
|
},
|
|
{
|
|
"epoch": 0.5472392638036809,
|
|
"grad_norm": 35.409001850251684,
|
|
"learning_rate": 4.1301652892561985e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.2334,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -3.875,
|
|
"step": 446
|
|
},
|
|
{
|
|
"epoch": 0.5484662576687117,
|
|
"grad_norm": 37.030505301448606,
|
|
"learning_rate": 4.128099173553719e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2524,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 447
|
|
},
|
|
{
|
|
"epoch": 0.5496932515337424,
|
|
"grad_norm": 40.14683673248258,
|
|
"learning_rate": 4.1260330578512394e-07,
|
|
"logits/chosen": -0.1435546875,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3068,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -4.0,
|
|
"step": 448
|
|
},
|
|
{
|
|
"epoch": 0.550920245398773,
|
|
"grad_norm": 36.49625976131608,
|
|
"learning_rate": 4.12396694214876e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.257,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 449
|
|
},
|
|
{
|
|
"epoch": 0.5521472392638037,
|
|
"grad_norm": 36.243221597677895,
|
|
"learning_rate": 4.121900826446281e-07,
|
|
"logits/chosen": -0.2890625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2672,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 0.5533742331288344,
|
|
"grad_norm": 32.1407046295721,
|
|
"learning_rate": 4.1198347107438014e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.353515625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.2477,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 451
|
|
},
|
|
{
|
|
"epoch": 0.554601226993865,
|
|
"grad_norm": 42.28873095520698,
|
|
"learning_rate": 4.1177685950413224e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -460.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2747,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9140625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.0,
|
|
"step": 452
|
|
},
|
|
{
|
|
"epoch": 0.5558282208588957,
|
|
"grad_norm": 32.36959074636312,
|
|
"learning_rate": 4.115702479338843e-07,
|
|
"logits/chosen": -0.29296875,
|
|
"logits/rejected": -0.5234375,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.1965,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 6.5,
|
|
"rewards/rejected": -4.375,
|
|
"step": 453
|
|
},
|
|
{
|
|
"epoch": 0.5570552147239264,
|
|
"grad_norm": 37.72144307940475,
|
|
"learning_rate": 4.1136363636363633e-07,
|
|
"logits/chosen": -0.1572265625,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -340.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2824,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 454
|
|
},
|
|
{
|
|
"epoch": 0.558282208588957,
|
|
"grad_norm": 36.8405329111992,
|
|
"learning_rate": 4.1115702479338843e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2387,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 455
|
|
},
|
|
{
|
|
"epoch": 0.5595092024539877,
|
|
"grad_norm": 38.28626472369992,
|
|
"learning_rate": 4.109504132231405e-07,
|
|
"logits/chosen": -0.1533203125,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2614,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.9921875,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 456
|
|
},
|
|
{
|
|
"epoch": 0.5607361963190184,
|
|
"grad_norm": 37.602299554825706,
|
|
"learning_rate": 4.1074380165289253e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.326171875,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2718,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -3.84375,
|
|
"step": 457
|
|
},
|
|
{
|
|
"epoch": 0.561963190184049,
|
|
"grad_norm": 34.248759383891915,
|
|
"learning_rate": 4.105371900826446e-07,
|
|
"logits/chosen": -0.203125,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2095,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.671875,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 458
|
|
},
|
|
{
|
|
"epoch": 0.5631901840490797,
|
|
"grad_norm": 36.02884511565117,
|
|
"learning_rate": 4.103305785123967e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2657,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 459
|
|
},
|
|
{
|
|
"epoch": 0.5644171779141104,
|
|
"grad_norm": 39.138490077598576,
|
|
"learning_rate": 4.101239669421487e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2605,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.8125,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 0.5656441717791411,
|
|
"grad_norm": 42.1658732163214,
|
|
"learning_rate": 4.099173553719008e-07,
|
|
"logits/chosen": -0.37890625,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2725,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.9765625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 461
|
|
},
|
|
{
|
|
"epoch": 0.5668711656441717,
|
|
"grad_norm": 41.91165523140955,
|
|
"learning_rate": 4.0971074380165287e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.3075,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.8203125,
|
|
"rewards/margins": 4.625,
|
|
"rewards/rejected": -3.8125,
|
|
"step": 462
|
|
},
|
|
{
|
|
"epoch": 0.5680981595092025,
|
|
"grad_norm": 38.69169820185617,
|
|
"learning_rate": 4.0950413223140497e-07,
|
|
"logits/chosen": -0.248046875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2766,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 463
|
|
},
|
|
{
|
|
"epoch": 0.5693251533742332,
|
|
"grad_norm": 43.91810675336078,
|
|
"learning_rate": 4.09297520661157e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2966,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 464
|
|
},
|
|
{
|
|
"epoch": 0.5705521472392638,
|
|
"grad_norm": 33.81460512528705,
|
|
"learning_rate": 4.090909090909091e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.1974,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 465
|
|
},
|
|
{
|
|
"epoch": 0.5717791411042945,
|
|
"grad_norm": 43.114669278637464,
|
|
"learning_rate": 4.088842975206611e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2803,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.75,
|
|
"step": 466
|
|
},
|
|
{
|
|
"epoch": 0.5730061349693252,
|
|
"grad_norm": 41.26777561271183,
|
|
"learning_rate": 4.086776859504132e-07,
|
|
"logits/chosen": -0.201171875,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2918,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 467
|
|
},
|
|
{
|
|
"epoch": 0.5742331288343558,
|
|
"grad_norm": 46.80452187024971,
|
|
"learning_rate": 4.0847107438016526e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2962,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 468
|
|
},
|
|
{
|
|
"epoch": 0.5754601226993865,
|
|
"grad_norm": 35.72783758477153,
|
|
"learning_rate": 4.0826446280991736e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.194,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 469
|
|
},
|
|
{
|
|
"epoch": 0.5766871165644172,
|
|
"grad_norm": 36.63750325467396,
|
|
"learning_rate": 4.080578512396694e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2875,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.625,
|
|
"step": 470
|
|
},
|
|
{
|
|
"epoch": 0.5779141104294478,
|
|
"grad_norm": 41.264612609731145,
|
|
"learning_rate": 4.0785123966942145e-07,
|
|
"logits/chosen": -0.07373046875,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2914,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 471
|
|
},
|
|
{
|
|
"epoch": 0.5791411042944785,
|
|
"grad_norm": 45.183894833304514,
|
|
"learning_rate": 4.0764462809917355e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2808,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 472
|
|
},
|
|
{
|
|
"epoch": 0.5803680981595092,
|
|
"grad_norm": 43.25551760993764,
|
|
"learning_rate": 4.074380165289256e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.4140625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.2992,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 473
|
|
},
|
|
{
|
|
"epoch": 0.5815950920245399,
|
|
"grad_norm": 36.874064189408784,
|
|
"learning_rate": 4.072314049586777e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.4609375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2515,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.984375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 474
|
|
},
|
|
{
|
|
"epoch": 0.5828220858895705,
|
|
"grad_norm": 33.97530110519338,
|
|
"learning_rate": 4.070247933884297e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2189,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.8828125,
|
|
"rewards/margins": 6.6875,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 475
|
|
},
|
|
{
|
|
"epoch": 0.5840490797546012,
|
|
"grad_norm": 32.31826742696185,
|
|
"learning_rate": 4.068181818181818e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2688,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.462890625,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 476
|
|
},
|
|
{
|
|
"epoch": 0.5852760736196319,
|
|
"grad_norm": 36.47831028156178,
|
|
"learning_rate": 4.0661157024793384e-07,
|
|
"logits/chosen": -0.287109375,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2274,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 477
|
|
},
|
|
{
|
|
"epoch": 0.5865030674846625,
|
|
"grad_norm": 38.16498737596341,
|
|
"learning_rate": 4.0640495867768594e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2857,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 478
|
|
},
|
|
{
|
|
"epoch": 0.5877300613496933,
|
|
"grad_norm": 41.81318374412193,
|
|
"learning_rate": 4.06198347107438e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.462890625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2644,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 479
|
|
},
|
|
{
|
|
"epoch": 0.588957055214724,
|
|
"grad_norm": 34.21678743382818,
|
|
"learning_rate": 4.059917355371901e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2258,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.68359375,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"epoch": 0.5901840490797546,
|
|
"grad_norm": 41.214709064982436,
|
|
"learning_rate": 4.0578512396694213e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3414,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.94921875,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 481
|
|
},
|
|
{
|
|
"epoch": 0.5914110429447853,
|
|
"grad_norm": 34.07376051032277,
|
|
"learning_rate": 4.0557851239669423e-07,
|
|
"logits/chosen": -0.2431640625,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.269,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 482
|
|
},
|
|
{
|
|
"epoch": 0.592638036809816,
|
|
"grad_norm": 37.03436000175101,
|
|
"learning_rate": 4.053719008264463e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -304.0,
|
|
"loss": 0.2869,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 483
|
|
},
|
|
{
|
|
"epoch": 0.5938650306748466,
|
|
"grad_norm": 37.50453607778557,
|
|
"learning_rate": 4.0516528925619833e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3017,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 484
|
|
},
|
|
{
|
|
"epoch": 0.5950920245398773,
|
|
"grad_norm": 35.082772186969954,
|
|
"learning_rate": 4.049586776859504e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2666,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.6796875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 485
|
|
},
|
|
{
|
|
"epoch": 0.596319018404908,
|
|
"grad_norm": 34.67777878301507,
|
|
"learning_rate": 4.047520661157025e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.33203125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2405,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 486
|
|
},
|
|
{
|
|
"epoch": 0.5975460122699386,
|
|
"grad_norm": 36.99824119978169,
|
|
"learning_rate": 4.045454545454545e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2534,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.96875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.0,
|
|
"step": 487
|
|
},
|
|
{
|
|
"epoch": 0.5987730061349693,
|
|
"grad_norm": 47.34405656815455,
|
|
"learning_rate": 4.0433884297520657e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2559,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 488
|
|
},
|
|
{
|
|
"epoch": 0.6,
|
|
"grad_norm": 32.18985964645197,
|
|
"learning_rate": 4.0413223140495867e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2603,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4609375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 489
|
|
},
|
|
{
|
|
"epoch": 0.6012269938650306,
|
|
"grad_norm": 41.3641809975054,
|
|
"learning_rate": 4.039256198347107e-07,
|
|
"logits/chosen": -0.14453125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2435,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.9609375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 490
|
|
},
|
|
{
|
|
"epoch": 0.6024539877300613,
|
|
"grad_norm": 36.929475951535615,
|
|
"learning_rate": 4.037190082644628e-07,
|
|
"logits/chosen": -0.1279296875,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2468,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.76953125,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 491
|
|
},
|
|
{
|
|
"epoch": 0.603680981595092,
|
|
"grad_norm": 34.25022073617326,
|
|
"learning_rate": 4.0351239669421486e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.2476,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.625,
|
|
"step": 492
|
|
},
|
|
{
|
|
"epoch": 0.6049079754601226,
|
|
"grad_norm": 44.528399584764976,
|
|
"learning_rate": 4.033057851239669e-07,
|
|
"logits/chosen": -0.07275390625,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3692,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.67578125,
|
|
"rewards/margins": 4.5625,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 493
|
|
},
|
|
{
|
|
"epoch": 0.6061349693251534,
|
|
"grad_norm": 33.0051523964255,
|
|
"learning_rate": 4.0309917355371896e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2292,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 494
|
|
},
|
|
{
|
|
"epoch": 0.6073619631901841,
|
|
"grad_norm": 36.62725418587262,
|
|
"learning_rate": 4.0289256198347106e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2572,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 495
|
|
},
|
|
{
|
|
"epoch": 0.6085889570552148,
|
|
"grad_norm": 42.16049484237724,
|
|
"learning_rate": 4.026859504132231e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.431640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3108,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.53515625,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 496
|
|
},
|
|
{
|
|
"epoch": 0.6098159509202454,
|
|
"grad_norm": 45.47382251482463,
|
|
"learning_rate": 4.024793388429752e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.3664,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.51953125,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 497
|
|
},
|
|
{
|
|
"epoch": 0.6110429447852761,
|
|
"grad_norm": 35.6433892815013,
|
|
"learning_rate": 4.0227272727272725e-07,
|
|
"logits/chosen": -0.28515625,
|
|
"logits/rejected": -0.46484375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2458,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.171875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 498
|
|
},
|
|
{
|
|
"epoch": 0.6122699386503068,
|
|
"grad_norm": 37.060767498257135,
|
|
"learning_rate": 4.0206611570247935e-07,
|
|
"logits/chosen": -0.1630859375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2685,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 499
|
|
},
|
|
{
|
|
"epoch": 0.6134969325153374,
|
|
"grad_norm": 35.21143709339605,
|
|
"learning_rate": 4.018595041322314e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2498,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.75,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.6147239263803681,
|
|
"grad_norm": 39.07038484978643,
|
|
"learning_rate": 4.0165289256198345e-07,
|
|
"logits/chosen": -0.240234375,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2402,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 501
|
|
},
|
|
{
|
|
"epoch": 0.6159509202453988,
|
|
"grad_norm": 34.95433271630032,
|
|
"learning_rate": 4.014462809917355e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2631,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 502
|
|
},
|
|
{
|
|
"epoch": 0.6171779141104294,
|
|
"grad_norm": 36.27046530070022,
|
|
"learning_rate": 4.012396694214876e-07,
|
|
"logits/chosen": -0.0498046875,
|
|
"logits/rejected": -0.248046875,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2571,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.8515625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 503
|
|
},
|
|
{
|
|
"epoch": 0.6184049079754601,
|
|
"grad_norm": 32.074615531320944,
|
|
"learning_rate": 4.0103305785123964e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.44140625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.253,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.416015625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 504
|
|
},
|
|
{
|
|
"epoch": 0.6196319018404908,
|
|
"grad_norm": 50.2956600108261,
|
|
"learning_rate": 4.0082644628099174e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.4921875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -384.0,
|
|
"loss": 0.3428,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.40234375,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 505
|
|
},
|
|
{
|
|
"epoch": 0.6208588957055214,
|
|
"grad_norm": 39.722486070303646,
|
|
"learning_rate": 4.006198347107438e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2585,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.93359375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 506
|
|
},
|
|
{
|
|
"epoch": 0.6220858895705521,
|
|
"grad_norm": 31.372949248189112,
|
|
"learning_rate": 4.0041322314049583e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2085,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 507
|
|
},
|
|
{
|
|
"epoch": 0.6233128834355828,
|
|
"grad_norm": 31.130400720786195,
|
|
"learning_rate": 4.0020661157024793e-07,
|
|
"logits/chosen": -0.062255859375,
|
|
"logits/rejected": -0.2431640625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2241,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 508
|
|
},
|
|
{
|
|
"epoch": 0.6245398773006134,
|
|
"grad_norm": 53.16460512064989,
|
|
"learning_rate": 4e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -464.0,
|
|
"logps/rejected": -374.0,
|
|
"loss": 0.3519,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 2.109375,
|
|
"rewards/margins": 7.0625,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 509
|
|
},
|
|
{
|
|
"epoch": 0.6257668711656442,
|
|
"grad_norm": 33.530586116590285,
|
|
"learning_rate": 3.9979338842975203e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2436,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 510
|
|
},
|
|
{
|
|
"epoch": 0.6269938650306749,
|
|
"grad_norm": 40.068124237732846,
|
|
"learning_rate": 3.995867768595041e-07,
|
|
"logits/chosen": -0.30078125,
|
|
"logits/rejected": -0.439453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2864,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 511
|
|
},
|
|
{
|
|
"epoch": 0.6282208588957056,
|
|
"grad_norm": 46.79446957798848,
|
|
"learning_rate": 3.993801652892562e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.3288,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 512
|
|
},
|
|
{
|
|
"epoch": 0.6294478527607362,
|
|
"grad_norm": 39.156027495649305,
|
|
"learning_rate": 3.991735537190082e-07,
|
|
"logits/chosen": -0.138671875,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.335,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.71875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 513
|
|
},
|
|
{
|
|
"epoch": 0.6306748466257669,
|
|
"grad_norm": 32.5774175879679,
|
|
"learning_rate": 3.989669421487603e-07,
|
|
"logits/chosen": -0.12353515625,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -278.0,
|
|
"loss": 0.2532,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.220703125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 514
|
|
},
|
|
{
|
|
"epoch": 0.6319018404907976,
|
|
"grad_norm": 40.41320321797794,
|
|
"learning_rate": 3.9876033057851237e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2856,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.2578125,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -5.0,
|
|
"step": 515
|
|
},
|
|
{
|
|
"epoch": 0.6331288343558282,
|
|
"grad_norm": 33.09690635668108,
|
|
"learning_rate": 3.9855371900826447e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2475,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.404296875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 516
|
|
},
|
|
{
|
|
"epoch": 0.6343558282208589,
|
|
"grad_norm": 45.00347742756972,
|
|
"learning_rate": 3.983471074380165e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3686,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 517
|
|
},
|
|
{
|
|
"epoch": 0.6355828220858896,
|
|
"grad_norm": 37.888666954836985,
|
|
"learning_rate": 3.981404958677686e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2069,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.328125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.625,
|
|
"step": 518
|
|
},
|
|
{
|
|
"epoch": 0.6368098159509202,
|
|
"grad_norm": 38.80463867569447,
|
|
"learning_rate": 3.979338842975206e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2446,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.921875,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.875,
|
|
"step": 519
|
|
},
|
|
{
|
|
"epoch": 0.6380368098159509,
|
|
"grad_norm": 40.986581892777515,
|
|
"learning_rate": 3.977272727272727e-07,
|
|
"logits/chosen": -0.259765625,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3191,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 520
|
|
},
|
|
{
|
|
"epoch": 0.6392638036809816,
|
|
"grad_norm": 36.04171128345913,
|
|
"learning_rate": 3.9752066115702476e-07,
|
|
"logits/chosen": -0.07763671875,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2642,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 521
|
|
},
|
|
{
|
|
"epoch": 0.6404907975460122,
|
|
"grad_norm": 36.91350515266316,
|
|
"learning_rate": 3.9731404958677686e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3094,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 522
|
|
},
|
|
{
|
|
"epoch": 0.6417177914110429,
|
|
"grad_norm": 42.88283547574401,
|
|
"learning_rate": 3.971074380165289e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3333,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.6171875,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 523
|
|
},
|
|
{
|
|
"epoch": 0.6429447852760736,
|
|
"grad_norm": 45.322523177242,
|
|
"learning_rate": 3.9690082644628095e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.3031,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 524
|
|
},
|
|
{
|
|
"epoch": 0.6441717791411042,
|
|
"grad_norm": 49.70289783132654,
|
|
"learning_rate": 3.9669421487603305e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.3567,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.578125,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 525
|
|
},
|
|
{
|
|
"epoch": 0.645398773006135,
|
|
"grad_norm": 36.78598991927111,
|
|
"learning_rate": 3.964876033057851e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2502,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.98828125,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 526
|
|
},
|
|
{
|
|
"epoch": 0.6466257668711657,
|
|
"grad_norm": 36.651686257226466,
|
|
"learning_rate": 3.962809917355372e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2789,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.85546875,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 527
|
|
},
|
|
{
|
|
"epoch": 0.6478527607361964,
|
|
"grad_norm": 35.31748764596186,
|
|
"learning_rate": 3.960743801652892e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2999,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 528
|
|
},
|
|
{
|
|
"epoch": 0.649079754601227,
|
|
"grad_norm": 36.86996860776575,
|
|
"learning_rate": 3.958677685950413e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3236,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 529
|
|
},
|
|
{
|
|
"epoch": 0.6503067484662577,
|
|
"grad_norm": 40.308596718463534,
|
|
"learning_rate": 3.9566115702479334e-07,
|
|
"logits/chosen": -0.1611328125,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.307,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 530
|
|
},
|
|
{
|
|
"epoch": 0.6515337423312884,
|
|
"grad_norm": 34.92823583852622,
|
|
"learning_rate": 3.9545454545454544e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2686,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.5,
|
|
"step": 531
|
|
},
|
|
{
|
|
"epoch": 0.652760736196319,
|
|
"grad_norm": 35.9537343386399,
|
|
"learning_rate": 3.952479338842975e-07,
|
|
"logits/chosen": -0.12158203125,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2733,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 532
|
|
},
|
|
{
|
|
"epoch": 0.6539877300613497,
|
|
"grad_norm": 33.35300325251816,
|
|
"learning_rate": 3.950413223140496e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.482421875,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2339,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 533
|
|
},
|
|
{
|
|
"epoch": 0.6552147239263804,
|
|
"grad_norm": 33.59071906368527,
|
|
"learning_rate": 3.9483471074380164e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2687,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 534
|
|
},
|
|
{
|
|
"epoch": 0.656441717791411,
|
|
"grad_norm": 37.966223525168424,
|
|
"learning_rate": 3.9462809917355374e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2745,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.359375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 535
|
|
},
|
|
{
|
|
"epoch": 0.6576687116564417,
|
|
"grad_norm": 33.781725934858414,
|
|
"learning_rate": 3.944214876033058e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2154,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 536
|
|
},
|
|
{
|
|
"epoch": 0.6588957055214724,
|
|
"grad_norm": 47.92234178996883,
|
|
"learning_rate": 3.9421487603305783e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.3881,
|
|
"rewards/accuracies": 0.7734375,
|
|
"rewards/chosen": 0.1796875,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 537
|
|
},
|
|
{
|
|
"epoch": 0.660122699386503,
|
|
"grad_norm": 42.47717470031941,
|
|
"learning_rate": 3.940082644628099e-07,
|
|
"logits/chosen": -0.150390625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3398,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 538
|
|
},
|
|
{
|
|
"epoch": 0.6613496932515337,
|
|
"grad_norm": 37.22523933086588,
|
|
"learning_rate": 3.93801652892562e-07,
|
|
"logits/chosen": -0.154296875,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2742,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.859375,
|
|
"step": 539
|
|
},
|
|
{
|
|
"epoch": 0.6625766871165644,
|
|
"grad_norm": 40.71616745650965,
|
|
"learning_rate": 3.93595041322314e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3006,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5,
|
|
"step": 540
|
|
},
|
|
{
|
|
"epoch": 0.6638036809815951,
|
|
"grad_norm": 36.00028869643927,
|
|
"learning_rate": 3.933884297520661e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2699,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.71875,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 541
|
|
},
|
|
{
|
|
"epoch": 0.6650306748466258,
|
|
"grad_norm": 36.856579052467744,
|
|
"learning_rate": 3.9318181818181817e-07,
|
|
"logits/chosen": -0.1845703125,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2808,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.265625,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.0,
|
|
"step": 542
|
|
},
|
|
{
|
|
"epoch": 0.6662576687116565,
|
|
"grad_norm": 37.076885533693634,
|
|
"learning_rate": 3.929752066115702e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -446.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2335,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.9296875,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 543
|
|
},
|
|
{
|
|
"epoch": 0.6674846625766871,
|
|
"grad_norm": 48.36100783202625,
|
|
"learning_rate": 3.927685950413223e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.3321,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 544
|
|
},
|
|
{
|
|
"epoch": 0.6687116564417178,
|
|
"grad_norm": 43.36042581975704,
|
|
"learning_rate": 3.9256198347107437e-07,
|
|
"logits/chosen": -0.25,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3606,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 545
|
|
},
|
|
{
|
|
"epoch": 0.6699386503067485,
|
|
"grad_norm": 39.63701590605686,
|
|
"learning_rate": 3.923553719008264e-07,
|
|
"logits/chosen": -0.1416015625,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3095,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 546
|
|
},
|
|
{
|
|
"epoch": 0.6711656441717792,
|
|
"grad_norm": 33.08408154520663,
|
|
"learning_rate": 3.9214876033057846e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2677,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.53125,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 547
|
|
},
|
|
{
|
|
"epoch": 0.6723926380368098,
|
|
"grad_norm": 41.01241845328248,
|
|
"learning_rate": 3.9194214876033056e-07,
|
|
"logits/chosen": -0.251953125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2427,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6640625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 548
|
|
},
|
|
{
|
|
"epoch": 0.6736196319018405,
|
|
"grad_norm": 38.04742407115687,
|
|
"learning_rate": 3.917355371900826e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2684,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3203125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 549
|
|
},
|
|
{
|
|
"epoch": 0.6748466257668712,
|
|
"grad_norm": 38.74485204677174,
|
|
"learning_rate": 3.915289256198347e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3275,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.78125,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 0.6760736196319018,
|
|
"grad_norm": 36.24315774996275,
|
|
"learning_rate": 3.9132231404958675e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2537,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.9375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 551
|
|
},
|
|
{
|
|
"epoch": 0.6773006134969325,
|
|
"grad_norm": 36.45208223894636,
|
|
"learning_rate": 3.9111570247933885e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2933,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -3.921875,
|
|
"step": 552
|
|
},
|
|
{
|
|
"epoch": 0.6785276073619632,
|
|
"grad_norm": 41.631610478201644,
|
|
"learning_rate": 3.909090909090909e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.3207,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.796875,
|
|
"step": 553
|
|
},
|
|
{
|
|
"epoch": 0.6797546012269938,
|
|
"grad_norm": 33.49412012982899,
|
|
"learning_rate": 3.90702479338843e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2289,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.84375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.796875,
|
|
"step": 554
|
|
},
|
|
{
|
|
"epoch": 0.6809815950920245,
|
|
"grad_norm": 37.66391393717185,
|
|
"learning_rate": 3.90495867768595e-07,
|
|
"logits/chosen": -0.03759765625,
|
|
"logits/rejected": -0.1650390625,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3404,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -3.71875,
|
|
"step": 555
|
|
},
|
|
{
|
|
"epoch": 0.6822085889570552,
|
|
"grad_norm": 30.301075310625425,
|
|
"learning_rate": 3.902892561983471e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2278,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 2.140625,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.0625,
|
|
"step": 556
|
|
},
|
|
{
|
|
"epoch": 0.6834355828220859,
|
|
"grad_norm": 33.83430112648751,
|
|
"learning_rate": 3.9008264462809914e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2682,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.65625,
|
|
"step": 557
|
|
},
|
|
{
|
|
"epoch": 0.6846625766871166,
|
|
"grad_norm": 30.75402635256423,
|
|
"learning_rate": 3.8987603305785124e-07,
|
|
"logits/chosen": -0.1298828125,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2312,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 558
|
|
},
|
|
{
|
|
"epoch": 0.6858895705521473,
|
|
"grad_norm": 40.61066869149253,
|
|
"learning_rate": 3.896694214876033e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3017,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 559
|
|
},
|
|
{
|
|
"epoch": 0.6871165644171779,
|
|
"grad_norm": 28.075176930111862,
|
|
"learning_rate": 3.8946280991735534e-07,
|
|
"logits/chosen": -0.34375,
|
|
"logits/rejected": -0.494140625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.1832,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6171875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 560
|
|
},
|
|
{
|
|
"epoch": 0.6883435582822086,
|
|
"grad_norm": 31.670561048270176,
|
|
"learning_rate": 3.8925619834710744e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.390625,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.2243,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 561
|
|
},
|
|
{
|
|
"epoch": 0.6895705521472393,
|
|
"grad_norm": 32.855068201542075,
|
|
"learning_rate": 3.890495867768595e-07,
|
|
"logits/chosen": -0.087890625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2422,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.84375,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 562
|
|
},
|
|
{
|
|
"epoch": 0.69079754601227,
|
|
"grad_norm": 35.79449426535959,
|
|
"learning_rate": 3.888429752066116e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.40625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2546,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7109375,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 563
|
|
},
|
|
{
|
|
"epoch": 0.6920245398773006,
|
|
"grad_norm": 29.77603709052077,
|
|
"learning_rate": 3.886363636363636e-07,
|
|
"logits/chosen": -0.255859375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2408,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 564
|
|
},
|
|
{
|
|
"epoch": 0.6932515337423313,
|
|
"grad_norm": 35.70334442848756,
|
|
"learning_rate": 3.884297520661157e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.400390625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3131,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 4.84375,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 565
|
|
},
|
|
{
|
|
"epoch": 0.694478527607362,
|
|
"grad_norm": 35.48298237028605,
|
|
"learning_rate": 3.882231404958677e-07,
|
|
"logits/chosen": -0.2109375,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2452,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 566
|
|
},
|
|
{
|
|
"epoch": 0.6957055214723926,
|
|
"grad_norm": 31.112271919353073,
|
|
"learning_rate": 3.880165289256198e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.251953125,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -266.0,
|
|
"loss": 0.2267,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 4.8125,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 567
|
|
},
|
|
{
|
|
"epoch": 0.6969325153374233,
|
|
"grad_norm": 38.89246029311522,
|
|
"learning_rate": 3.8780991735537187e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2865,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 568
|
|
},
|
|
{
|
|
"epoch": 0.698159509202454,
|
|
"grad_norm": 34.50521407178372,
|
|
"learning_rate": 3.8760330578512397e-07,
|
|
"logits/chosen": -0.1552734375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2647,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.7578125,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 569
|
|
},
|
|
{
|
|
"epoch": 0.6993865030674846,
|
|
"grad_norm": 27.986690455754786,
|
|
"learning_rate": 3.87396694214876e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.451171875,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.18,
|
|
"rewards/accuracies": 0.9453125,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 570
|
|
},
|
|
{
|
|
"epoch": 0.7006134969325153,
|
|
"grad_norm": 32.36101057946353,
|
|
"learning_rate": 3.871900826446281e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -326.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2221,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 5.0625,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 571
|
|
},
|
|
{
|
|
"epoch": 0.701840490797546,
|
|
"grad_norm": 36.69242569066846,
|
|
"learning_rate": 3.8698347107438017e-07,
|
|
"logits/chosen": -0.166015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2363,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 572
|
|
},
|
|
{
|
|
"epoch": 0.7030674846625767,
|
|
"grad_norm": 41.66310475024957,
|
|
"learning_rate": 3.867768595041322e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3484,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.51171875,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 573
|
|
},
|
|
{
|
|
"epoch": 0.7042944785276074,
|
|
"grad_norm": 39.39212100565405,
|
|
"learning_rate": 3.8657024793388426e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.376953125,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2719,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.625,
|
|
"step": 574
|
|
},
|
|
{
|
|
"epoch": 0.7055214723926381,
|
|
"grad_norm": 35.35137643401559,
|
|
"learning_rate": 3.8636363636363636e-07,
|
|
"logits/chosen": -0.283203125,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2178,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 6.65625,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 575
|
|
},
|
|
{
|
|
"epoch": 0.7067484662576687,
|
|
"grad_norm": 28.9202010136064,
|
|
"learning_rate": 3.861570247933884e-07,
|
|
"logits/chosen": -0.1494140625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.1958,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.59375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 576
|
|
},
|
|
{
|
|
"epoch": 0.7079754601226994,
|
|
"grad_norm": 36.46431827427722,
|
|
"learning_rate": 3.8595041322314045e-07,
|
|
"logits/chosen": -0.09033203125,
|
|
"logits/rejected": -0.23046875,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2593,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 577
|
|
},
|
|
{
|
|
"epoch": 0.7092024539877301,
|
|
"grad_norm": 40.66901563749969,
|
|
"learning_rate": 3.8574380165289255e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3148,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.0,
|
|
"step": 578
|
|
},
|
|
{
|
|
"epoch": 0.7104294478527607,
|
|
"grad_norm": 43.94460926088441,
|
|
"learning_rate": 3.855371900826446e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3599,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.498046875,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -4.375,
|
|
"step": 579
|
|
},
|
|
{
|
|
"epoch": 0.7116564417177914,
|
|
"grad_norm": 36.521177273018374,
|
|
"learning_rate": 3.853305785123967e-07,
|
|
"logits/chosen": -0.125,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2355,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.40625,
|
|
"step": 580
|
|
},
|
|
{
|
|
"epoch": 0.7128834355828221,
|
|
"grad_norm": 30.81358847829242,
|
|
"learning_rate": 3.8512396694214875e-07,
|
|
"logits/chosen": -0.2373046875,
|
|
"logits/rejected": -0.427734375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2127,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 581
|
|
},
|
|
{
|
|
"epoch": 0.7141104294478527,
|
|
"grad_norm": 46.46655474433081,
|
|
"learning_rate": 3.849173553719008e-07,
|
|
"logits/chosen": -0.046875,
|
|
"logits/rejected": -0.259765625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.31,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.9453125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 582
|
|
},
|
|
{
|
|
"epoch": 0.7153374233128834,
|
|
"grad_norm": 38.88117000055558,
|
|
"learning_rate": 3.8471074380165284e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3145,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4296875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 583
|
|
},
|
|
{
|
|
"epoch": 0.7165644171779141,
|
|
"grad_norm": 41.99819867075411,
|
|
"learning_rate": 3.8450413223140494e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.2099609375,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.3344,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 4.375,
|
|
"rewards/rejected": -3.734375,
|
|
"step": 584
|
|
},
|
|
{
|
|
"epoch": 0.7177914110429447,
|
|
"grad_norm": 44.1611316963058,
|
|
"learning_rate": 3.84297520661157e-07,
|
|
"logits/chosen": -0.038818359375,
|
|
"logits/rejected": -0.1630859375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.3664,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 4.90625,
|
|
"rewards/rejected": -3.609375,
|
|
"step": 585
|
|
},
|
|
{
|
|
"epoch": 0.7190184049079754,
|
|
"grad_norm": 39.0964035012581,
|
|
"learning_rate": 3.840909090909091e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.271484375,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2973,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.0,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 586
|
|
},
|
|
{
|
|
"epoch": 0.7202453987730061,
|
|
"grad_norm": 36.08960275580907,
|
|
"learning_rate": 3.8388429752066114e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2916,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 5.28125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 587
|
|
},
|
|
{
|
|
"epoch": 0.7214723926380369,
|
|
"grad_norm": 26.657846221925215,
|
|
"learning_rate": 3.8367768595041324e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2231,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.8515625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 588
|
|
},
|
|
{
|
|
"epoch": 0.7226993865030675,
|
|
"grad_norm": 41.32671053142075,
|
|
"learning_rate": 3.834710743801653e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3126,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.69140625,
|
|
"rewards/margins": 4.5,
|
|
"rewards/rejected": -3.828125,
|
|
"step": 589
|
|
},
|
|
{
|
|
"epoch": 0.7239263803680982,
|
|
"grad_norm": 31.50139516862032,
|
|
"learning_rate": 3.8326446280991733e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2457,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 590
|
|
},
|
|
{
|
|
"epoch": 0.7251533742331289,
|
|
"grad_norm": 41.923208067017086,
|
|
"learning_rate": 3.830578512396694e-07,
|
|
"logits/chosen": -0.328125,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2429,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 591
|
|
},
|
|
{
|
|
"epoch": 0.7263803680981595,
|
|
"grad_norm": 36.03509452521486,
|
|
"learning_rate": 3.828512396694215e-07,
|
|
"logits/chosen": -0.1904296875,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2519,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 592
|
|
},
|
|
{
|
|
"epoch": 0.7276073619631902,
|
|
"grad_norm": 42.073646570880065,
|
|
"learning_rate": 3.826446280991735e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -362.0,
|
|
"loss": 0.2844,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -3.984375,
|
|
"step": 593
|
|
},
|
|
{
|
|
"epoch": 0.7288343558282209,
|
|
"grad_norm": 32.37131856233386,
|
|
"learning_rate": 3.824380165289256e-07,
|
|
"logits/chosen": -0.193359375,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2431,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 594
|
|
},
|
|
{
|
|
"epoch": 0.7300613496932515,
|
|
"grad_norm": 40.766534902822876,
|
|
"learning_rate": 3.8223140495867767e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2581,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 595
|
|
},
|
|
{
|
|
"epoch": 0.7312883435582822,
|
|
"grad_norm": 39.26882224861265,
|
|
"learning_rate": 3.820247933884297e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2749,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 596
|
|
},
|
|
{
|
|
"epoch": 0.7325153374233129,
|
|
"grad_norm": 39.569676008461244,
|
|
"learning_rate": 3.818181818181818e-07,
|
|
"logits/chosen": -0.0849609375,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -282.0,
|
|
"loss": 0.2937,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2265625,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 597
|
|
},
|
|
{
|
|
"epoch": 0.7337423312883435,
|
|
"grad_norm": 32.53349158917794,
|
|
"learning_rate": 3.8161157024793387e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -438.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2181,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 598
|
|
},
|
|
{
|
|
"epoch": 0.7349693251533742,
|
|
"grad_norm": 37.462844674823714,
|
|
"learning_rate": 3.814049586776859e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.45703125,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2935,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 599
|
|
},
|
|
{
|
|
"epoch": 0.7361963190184049,
|
|
"grad_norm": 40.34755354050605,
|
|
"learning_rate": 3.8119834710743796e-07,
|
|
"logits/chosen": -0.3125,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3262,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.7374233128834355,
|
|
"grad_norm": 31.83131835226018,
|
|
"learning_rate": 3.8099173553719006e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.2587,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.5546875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 601
|
|
},
|
|
{
|
|
"epoch": 0.7386503067484662,
|
|
"grad_norm": 27.556154921143015,
|
|
"learning_rate": 3.807851239669421e-07,
|
|
"logits/chosen": -0.0517578125,
|
|
"logits/rejected": -0.2158203125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.1812,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.4765625,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 602
|
|
},
|
|
{
|
|
"epoch": 0.7398773006134969,
|
|
"grad_norm": 33.37496659841441,
|
|
"learning_rate": 3.805785123966942e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2746,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 603
|
|
},
|
|
{
|
|
"epoch": 0.7411042944785277,
|
|
"grad_norm": 33.47399818424688,
|
|
"learning_rate": 3.8037190082644626e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2518,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 604
|
|
},
|
|
{
|
|
"epoch": 0.7423312883435583,
|
|
"grad_norm": 31.794522300753904,
|
|
"learning_rate": 3.8016528925619836e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2352,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 605
|
|
},
|
|
{
|
|
"epoch": 0.743558282208589,
|
|
"grad_norm": 39.11777077485029,
|
|
"learning_rate": 3.799586776859504e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2993,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3671875,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 606
|
|
},
|
|
{
|
|
"epoch": 0.7447852760736197,
|
|
"grad_norm": 35.13403324428258,
|
|
"learning_rate": 3.797520661157025e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2528,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 607
|
|
},
|
|
{
|
|
"epoch": 0.7460122699386503,
|
|
"grad_norm": 34.622853683902335,
|
|
"learning_rate": 3.795454545454545e-07,
|
|
"logits/chosen": -0.06396484375,
|
|
"logits/rejected": -0.302734375,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2671,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.43359375,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 608
|
|
},
|
|
{
|
|
"epoch": 0.747239263803681,
|
|
"grad_norm": 33.44784961871993,
|
|
"learning_rate": 3.793388429752066e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2401,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.78515625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 609
|
|
},
|
|
{
|
|
"epoch": 0.7484662576687117,
|
|
"grad_norm": 30.3709205845726,
|
|
"learning_rate": 3.7913223140495864e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.2211,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.3125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 610
|
|
},
|
|
{
|
|
"epoch": 0.7496932515337423,
|
|
"grad_norm": 47.71192907247348,
|
|
"learning_rate": 3.7892561983471074e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.331,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.14453125,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 611
|
|
},
|
|
{
|
|
"epoch": 0.750920245398773,
|
|
"grad_norm": 40.03785494441384,
|
|
"learning_rate": 3.787190082644628e-07,
|
|
"logits/chosen": -0.2041015625,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -352.0,
|
|
"loss": 0.327,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.6875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 612
|
|
},
|
|
{
|
|
"epoch": 0.7521472392638037,
|
|
"grad_norm": 39.68756652628583,
|
|
"learning_rate": 3.7851239669421484e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3199,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.369140625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 613
|
|
},
|
|
{
|
|
"epoch": 0.7533742331288343,
|
|
"grad_norm": 41.644849602297185,
|
|
"learning_rate": 3.7830578512396694e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2887,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.80859375,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 614
|
|
},
|
|
{
|
|
"epoch": 0.754601226993865,
|
|
"grad_norm": 37.7145942345175,
|
|
"learning_rate": 3.78099173553719e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2656,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0078125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 615
|
|
},
|
|
{
|
|
"epoch": 0.7558282208588957,
|
|
"grad_norm": 33.41765706894794,
|
|
"learning_rate": 3.778925619834711e-07,
|
|
"logits/chosen": -0.1865234375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -286.0,
|
|
"loss": 0.2479,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.26953125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 616
|
|
},
|
|
{
|
|
"epoch": 0.7570552147239263,
|
|
"grad_norm": 40.974699543395324,
|
|
"learning_rate": 3.776859504132231e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.3367,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.609375,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 617
|
|
},
|
|
{
|
|
"epoch": 0.758282208588957,
|
|
"grad_norm": 46.934415155128356,
|
|
"learning_rate": 3.774793388429752e-07,
|
|
"logits/chosen": -0.099609375,
|
|
"logits/rejected": -0.275390625,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3508,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 4.9375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 618
|
|
},
|
|
{
|
|
"epoch": 0.7595092024539877,
|
|
"grad_norm": 35.432918707651034,
|
|
"learning_rate": 3.7727272727272723e-07,
|
|
"logits/chosen": -0.12890625,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.2261,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -5.125,
|
|
"step": 619
|
|
},
|
|
{
|
|
"epoch": 0.7607361963190185,
|
|
"grad_norm": 44.522688117502966,
|
|
"learning_rate": 3.7706611570247933e-07,
|
|
"logits/chosen": -0.1318359375,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2954,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.625,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.5,
|
|
"step": 620
|
|
},
|
|
{
|
|
"epoch": 0.7619631901840491,
|
|
"grad_norm": 35.21620215877052,
|
|
"learning_rate": 3.768595041322314e-07,
|
|
"logits/chosen": -0.2412109375,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -478.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2617,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -5.375,
|
|
"step": 621
|
|
},
|
|
{
|
|
"epoch": 0.7631901840490798,
|
|
"grad_norm": 32.97202434149592,
|
|
"learning_rate": 3.766528925619835e-07,
|
|
"logits/chosen": -0.13671875,
|
|
"logits/rejected": -0.310546875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2271,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.73828125,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 622
|
|
},
|
|
{
|
|
"epoch": 0.7644171779141105,
|
|
"grad_norm": 40.73770796587866,
|
|
"learning_rate": 3.764462809917355e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -452.0,
|
|
"logps/rejected": -370.0,
|
|
"loss": 0.3008,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.875,
|
|
"step": 623
|
|
},
|
|
{
|
|
"epoch": 0.7656441717791411,
|
|
"grad_norm": 33.6312460876679,
|
|
"learning_rate": 3.762396694214876e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.218,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.1328125,
|
|
"rewards/margins": 6.78125,
|
|
"rewards/rejected": -5.65625,
|
|
"step": 624
|
|
},
|
|
{
|
|
"epoch": 0.7668711656441718,
|
|
"grad_norm": 35.01728220554292,
|
|
"learning_rate": 3.7603305785123967e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2202,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.015625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 625
|
|
},
|
|
{
|
|
"epoch": 0.7680981595092025,
|
|
"grad_norm": 43.4178761562483,
|
|
"learning_rate": 3.758264462809917e-07,
|
|
"logits/chosen": -0.314453125,
|
|
"logits/rejected": -0.46875,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3158,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 626
|
|
},
|
|
{
|
|
"epoch": 0.7693251533742331,
|
|
"grad_norm": 35.14066305401538,
|
|
"learning_rate": 3.7561983471074376e-07,
|
|
"logits/chosen": -0.1162109375,
|
|
"logits/rejected": -0.21484375,
|
|
"logps/chosen": -308.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3225,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": -0.06884765625,
|
|
"rewards/margins": 4.53125,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 627
|
|
},
|
|
{
|
|
"epoch": 0.7705521472392638,
|
|
"grad_norm": 43.80833337888977,
|
|
"learning_rate": 3.7541322314049586e-07,
|
|
"logits/chosen": -0.15234375,
|
|
"logits/rejected": -0.294921875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.3054,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.90234375,
|
|
"rewards/margins": 5.15625,
|
|
"rewards/rejected": -4.25,
|
|
"step": 628
|
|
},
|
|
{
|
|
"epoch": 0.7717791411042945,
|
|
"grad_norm": 24.76129586575429,
|
|
"learning_rate": 3.752066115702479e-07,
|
|
"logits/chosen": -0.171875,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -276.0,
|
|
"loss": 0.1831,
|
|
"rewards/accuracies": 0.9375,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 629
|
|
},
|
|
{
|
|
"epoch": 0.7730061349693251,
|
|
"grad_norm": 34.904501482497786,
|
|
"learning_rate": 3.75e-07,
|
|
"logits/chosen": -0.11328125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2492,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 630
|
|
},
|
|
{
|
|
"epoch": 0.7742331288343558,
|
|
"grad_norm": 37.86286725268997,
|
|
"learning_rate": 3.7479338842975206e-07,
|
|
"logits/chosen": -0.298828125,
|
|
"logits/rejected": -0.404296875,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3021,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 631
|
|
},
|
|
{
|
|
"epoch": 0.7754601226993865,
|
|
"grad_norm": 39.56376346395484,
|
|
"learning_rate": 3.745867768595041e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2667,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.671875,
|
|
"step": 632
|
|
},
|
|
{
|
|
"epoch": 0.7766871165644171,
|
|
"grad_norm": 43.15920072260989,
|
|
"learning_rate": 3.743801652892562e-07,
|
|
"logits/chosen": -0.1484375,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.317,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.7734375,
|
|
"rewards/margins": 4.71875,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 633
|
|
},
|
|
{
|
|
"epoch": 0.7779141104294478,
|
|
"grad_norm": 34.76469344160583,
|
|
"learning_rate": 3.7417355371900825e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.443359375,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2151,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.4375,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 634
|
|
},
|
|
{
|
|
"epoch": 0.7791411042944786,
|
|
"grad_norm": 32.416601067780334,
|
|
"learning_rate": 3.739669421487603e-07,
|
|
"logits/chosen": -0.2421875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -288.0,
|
|
"loss": 0.217,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.9140625,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 635
|
|
},
|
|
{
|
|
"epoch": 0.7803680981595092,
|
|
"grad_norm": 38.28241269033868,
|
|
"learning_rate": 3.7376033057851234e-07,
|
|
"logits/chosen": -0.234375,
|
|
"logits/rejected": -0.36328125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2749,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 636
|
|
},
|
|
{
|
|
"epoch": 0.7815950920245399,
|
|
"grad_norm": 47.424162576758796,
|
|
"learning_rate": 3.7355371900826445e-07,
|
|
"logits/chosen": -0.0859375,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -292.0,
|
|
"loss": 0.2932,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.890625,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 637
|
|
},
|
|
{
|
|
"epoch": 0.7828220858895706,
|
|
"grad_norm": 33.14638387389286,
|
|
"learning_rate": 3.733471074380165e-07,
|
|
"logits/chosen": -0.2177734375,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.225,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.125,
|
|
"step": 638
|
|
},
|
|
{
|
|
"epoch": 0.7840490797546013,
|
|
"grad_norm": 40.33146789211734,
|
|
"learning_rate": 3.731404958677686e-07,
|
|
"logits/chosen": -0.1015625,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3027,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.75,
|
|
"step": 639
|
|
},
|
|
{
|
|
"epoch": 0.7852760736196319,
|
|
"grad_norm": 40.99608723585622,
|
|
"learning_rate": 3.7293388429752064e-07,
|
|
"logits/chosen": -0.24609375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -444.0,
|
|
"logps/rejected": -376.0,
|
|
"loss": 0.2653,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.890625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.125,
|
|
"step": 640
|
|
},
|
|
{
|
|
"epoch": 0.7865030674846626,
|
|
"grad_norm": 46.58650892643996,
|
|
"learning_rate": 3.7272727272727274e-07,
|
|
"logits/chosen": -0.27734375,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.3815,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.5703125,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.515625,
|
|
"step": 641
|
|
},
|
|
{
|
|
"epoch": 0.7877300613496933,
|
|
"grad_norm": 28.471925622053803,
|
|
"learning_rate": 3.725206611570248e-07,
|
|
"logits/chosen": -0.185546875,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.2363,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -3.890625,
|
|
"step": 642
|
|
},
|
|
{
|
|
"epoch": 0.7889570552147239,
|
|
"grad_norm": 40.149407361877095,
|
|
"learning_rate": 3.723140495867769e-07,
|
|
"logits/chosen": -0.30859375,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -364.0,
|
|
"loss": 0.2571,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 643
|
|
},
|
|
{
|
|
"epoch": 0.7901840490797546,
|
|
"grad_norm": 43.580660836250416,
|
|
"learning_rate": 3.721074380165289e-07,
|
|
"logits/chosen": -0.2314453125,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2481,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 644
|
|
},
|
|
{
|
|
"epoch": 0.7914110429447853,
|
|
"grad_norm": 36.964537896873544,
|
|
"learning_rate": 3.71900826446281e-07,
|
|
"logits/chosen": -0.1728515625,
|
|
"logits/rejected": -0.380859375,
|
|
"logps/chosen": -330.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.2416,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.6484375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 645
|
|
},
|
|
{
|
|
"epoch": 0.7926380368098159,
|
|
"grad_norm": 36.18484120423215,
|
|
"learning_rate": 3.7169421487603303e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.285,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.0,
|
|
"step": 646
|
|
},
|
|
{
|
|
"epoch": 0.7938650306748466,
|
|
"grad_norm": 32.85091829554944,
|
|
"learning_rate": 3.7148760330578513e-07,
|
|
"logits/chosen": -0.1826171875,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2558,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9609375,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 647
|
|
},
|
|
{
|
|
"epoch": 0.7950920245398773,
|
|
"grad_norm": 42.502304621111726,
|
|
"learning_rate": 3.712809917355372e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3343,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.90625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 648
|
|
},
|
|
{
|
|
"epoch": 0.7963190184049079,
|
|
"grad_norm": 41.7626560131211,
|
|
"learning_rate": 3.710743801652892e-07,
|
|
"logits/chosen": -0.1064453125,
|
|
"logits/rejected": -0.2109375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.4139,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 4.59375,
|
|
"rewards/rejected": -3.4375,
|
|
"step": 649
|
|
},
|
|
{
|
|
"epoch": 0.7975460122699386,
|
|
"grad_norm": 40.035923383900304,
|
|
"learning_rate": 3.708677685950413e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.296875,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.328,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -3.375,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 0.7987730061349694,
|
|
"grad_norm": 41.02033222140374,
|
|
"learning_rate": 3.7066115702479337e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2567,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.78125,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 651
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"grad_norm": 43.99996770243787,
|
|
"learning_rate": 3.7045454545454547e-07,
|
|
"logits/chosen": -0.1943359375,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.304,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.796875,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 652
|
|
},
|
|
{
|
|
"epoch": 0.8012269938650307,
|
|
"grad_norm": 42.280947809979125,
|
|
"learning_rate": 3.7024793388429746e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2642,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 653
|
|
},
|
|
{
|
|
"epoch": 0.8024539877300614,
|
|
"grad_norm": 33.69377308507155,
|
|
"learning_rate": 3.7004132231404956e-07,
|
|
"logits/chosen": -0.12255859375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2251,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 654
|
|
},
|
|
{
|
|
"epoch": 0.803680981595092,
|
|
"grad_norm": 37.83873855183703,
|
|
"learning_rate": 3.698347107438016e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.35546875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2549,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.375,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 655
|
|
},
|
|
{
|
|
"epoch": 0.8049079754601227,
|
|
"grad_norm": 39.726835256559795,
|
|
"learning_rate": 3.696280991735537e-07,
|
|
"logits/chosen": -0.1171875,
|
|
"logits/rejected": -0.27734375,
|
|
"logps/chosen": -364.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3072,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 5.1875,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 656
|
|
},
|
|
{
|
|
"epoch": 0.8061349693251534,
|
|
"grad_norm": 32.94464991591826,
|
|
"learning_rate": 3.6942148760330576e-07,
|
|
"logits/chosen": -0.2578125,
|
|
"logits/rejected": -0.416015625,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2901,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.421875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -3.953125,
|
|
"step": 657
|
|
},
|
|
{
|
|
"epoch": 0.807361963190184,
|
|
"grad_norm": 35.33216238810386,
|
|
"learning_rate": 3.6921487603305786e-07,
|
|
"logits/chosen": -0.265625,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.265,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 658
|
|
},
|
|
{
|
|
"epoch": 0.8085889570552147,
|
|
"grad_norm": 35.903067216165454,
|
|
"learning_rate": 3.690082644628099e-07,
|
|
"logits/chosen": -0.2060546875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2874,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.65625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.34375,
|
|
"step": 659
|
|
},
|
|
{
|
|
"epoch": 0.8098159509202454,
|
|
"grad_norm": 33.43274680085514,
|
|
"learning_rate": 3.68801652892562e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2533,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 660
|
|
},
|
|
{
|
|
"epoch": 0.811042944785276,
|
|
"grad_norm": 42.73024720052954,
|
|
"learning_rate": 3.6859504132231405e-07,
|
|
"logits/chosen": -0.158203125,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -284.0,
|
|
"loss": 0.3128,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.6640625,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 661
|
|
},
|
|
{
|
|
"epoch": 0.8122699386503067,
|
|
"grad_norm": 29.594222057082835,
|
|
"learning_rate": 3.683884297520661e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.211,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.5859375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 662
|
|
},
|
|
{
|
|
"epoch": 0.8134969325153374,
|
|
"grad_norm": 30.816389985498457,
|
|
"learning_rate": 3.6818181818181815e-07,
|
|
"logits/chosen": -0.11669921875,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2494,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.0,
|
|
"step": 663
|
|
},
|
|
{
|
|
"epoch": 0.8147239263803681,
|
|
"grad_norm": 27.861290032166615,
|
|
"learning_rate": 3.6797520661157025e-07,
|
|
"logits/chosen": -0.373046875,
|
|
"logits/rejected": -0.51953125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.1714,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.5390625,
|
|
"rewards/margins": 6.8125,
|
|
"rewards/rejected": -5.28125,
|
|
"step": 664
|
|
},
|
|
{
|
|
"epoch": 0.8159509202453987,
|
|
"grad_norm": 43.48547648980068,
|
|
"learning_rate": 3.677685950413223e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.267578125,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.2535,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.765625,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 665
|
|
},
|
|
{
|
|
"epoch": 0.8171779141104294,
|
|
"grad_norm": 41.230514254119846,
|
|
"learning_rate": 3.6756198347107434e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3107,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.4375,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 666
|
|
},
|
|
{
|
|
"epoch": 0.8184049079754602,
|
|
"grad_norm": 38.330931790926165,
|
|
"learning_rate": 3.6735537190082644e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.396484375,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2562,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.21875,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 667
|
|
},
|
|
{
|
|
"epoch": 0.8196319018404908,
|
|
"grad_norm": 36.01171777931778,
|
|
"learning_rate": 3.671487603305785e-07,
|
|
"logits/chosen": -0.19921875,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2577,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 668
|
|
},
|
|
{
|
|
"epoch": 0.8208588957055215,
|
|
"grad_norm": 37.805581553660694,
|
|
"learning_rate": 3.669421487603306e-07,
|
|
"logits/chosen": -0.25390625,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.257,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.9375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.625,
|
|
"step": 669
|
|
},
|
|
{
|
|
"epoch": 0.8220858895705522,
|
|
"grad_norm": 34.18026783004868,
|
|
"learning_rate": 3.6673553719008263e-07,
|
|
"logits/chosen": -0.0791015625,
|
|
"logits/rejected": -0.22265625,
|
|
"logps/chosen": -324.0,
|
|
"logps/rejected": -290.0,
|
|
"loss": 0.2365,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 0.9296875,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 670
|
|
},
|
|
{
|
|
"epoch": 0.8233128834355828,
|
|
"grad_norm": 31.80438827189515,
|
|
"learning_rate": 3.665289256198347e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.242,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 671
|
|
},
|
|
{
|
|
"epoch": 0.8245398773006135,
|
|
"grad_norm": 34.31886268675687,
|
|
"learning_rate": 3.6632231404958673e-07,
|
|
"logits/chosen": -0.1591796875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -448.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2443,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.875,
|
|
"rewards/margins": 6.53125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 672
|
|
},
|
|
{
|
|
"epoch": 0.8257668711656442,
|
|
"grad_norm": 35.357570263918994,
|
|
"learning_rate": 3.6611570247933883e-07,
|
|
"logits/chosen": -0.11083984375,
|
|
"logits/rejected": -0.3359375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2364,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.1484375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 673
|
|
},
|
|
{
|
|
"epoch": 0.8269938650306748,
|
|
"grad_norm": 33.53684126770598,
|
|
"learning_rate": 3.659090909090909e-07,
|
|
"logits/chosen": -0.0654296875,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2467,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 674
|
|
},
|
|
{
|
|
"epoch": 0.8282208588957055,
|
|
"grad_norm": 34.52915039204892,
|
|
"learning_rate": 3.65702479338843e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2621,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.453125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 675
|
|
},
|
|
{
|
|
"epoch": 0.8294478527607362,
|
|
"grad_norm": 39.20023525071254,
|
|
"learning_rate": 3.65495867768595e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2552,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.640625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 676
|
|
},
|
|
{
|
|
"epoch": 0.8306748466257668,
|
|
"grad_norm": 40.9061677699551,
|
|
"learning_rate": 3.652892561983471e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.291015625,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2803,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.4609375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 677
|
|
},
|
|
{
|
|
"epoch": 0.8319018404907975,
|
|
"grad_norm": 35.838195895535726,
|
|
"learning_rate": 3.6508264462809917e-07,
|
|
"logits/chosen": -0.15625,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2691,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.88671875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 678
|
|
},
|
|
{
|
|
"epoch": 0.8331288343558282,
|
|
"grad_norm": 39.33502920792321,
|
|
"learning_rate": 3.648760330578512e-07,
|
|
"logits/chosen": -0.07080078125,
|
|
"logits/rejected": -0.1845703125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.284,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.94140625,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 679
|
|
},
|
|
{
|
|
"epoch": 0.8343558282208589,
|
|
"grad_norm": 35.15449592029737,
|
|
"learning_rate": 3.6466942148760326e-07,
|
|
"logits/chosen": -0.1923828125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -454.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2298,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 680
|
|
},
|
|
{
|
|
"epoch": 0.8355828220858895,
|
|
"grad_norm": 31.78034807618272,
|
|
"learning_rate": 3.6446280991735536e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -346.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.1963,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 0.8984375,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 681
|
|
},
|
|
{
|
|
"epoch": 0.8368098159509203,
|
|
"grad_norm": 39.09021083370229,
|
|
"learning_rate": 3.642561983471074e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.301,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.74609375,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 682
|
|
},
|
|
{
|
|
"epoch": 0.838036809815951,
|
|
"grad_norm": 36.30997939377773,
|
|
"learning_rate": 3.640495867768595e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2667,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.15625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -5.125,
|
|
"step": 683
|
|
},
|
|
{
|
|
"epoch": 0.8392638036809816,
|
|
"grad_norm": 43.34454211125793,
|
|
"learning_rate": 3.6384297520661156e-07,
|
|
"logits/chosen": -0.06298828125,
|
|
"logits/rejected": -0.18359375,
|
|
"logps/chosen": -314.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3642,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": -0.0263671875,
|
|
"rewards/margins": 4.6875,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 684
|
|
},
|
|
{
|
|
"epoch": 0.8404907975460123,
|
|
"grad_norm": 34.04376739318291,
|
|
"learning_rate": 3.636363636363636e-07,
|
|
"logits/chosen": -0.1669921875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -396.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2432,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0546875,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.84375,
|
|
"step": 685
|
|
},
|
|
{
|
|
"epoch": 0.841717791411043,
|
|
"grad_norm": 32.60171699566285,
|
|
"learning_rate": 3.634297520661157e-07,
|
|
"logits/chosen": -0.1787109375,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2556,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 686
|
|
},
|
|
{
|
|
"epoch": 0.8429447852760736,
|
|
"grad_norm": 36.65354869863078,
|
|
"learning_rate": 3.6322314049586775e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2628,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 687
|
|
},
|
|
{
|
|
"epoch": 0.8441717791411043,
|
|
"grad_norm": 41.14195765537747,
|
|
"learning_rate": 3.630165289256198e-07,
|
|
"logits/chosen": -0.20703125,
|
|
"logits/rejected": -0.369140625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2638,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.177734375,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -5.34375,
|
|
"step": 688
|
|
},
|
|
{
|
|
"epoch": 0.845398773006135,
|
|
"grad_norm": 38.51353105415181,
|
|
"learning_rate": 3.6280991735537185e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2884,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 689
|
|
},
|
|
{
|
|
"epoch": 0.8466257668711656,
|
|
"grad_norm": 50.698015097303056,
|
|
"learning_rate": 3.6260330578512395e-07,
|
|
"logits/chosen": -0.044677734375,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.3436,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.5234375,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 690
|
|
},
|
|
{
|
|
"epoch": 0.8478527607361963,
|
|
"grad_norm": 30.814234506471102,
|
|
"learning_rate": 3.62396694214876e-07,
|
|
"logits/chosen": -0.3359375,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2217,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.2421875,
|
|
"rewards/margins": 6.375,
|
|
"rewards/rejected": -5.125,
|
|
"step": 691
|
|
},
|
|
{
|
|
"epoch": 0.849079754601227,
|
|
"grad_norm": 43.15048361460468,
|
|
"learning_rate": 3.621900826446281e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.337890625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.2766,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 692
|
|
},
|
|
{
|
|
"epoch": 0.8503067484662576,
|
|
"grad_norm": 35.3095511118675,
|
|
"learning_rate": 3.6198347107438014e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -442.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2852,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.0625,
|
|
"rewards/margins": 7.25,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 693
|
|
},
|
|
{
|
|
"epoch": 0.8515337423312883,
|
|
"grad_norm": 50.92757487577681,
|
|
"learning_rate": 3.6177685950413224e-07,
|
|
"logits/chosen": -0.18359375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3663,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.73828125,
|
|
"rewards/margins": 4.96875,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 694
|
|
},
|
|
{
|
|
"epoch": 0.852760736196319,
|
|
"grad_norm": 50.927014909943495,
|
|
"learning_rate": 3.615702479338843e-07,
|
|
"logits/chosen": -0.1689453125,
|
|
"logits/rejected": -0.26953125,
|
|
"logps/chosen": -328.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.4583,
|
|
"rewards/accuracies": 0.734375,
|
|
"rewards/chosen": 0.5625,
|
|
"rewards/margins": 4.34375,
|
|
"rewards/rejected": -3.765625,
|
|
"step": 695
|
|
},
|
|
{
|
|
"epoch": 0.8539877300613496,
|
|
"grad_norm": 39.49040375200759,
|
|
"learning_rate": 3.613636363636364e-07,
|
|
"logits/chosen": -0.130859375,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2881,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.2734375,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.40625,
|
|
"step": 696
|
|
},
|
|
{
|
|
"epoch": 0.8552147239263803,
|
|
"grad_norm": 45.57900690356607,
|
|
"learning_rate": 3.611570247933884e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.328125,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.2838,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.8671875,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 697
|
|
},
|
|
{
|
|
"epoch": 0.8564417177914111,
|
|
"grad_norm": 38.81941112945181,
|
|
"learning_rate": 3.609504132231405e-07,
|
|
"logits/chosen": -0.1328125,
|
|
"logits/rejected": -0.279296875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2619,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 698
|
|
},
|
|
{
|
|
"epoch": 0.8576687116564418,
|
|
"grad_norm": 42.966064611435726,
|
|
"learning_rate": 3.6074380165289253e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.30078125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3034,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.5859375,
|
|
"rewards/margins": 5.25,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 699
|
|
},
|
|
{
|
|
"epoch": 0.8588957055214724,
|
|
"grad_norm": 42.65113203498414,
|
|
"learning_rate": 3.6053719008264463e-07,
|
|
"logits/chosen": -0.11767578125,
|
|
"logits/rejected": -0.2353515625,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3109,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1015625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.125,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.8601226993865031,
|
|
"grad_norm": 45.17083911298038,
|
|
"learning_rate": 3.603305785123967e-07,
|
|
"logits/chosen": -0.08056640625,
|
|
"logits/rejected": -0.1982421875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3654,
|
|
"rewards/accuracies": 0.8046875,
|
|
"rewards/chosen": 0.75,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 701
|
|
},
|
|
{
|
|
"epoch": 0.8613496932515338,
|
|
"grad_norm": 31.973002163499245,
|
|
"learning_rate": 3.601239669421487e-07,
|
|
"logits/chosen": -0.2080078125,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.1933,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8125,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 702
|
|
},
|
|
{
|
|
"epoch": 0.8625766871165644,
|
|
"grad_norm": 36.36498243165029,
|
|
"learning_rate": 3.599173553719008e-07,
|
|
"logits/chosen": -0.10400390625,
|
|
"logits/rejected": -0.25,
|
|
"logps/chosen": -414.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3136,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.203125,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 703
|
|
},
|
|
{
|
|
"epoch": 0.8638036809815951,
|
|
"grad_norm": 35.79987216653264,
|
|
"learning_rate": 3.5971074380165287e-07,
|
|
"logits/chosen": -0.173828125,
|
|
"logits/rejected": -0.349609375,
|
|
"logps/chosen": -322.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2837,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.1171875,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.6875,
|
|
"step": 704
|
|
},
|
|
{
|
|
"epoch": 0.8650306748466258,
|
|
"grad_norm": 36.827745109710804,
|
|
"learning_rate": 3.5950413223140497e-07,
|
|
"logits/chosen": -0.279296875,
|
|
"logits/rejected": -0.41015625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -350.0,
|
|
"loss": 0.2403,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 2.09375,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 705
|
|
},
|
|
{
|
|
"epoch": 0.8662576687116564,
|
|
"grad_norm": 36.25592355489759,
|
|
"learning_rate": 3.5929752066115697e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.41796875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2682,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.3984375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 706
|
|
},
|
|
{
|
|
"epoch": 0.8674846625766871,
|
|
"grad_norm": 33.13517607643058,
|
|
"learning_rate": 3.5909090909090907e-07,
|
|
"logits/chosen": -0.228515625,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2352,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 707
|
|
},
|
|
{
|
|
"epoch": 0.8687116564417178,
|
|
"grad_norm": 29.108052314139623,
|
|
"learning_rate": 3.588842975206611e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.4296875,
|
|
"logps/chosen": -370.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2018,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.625,
|
|
"step": 708
|
|
},
|
|
{
|
|
"epoch": 0.8699386503067484,
|
|
"grad_norm": 26.864603379868992,
|
|
"learning_rate": 3.586776859504132e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.184,
|
|
"rewards/accuracies": 0.953125,
|
|
"rewards/chosen": 1.8046875,
|
|
"rewards/margins": 7.0,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 709
|
|
},
|
|
{
|
|
"epoch": 0.8711656441717791,
|
|
"grad_norm": 39.09283390375149,
|
|
"learning_rate": 3.5847107438016526e-07,
|
|
"logits/chosen": -0.0732421875,
|
|
"logits/rejected": -0.2109375,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3161,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 1.0859375,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 710
|
|
},
|
|
{
|
|
"epoch": 0.8723926380368098,
|
|
"grad_norm": 29.706815103224322,
|
|
"learning_rate": 3.5826446280991736e-07,
|
|
"logits/chosen": -0.09912109375,
|
|
"logits/rejected": -0.29296875,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2257,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.375,
|
|
"step": 711
|
|
},
|
|
{
|
|
"epoch": 0.8736196319018404,
|
|
"grad_norm": 44.764736935363494,
|
|
"learning_rate": 3.580578512396694e-07,
|
|
"logits/chosen": -0.08984375,
|
|
"logits/rejected": -0.244140625,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.3157,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.984375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.03125,
|
|
"step": 712
|
|
},
|
|
{
|
|
"epoch": 0.8748466257668711,
|
|
"grad_norm": 33.83006454123061,
|
|
"learning_rate": 3.578512396694215e-07,
|
|
"logits/chosen": -0.1748046875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2531,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.234375,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 713
|
|
},
|
|
{
|
|
"epoch": 0.8760736196319019,
|
|
"grad_norm": 37.79578296577771,
|
|
"learning_rate": 3.5764462809917355e-07,
|
|
"logits/chosen": -0.07177734375,
|
|
"logits/rejected": -0.28515625,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2529,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.21875,
|
|
"step": 714
|
|
},
|
|
{
|
|
"epoch": 0.8773006134969326,
|
|
"grad_norm": 33.397821944610435,
|
|
"learning_rate": 3.574380165289256e-07,
|
|
"logits/chosen": -0.26953125,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2448,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 715
|
|
},
|
|
{
|
|
"epoch": 0.8785276073619632,
|
|
"grad_norm": 38.599098942966314,
|
|
"learning_rate": 3.5723140495867765e-07,
|
|
"logits/chosen": -0.236328125,
|
|
"logits/rejected": -0.423828125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.3027,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.09375,
|
|
"step": 716
|
|
},
|
|
{
|
|
"epoch": 0.8797546012269939,
|
|
"grad_norm": 35.160888357665144,
|
|
"learning_rate": 3.5702479338842975e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -268.0,
|
|
"loss": 0.2911,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.92578125,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 717
|
|
},
|
|
{
|
|
"epoch": 0.8809815950920246,
|
|
"grad_norm": 26.612419306861238,
|
|
"learning_rate": 3.568181818181818e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.1966,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 1.859375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.3125,
|
|
"step": 718
|
|
},
|
|
{
|
|
"epoch": 0.8822085889570552,
|
|
"grad_norm": 39.83118714795231,
|
|
"learning_rate": 3.5661157024793384e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.3984375,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2762,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.203125,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 719
|
|
},
|
|
{
|
|
"epoch": 0.8834355828220859,
|
|
"grad_norm": 30.80915305318243,
|
|
"learning_rate": 3.5640495867768594e-07,
|
|
"logits/chosen": -0.267578125,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.218,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 2.171875,
|
|
"rewards/margins": 6.125,
|
|
"rewards/rejected": -3.96875,
|
|
"step": 720
|
|
},
|
|
{
|
|
"epoch": 0.8846625766871166,
|
|
"grad_norm": 27.499435353414444,
|
|
"learning_rate": 3.56198347107438e-07,
|
|
"logits/chosen": -0.23828125,
|
|
"logits/rejected": -0.498046875,
|
|
"logps/chosen": -412.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.1999,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.7890625,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 721
|
|
},
|
|
{
|
|
"epoch": 0.8858895705521472,
|
|
"grad_norm": 32.55228038139476,
|
|
"learning_rate": 3.559917355371901e-07,
|
|
"logits/chosen": -0.2353515625,
|
|
"logits/rejected": -0.333984375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.222,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.8359375,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.25,
|
|
"step": 722
|
|
},
|
|
{
|
|
"epoch": 0.8871165644171779,
|
|
"grad_norm": 38.73866617415686,
|
|
"learning_rate": 3.5578512396694214e-07,
|
|
"logits/chosen": -0.23046875,
|
|
"logits/rejected": -0.3515625,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2777,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.3046875,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 723
|
|
},
|
|
{
|
|
"epoch": 0.8883435582822086,
|
|
"grad_norm": 36.524786654594486,
|
|
"learning_rate": 3.555785123966942e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2442,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.6328125,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.375,
|
|
"step": 724
|
|
},
|
|
{
|
|
"epoch": 0.8895705521472392,
|
|
"grad_norm": 38.158440355258406,
|
|
"learning_rate": 3.5537190082644623e-07,
|
|
"logits/chosen": -0.10986328125,
|
|
"logits/rejected": -0.283203125,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.283,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.390625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 725
|
|
},
|
|
{
|
|
"epoch": 0.8907975460122699,
|
|
"grad_norm": 27.73103579824844,
|
|
"learning_rate": 3.5516528925619833e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -334.0,
|
|
"logps/rejected": -280.0,
|
|
"loss": 0.1949,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.0703125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 726
|
|
},
|
|
{
|
|
"epoch": 0.8920245398773006,
|
|
"grad_norm": 34.09534371161639,
|
|
"learning_rate": 3.549586776859504e-07,
|
|
"logits/chosen": -0.2001953125,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2172,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 5.78125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 727
|
|
},
|
|
{
|
|
"epoch": 0.8932515337423312,
|
|
"grad_norm": 39.39112103020417,
|
|
"learning_rate": 3.547520661157025e-07,
|
|
"logits/chosen": -0.208984375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2626,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.9765625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -4.59375,
|
|
"step": 728
|
|
},
|
|
{
|
|
"epoch": 0.894478527607362,
|
|
"grad_norm": 37.23604455297992,
|
|
"learning_rate": 3.545454545454545e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.37890625,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2356,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.640625,
|
|
"rewards/margins": 6.59375,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 729
|
|
},
|
|
{
|
|
"epoch": 0.8957055214723927,
|
|
"grad_norm": 32.56560137662825,
|
|
"learning_rate": 3.543388429752066e-07,
|
|
"logits/chosen": -0.294921875,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2002,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 2.1875,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -4.75,
|
|
"step": 730
|
|
},
|
|
{
|
|
"epoch": 0.8969325153374234,
|
|
"grad_norm": 37.581323988938905,
|
|
"learning_rate": 3.5413223140495867e-07,
|
|
"logits/chosen": -0.1044921875,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -430.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2709,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.40625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.75,
|
|
"step": 731
|
|
},
|
|
{
|
|
"epoch": 0.898159509202454,
|
|
"grad_norm": 42.424019347418515,
|
|
"learning_rate": 3.5392561983471077e-07,
|
|
"logits/chosen": -0.2451171875,
|
|
"logits/rejected": -0.419921875,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3232,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 732
|
|
},
|
|
{
|
|
"epoch": 0.8993865030674847,
|
|
"grad_norm": 30.46815348460911,
|
|
"learning_rate": 3.5371900826446277e-07,
|
|
"logits/chosen": -0.16015625,
|
|
"logits/rejected": -0.3203125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.205,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 733
|
|
},
|
|
{
|
|
"epoch": 0.9006134969325154,
|
|
"grad_norm": 40.11598742347428,
|
|
"learning_rate": 3.5351239669421487e-07,
|
|
"logits/chosen": -0.1337890625,
|
|
"logits/rejected": -0.322265625,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2734,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.125,
|
|
"step": 734
|
|
},
|
|
{
|
|
"epoch": 0.901840490797546,
|
|
"grad_norm": 30.987246843568794,
|
|
"learning_rate": 3.533057851239669e-07,
|
|
"logits/chosen": -0.2392578125,
|
|
"logits/rejected": -0.47265625,
|
|
"logps/chosen": -422.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.1867,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 1.4921875,
|
|
"rewards/margins": 6.96875,
|
|
"rewards/rejected": -5.46875,
|
|
"step": 735
|
|
},
|
|
{
|
|
"epoch": 0.9030674846625767,
|
|
"grad_norm": 39.665000609602814,
|
|
"learning_rate": 3.53099173553719e-07,
|
|
"logits/chosen": -0.08203125,
|
|
"logits/rejected": -0.2138671875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.3163,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.3671875,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 736
|
|
},
|
|
{
|
|
"epoch": 0.9042944785276074,
|
|
"grad_norm": 45.99636093708745,
|
|
"learning_rate": 3.5289256198347106e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.28125,
|
|
"logps/chosen": -350.0,
|
|
"logps/rejected": -322.0,
|
|
"loss": 0.3995,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.5546875,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -5.0,
|
|
"step": 737
|
|
},
|
|
{
|
|
"epoch": 0.905521472392638,
|
|
"grad_norm": 34.94091041398414,
|
|
"learning_rate": 3.526859504132231e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -440.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2497,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 6.5625,
|
|
"rewards/rejected": -5.0,
|
|
"step": 738
|
|
},
|
|
{
|
|
"epoch": 0.9067484662576687,
|
|
"grad_norm": 37.70067993632191,
|
|
"learning_rate": 3.524793388429752e-07,
|
|
"logits/chosen": -0.1650390625,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2834,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.671875,
|
|
"rewards/margins": 5.65625,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 739
|
|
},
|
|
{
|
|
"epoch": 0.9079754601226994,
|
|
"grad_norm": 33.58911831683584,
|
|
"learning_rate": 3.5227272727272725e-07,
|
|
"logits/chosen": -0.22265625,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2496,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.60546875,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.5,
|
|
"step": 740
|
|
},
|
|
{
|
|
"epoch": 0.90920245398773,
|
|
"grad_norm": 35.571601551599564,
|
|
"learning_rate": 3.5206611570247935e-07,
|
|
"logits/chosen": -0.2119140625,
|
|
"logits/rejected": -0.38671875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -306.0,
|
|
"loss": 0.2619,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.3515625,
|
|
"rewards/margins": 6.90625,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 741
|
|
},
|
|
{
|
|
"epoch": 0.9104294478527607,
|
|
"grad_norm": 31.812391536470038,
|
|
"learning_rate": 3.5185950413223135e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2088,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.484375,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.53125,
|
|
"step": 742
|
|
},
|
|
{
|
|
"epoch": 0.9116564417177914,
|
|
"grad_norm": 36.766990867423736,
|
|
"learning_rate": 3.5165289256198345e-07,
|
|
"logits/chosen": -0.19140625,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -368.0,
|
|
"loss": 0.2538,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.0390625,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 743
|
|
},
|
|
{
|
|
"epoch": 0.912883435582822,
|
|
"grad_norm": 30.797563569261442,
|
|
"learning_rate": 3.514462809917355e-07,
|
|
"logits/chosen": -0.1796875,
|
|
"logits/rejected": -0.373046875,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2174,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.5546875,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.5,
|
|
"step": 744
|
|
},
|
|
{
|
|
"epoch": 0.9141104294478528,
|
|
"grad_norm": 37.182331193903785,
|
|
"learning_rate": 3.512396694214876e-07,
|
|
"logits/chosen": -0.1376953125,
|
|
"logits/rejected": -0.2451171875,
|
|
"logps/chosen": -342.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2759,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.515625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.875,
|
|
"step": 745
|
|
},
|
|
{
|
|
"epoch": 0.9153374233128835,
|
|
"grad_norm": 42.37840427394414,
|
|
"learning_rate": 3.5103305785123964e-07,
|
|
"logits/chosen": -0.140625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.2818,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.96484375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -4.96875,
|
|
"step": 746
|
|
},
|
|
{
|
|
"epoch": 0.9165644171779141,
|
|
"grad_norm": 36.09849932082116,
|
|
"learning_rate": 3.5082644628099174e-07,
|
|
"logits/chosen": -0.306640625,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -338.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2741,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.51171875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -4.8125,
|
|
"step": 747
|
|
},
|
|
{
|
|
"epoch": 0.9177914110429448,
|
|
"grad_norm": 39.853022506318325,
|
|
"learning_rate": 3.506198347107438e-07,
|
|
"logits/chosen": -0.11376953125,
|
|
"logits/rejected": -0.2314453125,
|
|
"logps/chosen": -372.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.3302,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.205078125,
|
|
"rewards/margins": 4.875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 748
|
|
},
|
|
{
|
|
"epoch": 0.9190184049079755,
|
|
"grad_norm": 40.89419641030945,
|
|
"learning_rate": 3.504132231404959e-07,
|
|
"logits/chosen": -0.2275390625,
|
|
"logits/rejected": -0.4375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3306,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 1.046875,
|
|
"rewards/margins": 6.1875,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 749
|
|
},
|
|
{
|
|
"epoch": 0.9202453987730062,
|
|
"grad_norm": 48.1673918296563,
|
|
"learning_rate": 3.5020661157024794e-07,
|
|
"logits/chosen": -0.212890625,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -406.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3386,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.66796875,
|
|
"rewards/margins": 5.75,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 0.9214723926380368,
|
|
"grad_norm": 43.63193512219085,
|
|
"learning_rate": 3.5e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.384765625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2867,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 6.4375,
|
|
"rewards/rejected": -5.25,
|
|
"step": 751
|
|
},
|
|
{
|
|
"epoch": 0.9226993865030675,
|
|
"grad_norm": 35.77618155880579,
|
|
"learning_rate": 3.4979338842975203e-07,
|
|
"logits/chosen": -0.216796875,
|
|
"logits/rejected": -0.478515625,
|
|
"logps/chosen": -418.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2401,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.359375,
|
|
"rewards/margins": 6.875,
|
|
"rewards/rejected": -5.5,
|
|
"step": 752
|
|
},
|
|
{
|
|
"epoch": 0.9239263803680982,
|
|
"grad_norm": 38.24604411902878,
|
|
"learning_rate": 3.4958677685950413e-07,
|
|
"logits/chosen": -0.1181640625,
|
|
"logits/rejected": -0.265625,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2709,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 5.59375,
|
|
"rewards/rejected": -4.875,
|
|
"step": 753
|
|
},
|
|
{
|
|
"epoch": 0.9251533742331288,
|
|
"grad_norm": 43.12286562619324,
|
|
"learning_rate": 3.493801652892562e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.345703125,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.3289,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.66015625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 754
|
|
},
|
|
{
|
|
"epoch": 0.9263803680981595,
|
|
"grad_norm": 34.65065790431959,
|
|
"learning_rate": 3.491735537190082e-07,
|
|
"logits/chosen": -0.2216796875,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2314,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.87109375,
|
|
"rewards/margins": 5.90625,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 755
|
|
},
|
|
{
|
|
"epoch": 0.9276073619631902,
|
|
"grad_norm": 32.01115835369059,
|
|
"learning_rate": 3.489669421487603e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.2294921875,
|
|
"logps/chosen": -332.0,
|
|
"logps/rejected": -302.0,
|
|
"loss": 0.2247,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.390625,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -5.03125,
|
|
"step": 756
|
|
},
|
|
{
|
|
"epoch": 0.9288343558282208,
|
|
"grad_norm": 29.79853888885103,
|
|
"learning_rate": 3.4876033057851237e-07,
|
|
"logits/chosen": -0.087890625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.2143,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.4140625,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.875,
|
|
"step": 757
|
|
},
|
|
{
|
|
"epoch": 0.9300613496932515,
|
|
"grad_norm": 41.4434512511522,
|
|
"learning_rate": 3.4855371900826447e-07,
|
|
"logits/chosen": -0.142578125,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3204,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.494140625,
|
|
"rewards/margins": 5.40625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 758
|
|
},
|
|
{
|
|
"epoch": 0.9312883435582822,
|
|
"grad_norm": 33.101256433680135,
|
|
"learning_rate": 3.483471074380165e-07,
|
|
"logits/chosen": -0.1123046875,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2334,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.28125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 759
|
|
},
|
|
{
|
|
"epoch": 0.9325153374233128,
|
|
"grad_norm": 31.858866512033583,
|
|
"learning_rate": 3.4814049586776857e-07,
|
|
"logits/chosen": -0.275390625,
|
|
"logits/rejected": -0.53515625,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -342.0,
|
|
"loss": 0.2656,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 2.046875,
|
|
"rewards/margins": 7.03125,
|
|
"rewards/rejected": -5.0,
|
|
"step": 760
|
|
},
|
|
{
|
|
"epoch": 0.9337423312883436,
|
|
"grad_norm": 42.354453289827106,
|
|
"learning_rate": 3.479338842975206e-07,
|
|
"logits/chosen": -0.08251953125,
|
|
"logits/rejected": -0.232421875,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.258,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.3359375,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 761
|
|
},
|
|
{
|
|
"epoch": 0.9349693251533743,
|
|
"grad_norm": 38.24811318296336,
|
|
"learning_rate": 3.477272727272727e-07,
|
|
"logits/chosen": -0.0703125,
|
|
"logits/rejected": -0.25390625,
|
|
"logps/chosen": -420.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.2519,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.1953125,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 762
|
|
},
|
|
{
|
|
"epoch": 0.9361963190184049,
|
|
"grad_norm": 36.20016344003986,
|
|
"learning_rate": 3.4752066115702476e-07,
|
|
"logits/chosen": -0.1357421875,
|
|
"logits/rejected": -0.2333984375,
|
|
"logps/chosen": -388.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2628,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.578125,
|
|
"rewards/margins": 6.25,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 763
|
|
},
|
|
{
|
|
"epoch": 0.9374233128834356,
|
|
"grad_norm": 37.48666102403677,
|
|
"learning_rate": 3.4731404958677686e-07,
|
|
"logits/chosen": -0.330078125,
|
|
"logits/rejected": -0.490234375,
|
|
"logps/chosen": -320.0,
|
|
"logps/rejected": -294.0,
|
|
"loss": 0.2761,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.447265625,
|
|
"rewards/margins": 5.625,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 764
|
|
},
|
|
{
|
|
"epoch": 0.9386503067484663,
|
|
"grad_norm": 44.1796891457386,
|
|
"learning_rate": 3.471074380165289e-07,
|
|
"logits/chosen": -0.126953125,
|
|
"logits/rejected": -0.2578125,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.3388,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.89453125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.15625,
|
|
"step": 765
|
|
},
|
|
{
|
|
"epoch": 0.939877300613497,
|
|
"grad_norm": 36.01121716925078,
|
|
"learning_rate": 3.46900826446281e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.37109375,
|
|
"logps/chosen": -434.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.268,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.546875,
|
|
"rewards/margins": 6.21875,
|
|
"rewards/rejected": -4.65625,
|
|
"step": 766
|
|
},
|
|
{
|
|
"epoch": 0.9411042944785276,
|
|
"grad_norm": 36.56160162737348,
|
|
"learning_rate": 3.4669421487603306e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.33984375,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.3045,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.2890625,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -4.75,
|
|
"step": 767
|
|
},
|
|
{
|
|
"epoch": 0.9423312883435583,
|
|
"grad_norm": 40.47132685816318,
|
|
"learning_rate": 3.464876033057851e-07,
|
|
"logits/chosen": -0.2021484375,
|
|
"logits/rejected": -0.3828125,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.2909,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.765625,
|
|
"rewards/margins": 6.3125,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 768
|
|
},
|
|
{
|
|
"epoch": 0.943558282208589,
|
|
"grad_norm": 30.62201594248347,
|
|
"learning_rate": 3.4628099173553715e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.361328125,
|
|
"logps/chosen": -366.0,
|
|
"logps/rejected": -318.0,
|
|
"loss": 0.2247,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.09375,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 769
|
|
},
|
|
{
|
|
"epoch": 0.9447852760736196,
|
|
"grad_norm": 43.60373632493513,
|
|
"learning_rate": 3.4607438016528925e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.474609375,
|
|
"logps/chosen": -404.0,
|
|
"logps/rejected": -348.0,
|
|
"loss": 0.2772,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 2.078125,
|
|
"rewards/margins": 6.28125,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 770
|
|
},
|
|
{
|
|
"epoch": 0.9460122699386503,
|
|
"grad_norm": 45.17582091386918,
|
|
"learning_rate": 3.458677685950413e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.26171875,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.3922,
|
|
"rewards/accuracies": 0.78125,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -4.25,
|
|
"step": 771
|
|
},
|
|
{
|
|
"epoch": 0.947239263803681,
|
|
"grad_norm": 42.297058717421166,
|
|
"learning_rate": 3.456611570247934e-07,
|
|
"logits/chosen": -0.0947265625,
|
|
"logits/rejected": -0.31640625,
|
|
"logps/chosen": -410.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.295,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.734375,
|
|
"rewards/margins": 6.0,
|
|
"rewards/rejected": -4.28125,
|
|
"step": 772
|
|
},
|
|
{
|
|
"epoch": 0.9484662576687116,
|
|
"grad_norm": 36.5037310282473,
|
|
"learning_rate": 3.4545454545454544e-07,
|
|
"logits/chosen": -0.10009765625,
|
|
"logits/rejected": -0.287109375,
|
|
"logps/chosen": -360.0,
|
|
"logps/rejected": -298.0,
|
|
"loss": 0.2904,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.7890625,
|
|
"rewards/margins": 5.21875,
|
|
"rewards/rejected": -4.4375,
|
|
"step": 773
|
|
},
|
|
{
|
|
"epoch": 0.9496932515337423,
|
|
"grad_norm": 38.086980897558234,
|
|
"learning_rate": 3.452479338842975e-07,
|
|
"logits/chosen": -0.1240234375,
|
|
"logits/rejected": -0.263671875,
|
|
"logps/chosen": -344.0,
|
|
"logps/rejected": -296.0,
|
|
"loss": 0.3096,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.458984375,
|
|
"rewards/margins": 4.09375,
|
|
"rewards/rejected": -3.640625,
|
|
"step": 774
|
|
},
|
|
{
|
|
"epoch": 0.950920245398773,
|
|
"grad_norm": 36.85278680079692,
|
|
"learning_rate": 3.450413223140496e-07,
|
|
"logits/chosen": -0.03271484375,
|
|
"logits/rejected": -0.2275390625,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2823,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.515625,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -3.578125,
|
|
"step": 775
|
|
},
|
|
{
|
|
"epoch": 0.9521472392638037,
|
|
"grad_norm": 34.94073241862638,
|
|
"learning_rate": 3.4483471074380164e-07,
|
|
"logits/chosen": -0.2265625,
|
|
"logits/rejected": -0.412109375,
|
|
"logps/chosen": -398.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.2924,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.796875,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 776
|
|
},
|
|
{
|
|
"epoch": 0.9533742331288344,
|
|
"grad_norm": 36.22271519573166,
|
|
"learning_rate": 3.446280991735537e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.34375,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.2907,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 1.2109375,
|
|
"rewards/margins": 5.46875,
|
|
"rewards/rejected": -4.25,
|
|
"step": 777
|
|
},
|
|
{
|
|
"epoch": 0.9546012269938651,
|
|
"grad_norm": 38.284746149580194,
|
|
"learning_rate": 3.4442148760330573e-07,
|
|
"logits/chosen": -0.16796875,
|
|
"logits/rejected": -0.375,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2391,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 5.03125,
|
|
"rewards/rejected": -3.90625,
|
|
"step": 778
|
|
},
|
|
{
|
|
"epoch": 0.9558282208588957,
|
|
"grad_norm": 42.74163715791748,
|
|
"learning_rate": 3.4421487603305783e-07,
|
|
"logits/chosen": -0.2236328125,
|
|
"logits/rejected": -0.40234375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -332.0,
|
|
"loss": 0.2504,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 1.6015625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 779
|
|
},
|
|
{
|
|
"epoch": 0.9570552147239264,
|
|
"grad_norm": 39.028284362830085,
|
|
"learning_rate": 3.440082644628099e-07,
|
|
"logits/chosen": -0.1259765625,
|
|
"logits/rejected": -0.298828125,
|
|
"logps/chosen": -368.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.327,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.81640625,
|
|
"rewards/margins": 5.0,
|
|
"rewards/rejected": -4.1875,
|
|
"step": 780
|
|
},
|
|
{
|
|
"epoch": 0.9582822085889571,
|
|
"grad_norm": 29.49560218073861,
|
|
"learning_rate": 3.43801652892562e-07,
|
|
"logits/chosen": -0.2294921875,
|
|
"logits/rejected": -0.458984375,
|
|
"logps/chosen": -416.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.1947,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.921875,
|
|
"rewards/margins": 6.9375,
|
|
"rewards/rejected": -5.0,
|
|
"step": 781
|
|
},
|
|
{
|
|
"epoch": 0.9595092024539877,
|
|
"grad_norm": 30.862418691296618,
|
|
"learning_rate": 3.43595041322314e-07,
|
|
"logits/chosen": -0.1962890625,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -358.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2276,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 1.1640625,
|
|
"rewards/margins": 5.53125,
|
|
"rewards/rejected": -4.375,
|
|
"step": 782
|
|
},
|
|
{
|
|
"epoch": 0.9607361963190184,
|
|
"grad_norm": 32.30854631034835,
|
|
"learning_rate": 3.433884297520661e-07,
|
|
"logits/chosen": -0.2138671875,
|
|
"logits/rejected": -0.3671875,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -344.0,
|
|
"loss": 0.2037,
|
|
"rewards/accuracies": 0.8984375,
|
|
"rewards/chosen": 1.609375,
|
|
"rewards/margins": 6.46875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 783
|
|
},
|
|
{
|
|
"epoch": 0.9619631901840491,
|
|
"grad_norm": 44.933495412305525,
|
|
"learning_rate": 3.4318181818181817e-07,
|
|
"logits/chosen": -0.189453125,
|
|
"logits/rejected": -0.2412109375,
|
|
"logps/chosen": -376.0,
|
|
"logps/rejected": -330.0,
|
|
"loss": 0.3216,
|
|
"rewards/accuracies": 0.828125,
|
|
"rewards/chosen": 0.421875,
|
|
"rewards/margins": 5.125,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 784
|
|
},
|
|
{
|
|
"epoch": 0.9631901840490797,
|
|
"grad_norm": 35.45021237079088,
|
|
"learning_rate": 3.429752066115703e-07,
|
|
"logits/chosen": -0.146484375,
|
|
"logits/rejected": -0.314453125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2635,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.03125,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 785
|
|
},
|
|
{
|
|
"epoch": 0.9644171779141104,
|
|
"grad_norm": 31.822348766627872,
|
|
"learning_rate": 3.4276859504132227e-07,
|
|
"logits/chosen": -0.1396484375,
|
|
"logits/rejected": -0.34765625,
|
|
"logps/chosen": -362.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2296,
|
|
"rewards/accuracies": 0.9140625,
|
|
"rewards/chosen": 0.5703125,
|
|
"rewards/margins": 5.8125,
|
|
"rewards/rejected": -5.21875,
|
|
"step": 786
|
|
},
|
|
{
|
|
"epoch": 0.9656441717791411,
|
|
"grad_norm": 35.86043202347341,
|
|
"learning_rate": 3.4256198347107437e-07,
|
|
"logits/chosen": -0.21484375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -402.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2274,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 0.78125,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.625,
|
|
"step": 787
|
|
},
|
|
{
|
|
"epoch": 0.9668711656441717,
|
|
"grad_norm": 26.552313321059188,
|
|
"learning_rate": 3.423553719008264e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.408203125,
|
|
"logps/chosen": -382.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.1667,
|
|
"rewards/accuracies": 0.9296875,
|
|
"rewards/chosen": 1.1875,
|
|
"rewards/margins": 6.75,
|
|
"rewards/rejected": -5.5625,
|
|
"step": 788
|
|
},
|
|
{
|
|
"epoch": 0.9680981595092024,
|
|
"grad_norm": 33.30240522771134,
|
|
"learning_rate": 3.421487603305785e-07,
|
|
"logits/chosen": -0.1474609375,
|
|
"logits/rejected": -0.330078125,
|
|
"logps/chosen": -374.0,
|
|
"logps/rejected": -308.0,
|
|
"loss": 0.2769,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.419921875,
|
|
"rewards/margins": 5.4375,
|
|
"rewards/rejected": -5.0,
|
|
"step": 789
|
|
},
|
|
{
|
|
"epoch": 0.9693251533742331,
|
|
"grad_norm": 35.332506360333724,
|
|
"learning_rate": 3.4194214876033056e-07,
|
|
"logits/chosen": -0.263671875,
|
|
"logits/rejected": -0.484375,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -310.0,
|
|
"loss": 0.2259,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.71484375,
|
|
"rewards/margins": 6.03125,
|
|
"rewards/rejected": -5.3125,
|
|
"step": 790
|
|
},
|
|
{
|
|
"epoch": 0.9705521472392638,
|
|
"grad_norm": 35.242935109364,
|
|
"learning_rate": 3.417355371900826e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.4453125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2483,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 1.1796875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.6875,
|
|
"step": 791
|
|
},
|
|
{
|
|
"epoch": 0.9717791411042945,
|
|
"grad_norm": 38.00481733058289,
|
|
"learning_rate": 3.415289256198347e-07,
|
|
"logits/chosen": -0.232421875,
|
|
"logits/rejected": -0.392578125,
|
|
"logps/chosen": -432.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.3121,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 1.109375,
|
|
"rewards/margins": 5.5625,
|
|
"rewards/rejected": -4.46875,
|
|
"step": 792
|
|
},
|
|
{
|
|
"epoch": 0.9730061349693252,
|
|
"grad_norm": 39.05298404914802,
|
|
"learning_rate": 3.4132231404958676e-07,
|
|
"logits/chosen": -0.21875,
|
|
"logits/rejected": -0.365234375,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -340.0,
|
|
"loss": 0.3098,
|
|
"rewards/accuracies": 0.859375,
|
|
"rewards/chosen": 0.7109375,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -4.78125,
|
|
"step": 793
|
|
},
|
|
{
|
|
"epoch": 0.9742331288343559,
|
|
"grad_norm": 40.1763340425155,
|
|
"learning_rate": 3.4111570247933886e-07,
|
|
"logits/chosen": -0.169921875,
|
|
"logits/rejected": -0.30859375,
|
|
"logps/chosen": -348.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.3445,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 0.3046875,
|
|
"rewards/margins": 5.3125,
|
|
"rewards/rejected": -5.0,
|
|
"step": 794
|
|
},
|
|
{
|
|
"epoch": 0.9754601226993865,
|
|
"grad_norm": 31.061860595685847,
|
|
"learning_rate": 3.4090909090909085e-07,
|
|
"logits/chosen": -0.3359375,
|
|
"logits/rejected": -0.515625,
|
|
"logps/chosen": -384.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.2126,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 1.65625,
|
|
"rewards/margins": 7.59375,
|
|
"rewards/rejected": -5.9375,
|
|
"step": 795
|
|
},
|
|
{
|
|
"epoch": 0.9766871165644172,
|
|
"grad_norm": 43.66859590907784,
|
|
"learning_rate": 3.4070247933884295e-07,
|
|
"logits/chosen": -0.244140625,
|
|
"logits/rejected": -0.447265625,
|
|
"logps/chosen": -400.0,
|
|
"logps/rejected": -312.0,
|
|
"loss": 0.2891,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.0830078125,
|
|
"rewards/margins": 5.5,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 796
|
|
},
|
|
{
|
|
"epoch": 0.9779141104294479,
|
|
"grad_norm": 35.97056431665963,
|
|
"learning_rate": 3.40495867768595e-07,
|
|
"logits/chosen": -0.1767578125,
|
|
"logits/rejected": -0.357421875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -300.0,
|
|
"loss": 0.2689,
|
|
"rewards/accuracies": 0.84375,
|
|
"rewards/chosen": 0.8359375,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -4.84375,
|
|
"step": 797
|
|
},
|
|
{
|
|
"epoch": 0.9791411042944785,
|
|
"grad_norm": 34.63236728191056,
|
|
"learning_rate": 3.402892561983471e-07,
|
|
"logits/chosen": -0.2099609375,
|
|
"logits/rejected": -0.388671875,
|
|
"logps/chosen": -424.0,
|
|
"logps/rejected": -354.0,
|
|
"loss": 0.2477,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 0.91015625,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.25,
|
|
"step": 798
|
|
},
|
|
{
|
|
"epoch": 0.9803680981595092,
|
|
"grad_norm": 40.12810675669665,
|
|
"learning_rate": 3.4008264462809914e-07,
|
|
"logits/chosen": -0.2470703125,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -392.0,
|
|
"logps/rejected": -336.0,
|
|
"loss": 0.3047,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 1.125,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -4.9375,
|
|
"step": 799
|
|
},
|
|
{
|
|
"epoch": 0.9815950920245399,
|
|
"grad_norm": 36.02461663266261,
|
|
"learning_rate": 3.3987603305785124e-07,
|
|
"logits/chosen": -0.2255859375,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2414,
|
|
"rewards/accuracies": 0.890625,
|
|
"rewards/chosen": 0.7265625,
|
|
"rewards/margins": 5.96875,
|
|
"rewards/rejected": -5.25,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.9828220858895705,
|
|
"grad_norm": 42.912325824023604,
|
|
"learning_rate": 3.396694214876033e-07,
|
|
"logits/chosen": -0.2333984375,
|
|
"logits/rejected": -0.39453125,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3447,
|
|
"rewards/accuracies": 0.7890625,
|
|
"rewards/chosen": 0.82421875,
|
|
"rewards/margins": 5.375,
|
|
"rewards/rejected": -4.5625,
|
|
"step": 801
|
|
},
|
|
{
|
|
"epoch": 0.9840490797546012,
|
|
"grad_norm": 43.08820301261888,
|
|
"learning_rate": 3.394628099173554e-07,
|
|
"logits/chosen": -0.28125,
|
|
"logits/rejected": -0.42578125,
|
|
"logps/chosen": -380.0,
|
|
"logps/rejected": -328.0,
|
|
"loss": 0.3234,
|
|
"rewards/accuracies": 0.8203125,
|
|
"rewards/chosen": 0.58203125,
|
|
"rewards/margins": 5.6875,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 802
|
|
},
|
|
{
|
|
"epoch": 0.9852760736196319,
|
|
"grad_norm": 32.00066785635946,
|
|
"learning_rate": 3.3925619834710744e-07,
|
|
"logits/chosen": -0.333984375,
|
|
"logits/rejected": -0.5078125,
|
|
"logps/chosen": -390.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2588,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -4.90625,
|
|
"step": 803
|
|
},
|
|
{
|
|
"epoch": 0.9865030674846625,
|
|
"grad_norm": 34.73198448187894,
|
|
"learning_rate": 3.390495867768595e-07,
|
|
"logits/chosen": -0.1806640625,
|
|
"logits/rejected": -0.3046875,
|
|
"logps/chosen": -352.0,
|
|
"logps/rejected": -320.0,
|
|
"loss": 0.3089,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.357421875,
|
|
"rewards/margins": 5.09375,
|
|
"rewards/rejected": -4.71875,
|
|
"step": 804
|
|
},
|
|
{
|
|
"epoch": 0.9877300613496932,
|
|
"grad_norm": 40.00296686612965,
|
|
"learning_rate": 3.3884297520661153e-07,
|
|
"logits/chosen": -0.2734375,
|
|
"logits/rejected": -0.421875,
|
|
"logps/chosen": -426.0,
|
|
"logps/rejected": -372.0,
|
|
"loss": 0.3066,
|
|
"rewards/accuracies": 0.8515625,
|
|
"rewards/chosen": 1.078125,
|
|
"rewards/margins": 6.15625,
|
|
"rewards/rejected": -5.09375,
|
|
"step": 805
|
|
},
|
|
{
|
|
"epoch": 0.9889570552147239,
|
|
"grad_norm": 40.373803134623145,
|
|
"learning_rate": 3.3863636363636363e-07,
|
|
"logits/chosen": -0.07470703125,
|
|
"logits/rejected": -0.2119140625,
|
|
"logps/chosen": -408.0,
|
|
"logps/rejected": -360.0,
|
|
"loss": 0.3327,
|
|
"rewards/accuracies": 0.796875,
|
|
"rewards/chosen": 0.79296875,
|
|
"rewards/margins": 5.34375,
|
|
"rewards/rejected": -4.53125,
|
|
"step": 806
|
|
},
|
|
{
|
|
"epoch": 0.9901840490797545,
|
|
"grad_norm": 38.793240604287554,
|
|
"learning_rate": 3.384297520661157e-07,
|
|
"logits/chosen": -0.162109375,
|
|
"logits/rejected": -0.32421875,
|
|
"logps/chosen": -394.0,
|
|
"logps/rejected": -324.0,
|
|
"loss": 0.2214,
|
|
"rewards/accuracies": 0.8828125,
|
|
"rewards/chosen": 1.140625,
|
|
"rewards/margins": 6.34375,
|
|
"rewards/rejected": -5.1875,
|
|
"step": 807
|
|
},
|
|
{
|
|
"epoch": 0.9914110429447853,
|
|
"grad_norm": 34.36639607943694,
|
|
"learning_rate": 3.3822314049586773e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.3125,
|
|
"logps/chosen": -354.0,
|
|
"logps/rejected": -326.0,
|
|
"loss": 0.2778,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 0.34375,
|
|
"rewards/margins": 5.84375,
|
|
"rewards/rejected": -5.5,
|
|
"step": 808
|
|
},
|
|
{
|
|
"epoch": 0.992638036809816,
|
|
"grad_norm": 35.83788708559064,
|
|
"learning_rate": 3.3801652892561983e-07,
|
|
"logits/chosen": -0.205078125,
|
|
"logits/rejected": -0.43359375,
|
|
"logps/chosen": -428.0,
|
|
"logps/rejected": -334.0,
|
|
"loss": 0.2316,
|
|
"rewards/accuracies": 0.8671875,
|
|
"rewards/chosen": 1.6875,
|
|
"rewards/margins": 6.84375,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 809
|
|
},
|
|
{
|
|
"epoch": 0.9938650306748467,
|
|
"grad_norm": 33.992147771254764,
|
|
"learning_rate": 3.378099173553719e-07,
|
|
"logits/chosen": -0.30078125,
|
|
"logits/rejected": -0.486328125,
|
|
"logps/chosen": -436.0,
|
|
"logps/rejected": -358.0,
|
|
"loss": 0.232,
|
|
"rewards/accuracies": 0.875,
|
|
"rewards/chosen": 1.34375,
|
|
"rewards/margins": 6.40625,
|
|
"rewards/rejected": -5.0625,
|
|
"step": 810
|
|
},
|
|
{
|
|
"epoch": 0.9950920245398773,
|
|
"grad_norm": 33.283496414356925,
|
|
"learning_rate": 3.37603305785124e-07,
|
|
"logits/chosen": -0.291015625,
|
|
"logits/rejected": -0.44921875,
|
|
"logps/chosen": -386.0,
|
|
"logps/rejected": -356.0,
|
|
"loss": 0.2725,
|
|
"rewards/accuracies": 0.8359375,
|
|
"rewards/chosen": 1.2578125,
|
|
"rewards/margins": 5.875,
|
|
"rewards/rejected": -4.625,
|
|
"step": 811
|
|
},
|
|
{
|
|
"epoch": 0.996319018404908,
|
|
"grad_norm": 37.39709126950234,
|
|
"learning_rate": 3.37396694214876e-07,
|
|
"logits/chosen": -0.271484375,
|
|
"logits/rejected": -0.359375,
|
|
"logps/chosen": -356.0,
|
|
"logps/rejected": -316.0,
|
|
"loss": 0.2424,
|
|
"rewards/accuracies": 0.921875,
|
|
"rewards/chosen": 0.65625,
|
|
"rewards/margins": 6.09375,
|
|
"rewards/rejected": -5.4375,
|
|
"step": 812
|
|
},
|
|
{
|
|
"epoch": 0.9975460122699387,
|
|
"grad_norm": 33.804195099429606,
|
|
"learning_rate": 3.3719008264462807e-07,
|
|
"logits/chosen": -0.1953125,
|
|
"logits/rejected": -0.341796875,
|
|
"logps/chosen": -336.0,
|
|
"logps/rejected": -314.0,
|
|
"loss": 0.1865,
|
|
"rewards/accuracies": 0.90625,
|
|
"rewards/chosen": 0.9140625,
|
|
"rewards/margins": 6.0625,
|
|
"rewards/rejected": -5.15625,
|
|
"step": 813
|
|
},
|
|
{
|
|
"epoch": 0.9987730061349693,
|
|
"grad_norm": 38.231813987979244,
|
|
"learning_rate": 3.369834710743801e-07,
|
|
"logits/chosen": -0.1884765625,
|
|
"logits/rejected": -0.2734375,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -346.0,
|
|
"loss": 0.3288,
|
|
"rewards/accuracies": 0.8125,
|
|
"rewards/chosen": 0.162109375,
|
|
"rewards/margins": 4.65625,
|
|
"rewards/rejected": -4.5,
|
|
"step": 814
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"grad_norm": 37.3431869316195,
|
|
"learning_rate": 3.367768595041322e-07,
|
|
"logits/chosen": -0.11572265625,
|
|
"logits/rejected": -0.306640625,
|
|
"logps/chosen": -378.0,
|
|
"logps/rejected": -338.0,
|
|
"loss": 0.203,
|
|
"rewards/accuracies": 0.8571429252624512,
|
|
"rewards/chosen": 0.859375,
|
|
"rewards/margins": 5.9375,
|
|
"rewards/rejected": -5.0625,
|
|
"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
|
|
}
|