| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 939, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003194888178913738, | |
| "grad_norm": 10.744781344666563, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -90.0, | |
| "logps/rejected": -92.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03194888178913738, | |
| "grad_norm": 9.082097631913506, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -1.265625, | |
| "logits/rejected": -1.3046875, | |
| "logps/chosen": -95.5, | |
| "logps/rejected": -109.0, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.0045166015625, | |
| "rewards/margins": -0.00138092041015625, | |
| "rewards/rejected": -0.0031280517578125, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 11.293929203635875, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -1.328125, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -115.0, | |
| "logps/rejected": -124.5, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.012500000186264515, | |
| "rewards/chosen": -0.0015716552734375, | |
| "rewards/margins": -0.00093841552734375, | |
| "rewards/rejected": -0.000629425048828125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 11.500151336259389, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -110.0, | |
| "logps/rejected": -114.5, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.02783203125, | |
| "rewards/margins": -0.0021820068359375, | |
| "rewards/rejected": 0.030029296875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 6.587341461970503, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -116.5, | |
| "logps/rejected": -114.0, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.11249999701976776, | |
| "rewards/chosen": 0.0341796875, | |
| "rewards/margins": -0.0028076171875, | |
| "rewards/rejected": 0.036865234375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 13.242551025742099, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -1.3046875, | |
| "logits/rejected": -1.328125, | |
| "logps/chosen": -99.0, | |
| "logps/rejected": -111.5, | |
| "loss": 0.6897, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": 0.03662109375, | |
| "rewards/margins": 0.00750732421875, | |
| "rewards/rejected": 0.029052734375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 12.143173292039066, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -1.3203125, | |
| "logits/rejected": -1.3125, | |
| "logps/chosen": -106.0, | |
| "logps/rejected": -106.0, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.12109375, | |
| "rewards/margins": 0.013427734375, | |
| "rewards/rejected": 0.10791015625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 9.586949276923757, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -1.296875, | |
| "logits/rejected": -1.3203125, | |
| "logps/chosen": -96.5, | |
| "logps/rejected": -102.5, | |
| "loss": 0.6809, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": 0.2060546875, | |
| "rewards/margins": 0.0274658203125, | |
| "rewards/rejected": 0.1787109375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 15.537860332105156, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -1.328125, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -110.5, | |
| "logps/rejected": -123.0, | |
| "loss": 0.6772, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.29296875, | |
| "rewards/margins": 0.056640625, | |
| "rewards/rejected": 0.236328125, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 11.737281018212952, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -109.5, | |
| "logps/rejected": -117.5, | |
| "loss": 0.6749, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.251953125, | |
| "rewards/margins": 0.0751953125, | |
| "rewards/rejected": 0.177734375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 12.144764025677807, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -1.34375, | |
| "logits/rejected": -1.328125, | |
| "logps/chosen": -96.5, | |
| "logps/rejected": -97.5, | |
| "loss": 0.6715, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.400390625, | |
| "rewards/margins": 0.047119140625, | |
| "rewards/rejected": 0.35546875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 8.02976983372895, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -1.34375, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -105.5, | |
| "logps/rejected": -113.5, | |
| "loss": 0.6705, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": 0.3515625, | |
| "rewards/margins": 0.11572265625, | |
| "rewards/rejected": 0.2373046875, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 10.631432399065563, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -1.3203125, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -97.5, | |
| "logps/rejected": -101.0, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.5546875, | |
| "rewards/margins": 0.045166015625, | |
| "rewards/rejected": 0.51171875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 13.49563173839027, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -1.296875, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -91.0, | |
| "logps/rejected": -114.0, | |
| "loss": 0.6711, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": 0.61328125, | |
| "rewards/margins": 0.197265625, | |
| "rewards/rejected": 0.416015625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 16.12741791493388, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -1.3515625, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -102.5, | |
| "logps/rejected": -110.5, | |
| "loss": 0.6694, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.380859375, | |
| "rewards/margins": 0.061279296875, | |
| "rewards/rejected": 0.3203125, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 12.019867226636407, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -1.34375, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -100.5, | |
| "logps/rejected": -106.0, | |
| "loss": 0.6588, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.5, | |
| "rewards/margins": 0.1806640625, | |
| "rewards/rejected": 0.3203125, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 9.291388851043736, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -1.34375, | |
| "logits/rejected": -1.3515625, | |
| "logps/chosen": -100.5, | |
| "logps/rejected": -109.0, | |
| "loss": 0.6528, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.65625, | |
| "rewards/margins": 0.1552734375, | |
| "rewards/rejected": 0.5, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 7.795956096986048, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -1.2890625, | |
| "logits/rejected": -1.296875, | |
| "logps/chosen": -86.5, | |
| "logps/rejected": -88.0, | |
| "loss": 0.6332, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.33203125, | |
| "rewards/margins": 0.2021484375, | |
| "rewards/rejected": 0.1298828125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 13.033638428168622, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -1.28125, | |
| "logits/rejected": -1.3125, | |
| "logps/chosen": -95.5, | |
| "logps/rejected": -102.0, | |
| "loss": 0.6414, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.0751953125, | |
| "rewards/margins": 0.30078125, | |
| "rewards/rejected": -0.376953125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 9.06131966560706, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -1.328125, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -100.5, | |
| "logps/rejected": -113.5, | |
| "loss": 0.6389, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -0.1884765625, | |
| "rewards/margins": 0.240234375, | |
| "rewards/rejected": -0.427734375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 10.830447939806625, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -1.328125, | |
| "logits/rejected": -1.3125, | |
| "logps/chosen": -92.5, | |
| "logps/rejected": -99.5, | |
| "loss": 0.6526, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.333984375, | |
| "rewards/margins": 0.234375, | |
| "rewards/rejected": 0.09765625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 25.689913958994083, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -1.3671875, | |
| "logits/rejected": -1.3671875, | |
| "logps/chosen": -97.5, | |
| "logps/rejected": -100.0, | |
| "loss": 0.6613, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.31640625, | |
| "rewards/margins": 0.16796875, | |
| "rewards/rejected": 0.1494140625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 4.168917801938765, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -1.3125, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -114.0, | |
| "loss": 0.6363, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.054931640625, | |
| "rewards/margins": 0.1455078125, | |
| "rewards/rejected": -0.0908203125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 12.774310599728109, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -119.0, | |
| "loss": 0.6926, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.19140625, | |
| "rewards/margins": 0.09033203125, | |
| "rewards/rejected": -0.28125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 9.248819996807187, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -1.3515625, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -98.0, | |
| "logps/rejected": -110.5, | |
| "loss": 0.6155, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.1171875, | |
| "rewards/margins": 0.2177734375, | |
| "rewards/rejected": -0.333984375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 15.168026248041878, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -89.0, | |
| "logps/rejected": -99.0, | |
| "loss": 0.6792, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.73828125, | |
| "rewards/margins": 0.2265625, | |
| "rewards/rejected": 0.51171875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 15.007736942940102, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -104.5, | |
| "logps/rejected": -115.0, | |
| "loss": 0.677, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.6875, | |
| "rewards/margins": 0.158203125, | |
| "rewards/rejected": 0.52734375, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 11.752375867792377, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -1.328125, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -89.0, | |
| "logps/rejected": -93.5, | |
| "loss": 0.6772, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.5546875, | |
| "rewards/margins": 0.09521484375, | |
| "rewards/rejected": 0.45703125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 8.47412587415608, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -1.3515625, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -94.0, | |
| "logps/rejected": -96.5, | |
| "loss": 0.6545, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.6171875, | |
| "rewards/margins": 0.109375, | |
| "rewards/rejected": 0.5078125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 14.053798791185852, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.3515625, | |
| "logps/chosen": -108.0, | |
| "logps/rejected": -116.0, | |
| "loss": 0.678, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.04931640625, | |
| "rewards/margins": 0.15234375, | |
| "rewards/rejected": -0.201171875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 10.194489389821156, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -1.3828125, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -112.0, | |
| "logps/rejected": -117.0, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": -0.1298828125, | |
| "rewards/margins": 0.06005859375, | |
| "rewards/rejected": -0.189453125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 11.274044326460443, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -114.0, | |
| "logps/rejected": -117.0, | |
| "loss": 0.6622, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.0888671875, | |
| "rewards/margins": 0.042724609375, | |
| "rewards/rejected": -0.130859375, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.2734375, | |
| "eval_logits/rejected": -1.2734375, | |
| "eval_logps/chosen": -85.5, | |
| "eval_logps/rejected": -91.0, | |
| "eval_loss": 0.6657812595367432, | |
| "eval_rewards/accuracies": 0.1785714328289032, | |
| "eval_rewards/chosen": 0.1455078125, | |
| "eval_rewards/margins": 0.169921875, | |
| "eval_rewards/rejected": -0.0244140625, | |
| "eval_runtime": 12.1715, | |
| "eval_samples_per_second": 16.432, | |
| "eval_steps_per_second": 0.575, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 9.961349097417994, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -118.5, | |
| "logps/rejected": -121.5, | |
| "loss": 0.6499, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.03125, | |
| "rewards/margins": 0.185546875, | |
| "rewards/rejected": -0.1533203125, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 6.937402354627386, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -1.375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -102.5, | |
| "logps/rejected": -114.0, | |
| "loss": 0.6044, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.1220703125, | |
| "rewards/margins": 0.275390625, | |
| "rewards/rejected": -0.396484375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 8.504289129675827, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -107.5, | |
| "loss": 0.6038, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": 0.181640625, | |
| "rewards/margins": 0.28125, | |
| "rewards/rejected": -0.0986328125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 12.12367798974703, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -1.375, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -100.5, | |
| "logps/rejected": -120.0, | |
| "loss": 0.5815, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.00567626953125, | |
| "rewards/margins": 0.419921875, | |
| "rewards/rejected": -0.423828125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 7.691109026647345, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -102.5, | |
| "logps/rejected": -126.5, | |
| "loss": 0.6036, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.375, | |
| "rewards/margins": 0.341796875, | |
| "rewards/rejected": 0.033203125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 9.407581984350474, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -112.5, | |
| "logps/rejected": -117.5, | |
| "loss": 0.6071, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.2060546875, | |
| "rewards/margins": 0.296875, | |
| "rewards/rejected": -0.09033203125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 7.491889072000949, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -1.3515625, | |
| "logits/rejected": -1.359375, | |
| "logps/chosen": -93.5, | |
| "logps/rejected": -97.5, | |
| "loss": 0.5994, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.1611328125, | |
| "rewards/margins": 0.240234375, | |
| "rewards/rejected": -0.07958984375, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 6.076378075960898, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -105.0, | |
| "logps/rejected": -115.0, | |
| "loss": 0.6212, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.416015625, | |
| "rewards/margins": 0.287109375, | |
| "rewards/rejected": 0.1279296875, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 3.8330692877398334, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -117.0, | |
| "logps/rejected": -123.5, | |
| "loss": 0.5816, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.0067138671875, | |
| "rewards/margins": 0.35546875, | |
| "rewards/rejected": -0.349609375, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 9.190898997997795, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -109.0, | |
| "logps/rejected": -118.0, | |
| "loss": 0.5719, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": 0.095703125, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": -0.455078125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 9.390187857025627, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -109.0, | |
| "logps/rejected": -124.0, | |
| "loss": 0.5756, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.1962890625, | |
| "rewards/margins": 0.5703125, | |
| "rewards/rejected": -0.373046875, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 6.190589789849267, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -97.5, | |
| "logps/rejected": -115.0, | |
| "loss": 0.5975, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": 0.1162109375, | |
| "rewards/margins": 0.51171875, | |
| "rewards/rejected": -0.396484375, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 18.266289420442742, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -105.5, | |
| "logps/rejected": -112.5, | |
| "loss": 0.5727, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.044189453125, | |
| "rewards/margins": 0.404296875, | |
| "rewards/rejected": -0.361328125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 12.220726704381809, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -101.5, | |
| "logps/rejected": -117.5, | |
| "loss": 0.5943, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": 0.4609375, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": -0.07861328125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 9.286364410172531, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -106.5, | |
| "logps/rejected": -109.0, | |
| "loss": 0.5668, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.23828125, | |
| "rewards/margins": 0.287109375, | |
| "rewards/rejected": -0.0478515625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 9.489457520832659, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -1.3828125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -96.0, | |
| "logps/rejected": -110.0, | |
| "loss": 0.5897, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -0.03564453125, | |
| "rewards/margins": 0.6015625, | |
| "rewards/rejected": -0.63671875, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 10.08281709290549, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -102.0, | |
| "logps/rejected": -111.5, | |
| "loss": 0.5867, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.41796875, | |
| "rewards/margins": 0.28125, | |
| "rewards/rejected": -0.69921875, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 11.966493723685009, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -108.5, | |
| "logps/rejected": -116.0, | |
| "loss": 0.6152, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.60546875, | |
| "rewards/margins": 0.42578125, | |
| "rewards/rejected": -1.03125, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 9.087229848717023, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -124.0, | |
| "logps/rejected": -141.0, | |
| "loss": 0.5946, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.271484375, | |
| "rewards/margins": 0.380859375, | |
| "rewards/rejected": -0.65234375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 16.95351537988852, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -117.0, | |
| "logps/rejected": -120.0, | |
| "loss": 0.5911, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.035888671875, | |
| "rewards/margins": 0.33984375, | |
| "rewards/rejected": -0.375, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 14.925346317485197, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -93.5, | |
| "logps/rejected": -108.5, | |
| "loss": 0.5565, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": 0.0908203125, | |
| "rewards/margins": 0.66796875, | |
| "rewards/rejected": -0.57421875, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 17.773561663409936, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -104.5, | |
| "logps/rejected": -110.5, | |
| "loss": 0.5835, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.0194091796875, | |
| "rewards/margins": 0.38671875, | |
| "rewards/rejected": -0.3671875, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 8.539790110888523, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -108.5, | |
| "logps/rejected": -119.5, | |
| "loss": 0.5923, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.045166015625, | |
| "rewards/margins": 0.341796875, | |
| "rewards/rejected": -0.294921875, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 7.9761645140123845, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -122.0, | |
| "loss": 0.5527, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.12890625, | |
| "rewards/margins": 0.64453125, | |
| "rewards/rejected": -0.7734375, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 11.005045009756817, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -112.0, | |
| "logps/rejected": -143.0, | |
| "loss": 0.5722, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.765625, | |
| "rewards/margins": 0.6953125, | |
| "rewards/rejected": -1.4609375, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 8.061641195431347, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -100.0, | |
| "logps/rejected": -119.0, | |
| "loss": 0.58, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -0.28515625, | |
| "rewards/margins": 0.4921875, | |
| "rewards/rejected": -0.77734375, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 7.157664665562145, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -99.0, | |
| "logps/rejected": -98.5, | |
| "loss": 0.5778, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -0.06640625, | |
| "rewards/margins": 0.455078125, | |
| "rewards/rejected": -0.51953125, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 8.622794557745845, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -111.0, | |
| "logps/rejected": -115.5, | |
| "loss": 0.5849, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -0.33984375, | |
| "rewards/margins": 0.486328125, | |
| "rewards/rejected": -0.82421875, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 6.335368409957574, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -119.5, | |
| "loss": 0.5714, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.60546875, | |
| "rewards/rejected": -1.3984375, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 8.606498505467815, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -107.5, | |
| "logps/rejected": -107.5, | |
| "loss": 0.6085, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.578125, | |
| "rewards/margins": 0.419921875, | |
| "rewards/rejected": -0.99609375, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 10.14671010681783, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -108.0, | |
| "logps/rejected": -113.5, | |
| "loss": 0.6055, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -0.58984375, | |
| "rewards/margins": 0.314453125, | |
| "rewards/rejected": -0.90625, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.375, | |
| "eval_logits/rejected": -1.3828125, | |
| "eval_logps/chosen": -91.5, | |
| "eval_logps/rejected": -98.0, | |
| "eval_loss": 0.6832812428474426, | |
| "eval_rewards/accuracies": 0.1785714328289032, | |
| "eval_rewards/chosen": -0.4453125, | |
| "eval_rewards/margins": 0.22265625, | |
| "eval_rewards/rejected": -0.66796875, | |
| "eval_runtime": 12.162, | |
| "eval_samples_per_second": 16.445, | |
| "eval_steps_per_second": 0.576, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 7.1861162457197985, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -112.5, | |
| "logps/rejected": -120.0, | |
| "loss": 0.5679, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.56640625, | |
| "rewards/margins": 0.466796875, | |
| "rewards/rejected": -1.03125, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 7.26251570475917, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -108.0, | |
| "logps/rejected": -116.5, | |
| "loss": 0.5206, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -0.3984375, | |
| "rewards/margins": 0.7734375, | |
| "rewards/rejected": -1.171875, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 7.42279503637195, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -115.5, | |
| "loss": 0.533, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.72265625, | |
| "rewards/margins": 0.7734375, | |
| "rewards/rejected": -1.5, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 3.8915941029027423, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -123.5, | |
| "loss": 0.5438, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.625, | |
| "rewards/margins": 0.859375, | |
| "rewards/rejected": -1.484375, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 11.244606303188515, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -109.5, | |
| "logps/rejected": -123.0, | |
| "loss": 0.5285, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 0.7578125, | |
| "rewards/rejected": -1.59375, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 6.202211746206594, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -123.5, | |
| "loss": 0.5452, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -1.03125, | |
| "rewards/margins": 0.6328125, | |
| "rewards/rejected": -1.65625, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 2.203584000423544, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -124.5, | |
| "loss": 0.5582, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -1.1015625, | |
| "rewards/margins": 0.95703125, | |
| "rewards/rejected": -2.0625, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 7.599809646506574, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -114.0, | |
| "logps/rejected": -128.0, | |
| "loss": 0.5152, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -0.41796875, | |
| "rewards/margins": 0.8359375, | |
| "rewards/rejected": -1.2578125, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 16.13764475579174, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -114.5, | |
| "logps/rejected": -137.0, | |
| "loss": 0.5064, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 0.90234375, | |
| "rewards/rejected": -1.75, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 9.9694069692221, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -117.5, | |
| "logps/rejected": -138.0, | |
| "loss": 0.5356, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.7734375, | |
| "rewards/margins": 0.9140625, | |
| "rewards/rejected": -1.6875, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 15.80088733951728, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -124.5, | |
| "loss": 0.5391, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -0.59765625, | |
| "rewards/margins": 0.890625, | |
| "rewards/rejected": -1.4921875, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 7.588122085996054, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -114.5, | |
| "logps/rejected": -116.0, | |
| "loss": 0.5511, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.73046875, | |
| "rewards/margins": 0.59765625, | |
| "rewards/rejected": -1.328125, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 27.36534088884122, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -124.5, | |
| "logps/rejected": -139.0, | |
| "loss": 0.5164, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -1.1640625, | |
| "rewards/margins": 0.90625, | |
| "rewards/rejected": -2.078125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 2.522210813631439, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -114.0, | |
| "logps/rejected": -117.5, | |
| "loss": 0.5472, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -1.453125, | |
| "rewards/margins": 0.57421875, | |
| "rewards/rejected": -2.03125, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 6.690206445009991, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -96.0, | |
| "logps/rejected": -122.0, | |
| "loss": 0.5059, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.92578125, | |
| "rewards/margins": 0.8515625, | |
| "rewards/rejected": -1.78125, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 4.411569666492494, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -112.0, | |
| "logps/rejected": -135.0, | |
| "loss": 0.5082, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.859375, | |
| "rewards/margins": 1.3046875, | |
| "rewards/rejected": -2.15625, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 6.232058034991466, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -113.5, | |
| "logps/rejected": -130.0, | |
| "loss": 0.5355, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 0.78515625, | |
| "rewards/rejected": -2.3125, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 4.628852449599637, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -127.5, | |
| "logps/rejected": -145.0, | |
| "loss": 0.5347, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 0.8515625, | |
| "rewards/rejected": -2.125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 3.2274448332025334, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -110.5, | |
| "logps/rejected": -120.5, | |
| "loss": 0.5221, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -1.25, | |
| "rewards/margins": 0.9375, | |
| "rewards/rejected": -2.1875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 8.1074535346936, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -123.0, | |
| "logps/rejected": -138.0, | |
| "loss": 0.5326, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -1.515625, | |
| "rewards/margins": 0.671875, | |
| "rewards/rejected": -2.1875, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 6.483448432938418, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -120.5, | |
| "logps/rejected": -134.0, | |
| "loss": 0.5404, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -1.1640625, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -1.9765625, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 17.578564819223217, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -106.5, | |
| "logps/rejected": -125.5, | |
| "loss": 0.5345, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.67578125, | |
| "rewards/margins": 0.99609375, | |
| "rewards/rejected": -1.671875, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 1.7427437569760602, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -116.0, | |
| "logps/rejected": -137.0, | |
| "loss": 0.519, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -0.9453125, | |
| "rewards/margins": 1.203125, | |
| "rewards/rejected": -2.140625, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 9.189331850778302, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -116.0, | |
| "logps/rejected": -137.0, | |
| "loss": 0.5364, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -1.1875, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": -2.234375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 20.289715925590105, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -129.0, | |
| "logps/rejected": -145.0, | |
| "loss": 0.5214, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -1.6015625, | |
| "rewards/margins": 0.8359375, | |
| "rewards/rejected": -2.4375, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 5.161559692436978, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -109.0, | |
| "logps/rejected": -118.5, | |
| "loss": 0.555, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 0.58984375, | |
| "rewards/rejected": -2.109375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 11.491208251197875, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -124.5, | |
| "logps/rejected": -128.0, | |
| "loss": 0.5216, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -1.2265625, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -2.03125, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 10.936109944231491, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -119.0, | |
| "logps/rejected": -140.0, | |
| "loss": 0.5305, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 1.21875, | |
| "rewards/rejected": -2.484375, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 13.859739962430716, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.5390625, | |
| "logps/chosen": -134.0, | |
| "logps/rejected": -149.0, | |
| "loss": 0.5595, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -1.5078125, | |
| "rewards/margins": 0.78515625, | |
| "rewards/rejected": -2.296875, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 6.9996786378683105, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -111.5, | |
| "logps/rejected": -118.5, | |
| "loss": 0.5504, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -0.98828125, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": -1.8515625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 9.12819987781117, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -112.0, | |
| "logps/rejected": -119.5, | |
| "loss": 0.539, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.82421875, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -1.6328125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -1.3828125, | |
| "eval_logits/rejected": -1.3828125, | |
| "eval_logps/chosen": -99.0, | |
| "eval_logps/rejected": -105.5, | |
| "eval_loss": 0.731249988079071, | |
| "eval_rewards/accuracies": 0.1428571492433548, | |
| "eval_rewards/chosen": -1.171875, | |
| "eval_rewards/margins": 0.248046875, | |
| "eval_rewards/rejected": -1.421875, | |
| "eval_runtime": 14.6078, | |
| "eval_samples_per_second": 13.691, | |
| "eval_steps_per_second": 0.479, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5970425184288471, | |
| "train_runtime": 4683.6892, | |
| "train_samples_per_second": 6.405, | |
| "train_steps_per_second": 0.2 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 939, | |
| "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": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |