| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 542, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0018450184501845018, | |
| "grad_norm": 246.95624117590273, | |
| "learning_rate": 9.09090909090909e-09, | |
| "logits/chosen": 0.291015625, | |
| "logits/rejected": 0.058349609375, | |
| "logps/chosen": -532.0, | |
| "logps/rejected": -424.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01845018450184502, | |
| "grad_norm": 261.1507355895937, | |
| "learning_rate": 9.09090909090909e-08, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -1.03125, | |
| "logps/chosen": -490.0, | |
| "logps/rejected": -356.0, | |
| "loss": 0.6782, | |
| "rewards/accuracies": 0.3333333432674408, | |
| "rewards/chosen": 0.050048828125, | |
| "rewards/margins": -0.00543212890625, | |
| "rewards/rejected": 0.0556640625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03690036900369004, | |
| "grad_norm": 214.53455363826293, | |
| "learning_rate": 1.818181818181818e-07, | |
| "logits/chosen": -1.203125, | |
| "logits/rejected": -1.15625, | |
| "logps/chosen": -480.0, | |
| "logps/rejected": -372.0, | |
| "loss": 0.5618, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.828125, | |
| "rewards/margins": 0.48046875, | |
| "rewards/rejected": 0.349609375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.055350553505535055, | |
| "grad_norm": 85.91357219148367, | |
| "learning_rate": 2.727272727272727e-07, | |
| "logits/chosen": -0.55078125, | |
| "logits/rejected": -0.95703125, | |
| "logps/chosen": -492.0, | |
| "logps/rejected": -368.0, | |
| "loss": 0.4012, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.8828125, | |
| "rewards/margins": 1.171875, | |
| "rewards/rejected": 0.7109375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.07380073800738007, | |
| "grad_norm": 48.60414747910648, | |
| "learning_rate": 3.636363636363636e-07, | |
| "logits/chosen": -0.94140625, | |
| "logits/rejected": -0.9296875, | |
| "logps/chosen": -456.0, | |
| "logps/rejected": -362.0, | |
| "loss": 0.2995, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 2.703125, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": 0.6171875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.09225092250922509, | |
| "grad_norm": 268.4596876602969, | |
| "learning_rate": 4.545454545454545e-07, | |
| "logits/chosen": -0.435546875, | |
| "logits/rejected": -0.49609375, | |
| "logps/chosen": -736.0, | |
| "logps/rejected": -378.0, | |
| "loss": 0.2394, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.09375, | |
| "rewards/margins": 3.859375, | |
| "rewards/rejected": 1.2421875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.11070110701107011, | |
| "grad_norm": 28.519035791504386, | |
| "learning_rate": 4.948665297741273e-07, | |
| "logits/chosen": -0.5625, | |
| "logits/rejected": -0.55859375, | |
| "logps/chosen": -512.0, | |
| "logps/rejected": -430.0, | |
| "loss": 0.1314, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 4.59375, | |
| "rewards/margins": 3.8125, | |
| "rewards/rejected": 0.79296875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.12915129151291513, | |
| "grad_norm": 2.758408343392576, | |
| "learning_rate": 4.845995893223819e-07, | |
| "logits/chosen": -0.3671875, | |
| "logits/rejected": -0.494140625, | |
| "logps/chosen": -408.0, | |
| "logps/rejected": -326.0, | |
| "loss": 0.1174, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 4.90625, | |
| "rewards/margins": 4.96875, | |
| "rewards/rejected": -0.06494140625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.14760147601476015, | |
| "grad_norm": 123.36663605539277, | |
| "learning_rate": 4.743326488706365e-07, | |
| "logits/chosen": -0.419921875, | |
| "logits/rejected": -0.59375, | |
| "logps/chosen": -490.0, | |
| "logps/rejected": -406.0, | |
| "loss": 0.1084, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 5.125, | |
| "rewards/margins": 5.75, | |
| "rewards/rejected": -0.609375, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.16605166051660517, | |
| "grad_norm": 52.04622111574895, | |
| "learning_rate": 4.6406570841889114e-07, | |
| "logits/chosen": -0.3359375, | |
| "logits/rejected": -0.267578125, | |
| "logps/chosen": -494.0, | |
| "logps/rejected": -516.0, | |
| "loss": 0.1005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.78125, | |
| "rewards/margins": 6.46875, | |
| "rewards/rejected": -0.66015625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.18450184501845018, | |
| "grad_norm": 7.266800454502919, | |
| "learning_rate": 4.537987679671458e-07, | |
| "logits/chosen": -0.1279296875, | |
| "logits/rejected": -0.0986328125, | |
| "logps/chosen": -528.0, | |
| "logps/rejected": -422.0, | |
| "loss": 0.0206, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.8125, | |
| "rewards/margins": 7.5625, | |
| "rewards/rejected": -0.73828125, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2029520295202952, | |
| "grad_norm": 17.639923775576452, | |
| "learning_rate": 4.435318275154004e-07, | |
| "logits/chosen": -0.017822265625, | |
| "logits/rejected": -0.08447265625, | |
| "logps/chosen": -360.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.1239, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.71875, | |
| "rewards/margins": 8.375, | |
| "rewards/rejected": -1.6484375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.22140221402214022, | |
| "grad_norm": 100.4656637665847, | |
| "learning_rate": 4.33264887063655e-07, | |
| "logits/chosen": -0.232421875, | |
| "logits/rejected": -0.58203125, | |
| "logps/chosen": -476.0, | |
| "logps/rejected": -548.0, | |
| "loss": 0.0276, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.875, | |
| "rewards/margins": 8.0625, | |
| "rewards/rejected": -2.15625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.23985239852398524, | |
| "grad_norm": 2.5639802736905364, | |
| "learning_rate": 4.2299794661190965e-07, | |
| "logits/chosen": -0.019775390625, | |
| "logits/rejected": -0.03662109375, | |
| "logps/chosen": -524.0, | |
| "logps/rejected": -418.0, | |
| "loss": 0.0364, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.59375, | |
| "rewards/margins": 9.25, | |
| "rewards/rejected": -2.640625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.25830258302583026, | |
| "grad_norm": 2.322522622647086, | |
| "learning_rate": 4.1273100616016426e-07, | |
| "logits/chosen": 0.0277099609375, | |
| "logits/rejected": -0.2041015625, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -450.0, | |
| "loss": 0.0647, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 6.0, | |
| "rewards/margins": 8.25, | |
| "rewards/rejected": -2.265625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.2767527675276753, | |
| "grad_norm": 16.567678257558736, | |
| "learning_rate": 4.024640657084189e-07, | |
| "logits/chosen": 0.314453125, | |
| "logits/rejected": -0.0361328125, | |
| "logps/chosen": -422.0, | |
| "logps/rejected": -476.0, | |
| "loss": 0.0139, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 5.625, | |
| "rewards/margins": 7.5, | |
| "rewards/rejected": -1.859375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2952029520295203, | |
| "grad_norm": 156.39101191132625, | |
| "learning_rate": 3.9219712525667346e-07, | |
| "logits/chosen": 0.2734375, | |
| "logits/rejected": 0.031494140625, | |
| "logps/chosen": -422.0, | |
| "logps/rejected": -504.0, | |
| "loss": 0.0334, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 7.65625, | |
| "rewards/margins": 8.5625, | |
| "rewards/rejected": -0.859375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.31365313653136534, | |
| "grad_norm": 1.346956269793295, | |
| "learning_rate": 3.819301848049281e-07, | |
| "logits/chosen": 0.263671875, | |
| "logits/rejected": 0.07568359375, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -516.0, | |
| "loss": 0.0219, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.03125, | |
| "rewards/margins": 9.5625, | |
| "rewards/rejected": -2.5, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.33210332103321033, | |
| "grad_norm": 0.08324608481285954, | |
| "learning_rate": 3.7166324435318277e-07, | |
| "logits/chosen": 0.337890625, | |
| "logits/rejected": 0.259765625, | |
| "logps/chosen": -434.0, | |
| "logps/rejected": -528.0, | |
| "loss": 0.0257, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.0, | |
| "rewards/margins": 10.8125, | |
| "rewards/rejected": -2.8125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3505535055350554, | |
| "grad_norm": 2.8470520017552157, | |
| "learning_rate": 3.6139630390143737e-07, | |
| "logits/chosen": -0.068359375, | |
| "logits/rejected": 0.1044921875, | |
| "logps/chosen": -552.0, | |
| "logps/rejected": -458.0, | |
| "loss": 0.019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.0625, | |
| "rewards/margins": 9.8125, | |
| "rewards/rejected": -0.796875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.36900369003690037, | |
| "grad_norm": 0.33788257296911867, | |
| "learning_rate": 3.51129363449692e-07, | |
| "logits/chosen": 0.294921875, | |
| "logits/rejected": 0.0186767578125, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -412.0, | |
| "loss": 0.0248, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.21875, | |
| "rewards/margins": 9.3125, | |
| "rewards/rejected": -2.125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.3874538745387454, | |
| "grad_norm": 0.041944694451615326, | |
| "learning_rate": 3.408624229979466e-07, | |
| "logits/chosen": -0.00830078125, | |
| "logits/rejected": -0.2294921875, | |
| "logps/chosen": -700.0, | |
| "logps/rejected": -556.0, | |
| "loss": 0.0028, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.3125, | |
| "rewards/margins": 10.8125, | |
| "rewards/rejected": -2.484375, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4059040590405904, | |
| "grad_norm": 0.413539769640826, | |
| "learning_rate": 3.3059548254620123e-07, | |
| "logits/chosen": 0.10205078125, | |
| "logits/rejected": -0.2158203125, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -422.0, | |
| "loss": 0.0332, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 7.03125, | |
| "rewards/margins": 9.1875, | |
| "rewards/rejected": -2.140625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.42435424354243545, | |
| "grad_norm": 16.188053914068192, | |
| "learning_rate": 3.2032854209445584e-07, | |
| "logits/chosen": 0.267578125, | |
| "logits/rejected": 0.03466796875, | |
| "logps/chosen": -420.0, | |
| "logps/rejected": -484.0, | |
| "loss": 0.0047, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.3125, | |
| "rewards/margins": 12.375, | |
| "rewards/rejected": -3.03125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.44280442804428044, | |
| "grad_norm": 98.96986455098327, | |
| "learning_rate": 3.1006160164271044e-07, | |
| "logits/chosen": 0.2158203125, | |
| "logits/rejected": 0.267578125, | |
| "logps/chosen": -548.0, | |
| "logps/rejected": -504.0, | |
| "loss": 0.013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.03125, | |
| "rewards/margins": 10.4375, | |
| "rewards/rejected": -3.390625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.4612546125461255, | |
| "grad_norm": 1.4122764971682638, | |
| "learning_rate": 2.997946611909651e-07, | |
| "logits/chosen": 0.5234375, | |
| "logits/rejected": 0.302734375, | |
| "logps/chosen": -540.0, | |
| "logps/rejected": -588.0, | |
| "loss": 0.0284, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.0625, | |
| "rewards/margins": 14.3125, | |
| "rewards/rejected": -6.28125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.4797047970479705, | |
| "grad_norm": 2.39173502123089, | |
| "learning_rate": 2.895277207392197e-07, | |
| "logits/chosen": 0.0294189453125, | |
| "logits/rejected": 0.26953125, | |
| "logps/chosen": -420.0, | |
| "logps/rejected": -532.0, | |
| "loss": 0.0051, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.53125, | |
| "rewards/margins": 12.6875, | |
| "rewards/rejected": -5.1875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.4981549815498155, | |
| "grad_norm": 2.2168536993734294, | |
| "learning_rate": 2.792607802874743e-07, | |
| "logits/chosen": 0.142578125, | |
| "logits/rejected": 0.0908203125, | |
| "logps/chosen": -468.0, | |
| "logps/rejected": -444.0, | |
| "loss": 0.0395, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.1875, | |
| "rewards/margins": 12.625, | |
| "rewards/rejected": -4.4375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5166051660516605, | |
| "grad_norm": 0.037373864558181605, | |
| "learning_rate": 2.6899383983572895e-07, | |
| "logits/chosen": 0.326171875, | |
| "logits/rejected": 0.3046875, | |
| "logps/chosen": -484.0, | |
| "logps/rejected": -510.0, | |
| "loss": 0.0028, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.3125, | |
| "rewards/margins": 13.875, | |
| "rewards/rejected": -4.59375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.5350553505535055, | |
| "grad_norm": 6.827154453731148, | |
| "learning_rate": 2.5872689938398356e-07, | |
| "logits/chosen": 0.33984375, | |
| "logits/rejected": 0.2021484375, | |
| "logps/chosen": -462.0, | |
| "logps/rejected": -516.0, | |
| "loss": 0.0181, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.9375, | |
| "rewards/margins": 12.125, | |
| "rewards/rejected": -4.25, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.5535055350553506, | |
| "grad_norm": 0.14748708554887247, | |
| "learning_rate": 2.4845995893223816e-07, | |
| "logits/chosen": 0.345703125, | |
| "logits/rejected": 0.173828125, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -504.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.375, | |
| "rewards/margins": 12.125, | |
| "rewards/rejected": -3.75, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5719557195571956, | |
| "grad_norm": 0.231487702191985, | |
| "learning_rate": 2.381930184804928e-07, | |
| "logits/chosen": 0.0216064453125, | |
| "logits/rejected": 0.11865234375, | |
| "logps/chosen": -480.0, | |
| "logps/rejected": -426.0, | |
| "loss": 0.0308, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 8.0625, | |
| "rewards/margins": 12.25, | |
| "rewards/rejected": -4.15625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.5904059040590406, | |
| "grad_norm": 0.5592057504990625, | |
| "learning_rate": 2.2792607802874742e-07, | |
| "logits/chosen": 0.4296875, | |
| "logits/rejected": 0.1904296875, | |
| "logps/chosen": -444.0, | |
| "logps/rejected": -516.0, | |
| "loss": 0.0081, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.5625, | |
| "rewards/margins": 11.3125, | |
| "rewards/rejected": -3.765625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6088560885608856, | |
| "grad_norm": 0.010255339682818812, | |
| "learning_rate": 2.1765913757700204e-07, | |
| "logits/chosen": 0.24609375, | |
| "logits/rejected": 0.349609375, | |
| "logps/chosen": -440.0, | |
| "logps/rejected": -402.0, | |
| "loss": 0.0444, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.6875, | |
| "rewards/margins": 13.25, | |
| "rewards/rejected": -6.53125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.6273062730627307, | |
| "grad_norm": 0.15764226293605477, | |
| "learning_rate": 2.0739219712525665e-07, | |
| "logits/chosen": 0.7109375, | |
| "logits/rejected": 0.578125, | |
| "logps/chosen": -516.0, | |
| "logps/rejected": -496.0, | |
| "loss": 0.018, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 8.125, | |
| "rewards/margins": 14.75, | |
| "rewards/rejected": -6.6875, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.6457564575645757, | |
| "grad_norm": 0.48399973225732595, | |
| "learning_rate": 1.971252566735113e-07, | |
| "logits/chosen": 0.423828125, | |
| "logits/rejected": 0.3125, | |
| "logps/chosen": -632.0, | |
| "logps/rejected": -552.0, | |
| "loss": 0.0187, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.875, | |
| "rewards/margins": 15.125, | |
| "rewards/rejected": -6.3125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.6642066420664207, | |
| "grad_norm": 38.899685751981636, | |
| "learning_rate": 1.868583162217659e-07, | |
| "logits/chosen": 0.396484375, | |
| "logits/rejected": 0.330078125, | |
| "logps/chosen": -458.0, | |
| "logps/rejected": -404.0, | |
| "loss": 0.0089, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.5625, | |
| "rewards/margins": 11.0, | |
| "rewards/rejected": -4.4375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.6826568265682657, | |
| "grad_norm": 0.14172083131050536, | |
| "learning_rate": 1.7659137577002053e-07, | |
| "logits/chosen": 0.578125, | |
| "logits/rejected": 0.3125, | |
| "logps/chosen": -548.0, | |
| "logps/rejected": -532.0, | |
| "loss": 0.0081, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 10.25, | |
| "rewards/margins": 16.125, | |
| "rewards/rejected": -5.875, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7011070110701108, | |
| "grad_norm": 1.3673771780784945, | |
| "learning_rate": 1.6632443531827514e-07, | |
| "logits/chosen": 0.314453125, | |
| "logits/rejected": 0.2412109375, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -504.0, | |
| "loss": 0.024, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.1875, | |
| "rewards/margins": 12.75, | |
| "rewards/rejected": -3.5625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.7195571955719557, | |
| "grad_norm": 0.0008789882379138907, | |
| "learning_rate": 1.5605749486652976e-07, | |
| "logits/chosen": 0.150390625, | |
| "logits/rejected": 0.216796875, | |
| "logps/chosen": -416.0, | |
| "logps/rejected": -584.0, | |
| "loss": 0.0028, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 6.03125, | |
| "rewards/margins": 16.625, | |
| "rewards/rejected": -10.625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.7380073800738007, | |
| "grad_norm": 3.326884714941732, | |
| "learning_rate": 1.457905544147844e-07, | |
| "logits/chosen": 0.322265625, | |
| "logits/rejected": 0.451171875, | |
| "logps/chosen": -510.0, | |
| "logps/rejected": -508.0, | |
| "loss": 0.0576, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.125, | |
| "rewards/margins": 17.375, | |
| "rewards/rejected": -8.1875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.7564575645756457, | |
| "grad_norm": 0.5212668444634995, | |
| "learning_rate": 1.3552361396303902e-07, | |
| "logits/chosen": 0.7734375, | |
| "logits/rejected": 0.5546875, | |
| "logps/chosen": -426.0, | |
| "logps/rejected": -372.0, | |
| "loss": 0.009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.8125, | |
| "rewards/margins": 13.0, | |
| "rewards/rejected": -5.1875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.7749077490774908, | |
| "grad_norm": 0.00044257865962132295, | |
| "learning_rate": 1.2525667351129362e-07, | |
| "logits/chosen": 0.13671875, | |
| "logits/rejected": 0.064453125, | |
| "logps/chosen": -600.0, | |
| "logps/rejected": -444.0, | |
| "loss": 0.0188, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.0625, | |
| "rewards/margins": 13.4375, | |
| "rewards/rejected": -5.34375, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.7933579335793358, | |
| "grad_norm": 11.997474446041776, | |
| "learning_rate": 1.1498973305954825e-07, | |
| "logits/chosen": 0.318359375, | |
| "logits/rejected": 0.322265625, | |
| "logps/chosen": -556.0, | |
| "logps/rejected": -400.0, | |
| "loss": 0.0023, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.0, | |
| "rewards/margins": 14.3125, | |
| "rewards/rejected": -5.3125, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.8118081180811808, | |
| "grad_norm": 5.233979144042266, | |
| "learning_rate": 1.0472279260780287e-07, | |
| "logits/chosen": 0.35546875, | |
| "logits/rejected": 0.384765625, | |
| "logps/chosen": -508.0, | |
| "logps/rejected": -512.0, | |
| "loss": 0.0072, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.3125, | |
| "rewards/margins": 14.4375, | |
| "rewards/rejected": -6.125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.8302583025830258, | |
| "grad_norm": 0.0019388799655296358, | |
| "learning_rate": 9.445585215605748e-08, | |
| "logits/chosen": 0.412109375, | |
| "logits/rejected": 0.4140625, | |
| "logps/chosen": -490.0, | |
| "logps/rejected": -402.0, | |
| "loss": 0.0082, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.53125, | |
| "rewards/margins": 13.0, | |
| "rewards/rejected": -5.46875, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.8487084870848709, | |
| "grad_norm": 0.0037524190968933756, | |
| "learning_rate": 8.418891170431211e-08, | |
| "logits/chosen": 0.64453125, | |
| "logits/rejected": 0.47265625, | |
| "logps/chosen": -568.0, | |
| "logps/rejected": -424.0, | |
| "loss": 0.0029, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 10.5, | |
| "rewards/margins": 17.125, | |
| "rewards/rejected": -6.65625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.8671586715867159, | |
| "grad_norm": 0.23207842387009886, | |
| "learning_rate": 7.392197125256673e-08, | |
| "logits/chosen": 0.18359375, | |
| "logits/rejected": 0.2333984375, | |
| "logps/chosen": -458.0, | |
| "logps/rejected": -528.0, | |
| "loss": 0.0091, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.25, | |
| "rewards/margins": 16.5, | |
| "rewards/rejected": -8.1875, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.8856088560885609, | |
| "grad_norm": 4.488121519968395, | |
| "learning_rate": 6.365503080082136e-08, | |
| "logits/chosen": 0.72265625, | |
| "logits/rejected": 0.369140625, | |
| "logps/chosen": -366.0, | |
| "logps/rejected": -492.0, | |
| "loss": 0.0183, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 8.1875, | |
| "rewards/margins": 16.25, | |
| "rewards/rejected": -8.0625, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.9040590405904059, | |
| "grad_norm": 4.003085746637136, | |
| "learning_rate": 5.338809034907597e-08, | |
| "logits/chosen": 0.42578125, | |
| "logits/rejected": 0.310546875, | |
| "logps/chosen": -536.0, | |
| "logps/rejected": -392.0, | |
| "loss": 0.0036, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.0, | |
| "rewards/margins": 14.0, | |
| "rewards/rejected": -4.96875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.922509225092251, | |
| "grad_norm": 0.003980221494967455, | |
| "learning_rate": 4.3121149897330595e-08, | |
| "logits/chosen": 0.396484375, | |
| "logits/rejected": 0.30078125, | |
| "logps/chosen": -592.0, | |
| "logps/rejected": -612.0, | |
| "loss": 0.0031, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.15625, | |
| "rewards/margins": 13.125, | |
| "rewards/rejected": -6.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.940959409594096, | |
| "grad_norm": 54.57112101614215, | |
| "learning_rate": 3.285420944558522e-08, | |
| "logits/chosen": 0.345703125, | |
| "logits/rejected": 0.419921875, | |
| "logps/chosen": -438.0, | |
| "logps/rejected": -540.0, | |
| "loss": 0.0349, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.28125, | |
| "rewards/margins": 16.375, | |
| "rewards/rejected": -9.125, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.959409594095941, | |
| "grad_norm": 4.639121263815892, | |
| "learning_rate": 2.2587268993839836e-08, | |
| "logits/chosen": 0.390625, | |
| "logits/rejected": 0.1826171875, | |
| "logps/chosen": -436.0, | |
| "logps/rejected": -632.0, | |
| "loss": 0.0075, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.5, | |
| "rewards/margins": 16.0, | |
| "rewards/rejected": -8.4375, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.977859778597786, | |
| "grad_norm": 0.1326299403468037, | |
| "learning_rate": 1.2320328542094456e-08, | |
| "logits/chosen": 0.26171875, | |
| "logits/rejected": 0.3359375, | |
| "logps/chosen": -422.0, | |
| "logps/rejected": -394.0, | |
| "loss": 0.0043, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 9.1875, | |
| "rewards/margins": 15.5, | |
| "rewards/rejected": -6.28125, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.996309963099631, | |
| "grad_norm": 0.03353244661943473, | |
| "learning_rate": 2.053388090349076e-09, | |
| "logits/chosen": 0.61328125, | |
| "logits/rejected": 0.39453125, | |
| "logps/chosen": -426.0, | |
| "logps/rejected": -484.0, | |
| "loss": 0.0247, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 7.0, | |
| "rewards/margins": 13.5, | |
| "rewards/rejected": -6.5, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": 0.7109375, | |
| "eval_logits/rejected": 0.55859375, | |
| "eval_logps/chosen": -376.0, | |
| "eval_logps/rejected": -412.0, | |
| "eval_loss": 0.012509898282587528, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": 7.90625, | |
| "eval_rewards/margins": 15.4375, | |
| "eval_rewards/rejected": -7.5, | |
| "eval_runtime": 7.1353, | |
| "eval_samples_per_second": 14.015, | |
| "eval_steps_per_second": 1.822, | |
| "step": 542 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 542, | |
| "total_flos": 0.0, | |
| "train_loss": 0.06598214372800065, | |
| "train_runtime": 921.9288, | |
| "train_samples_per_second": 4.699, | |
| "train_steps_per_second": 0.588 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 542, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |