| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020920502092050207, | |
| "grad_norm": 9.099456092432153, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -3.359375, | |
| "logits/rejected": -3.3125, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -360.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02092050209205021, | |
| "grad_norm": 8.217366878859176, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -3.15625, | |
| "logits/rejected": -3.1875, | |
| "logps/chosen": -262.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.2708333432674408, | |
| "rewards/chosen": 0.000598907470703125, | |
| "rewards/margins": 0.00058746337890625, | |
| "rewards/rejected": 7.867813110351562e-06, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04184100418410042, | |
| "grad_norm": 8.628513952797078, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -3.21875, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.6909, | |
| "rewards/accuracies": 0.3687500059604645, | |
| "rewards/chosen": -0.00173187255859375, | |
| "rewards/margins": 0.0021820068359375, | |
| "rewards/rejected": -0.00390625, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06276150627615062, | |
| "grad_norm": 9.169127445477065, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -3.25, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.6832, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.0167236328125, | |
| "rewards/margins": 0.01611328125, | |
| "rewards/rejected": -0.032958984375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08368200836820083, | |
| "grad_norm": 9.282045835045837, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -3.09375, | |
| "logits/rejected": -3.1875, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.6603, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.051513671875, | |
| "rewards/margins": 0.0771484375, | |
| "rewards/rejected": -0.12890625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10460251046025104, | |
| "grad_norm": 12.0230941013448, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -320.0, | |
| "loss": 0.6096, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.056396484375, | |
| "rewards/margins": 0.177734375, | |
| "rewards/rejected": -0.2333984375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12552301255230125, | |
| "grad_norm": 19.700932387600457, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -3.1875, | |
| "logits/rejected": -3.140625, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -354.0, | |
| "loss": 0.5051, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -0.1455078125, | |
| "rewards/margins": 0.443359375, | |
| "rewards/rejected": -0.5859375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14644351464435146, | |
| "grad_norm": 25.122385996985045, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.984375, | |
| "logits/rejected": -3.0, | |
| "logps/chosen": -306.0, | |
| "logps/rejected": -422.0, | |
| "loss": 0.3824, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.4140625, | |
| "rewards/margins": 1.0234375, | |
| "rewards/rejected": -1.4375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16736401673640167, | |
| "grad_norm": 29.28714176471108, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -3.09375, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -406.0, | |
| "logps/rejected": -544.0, | |
| "loss": 0.2788, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.765625, | |
| "rewards/margins": 1.7421875, | |
| "rewards/rejected": -2.5, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18828451882845187, | |
| "grad_norm": 30.45351538248815, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -3.015625, | |
| "logits/rejected": -2.953125, | |
| "logps/chosen": -356.0, | |
| "logps/rejected": -560.0, | |
| "loss": 0.251, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.91015625, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -3.09375, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "grad_norm": 25.879079240835583, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.875, | |
| "logps/chosen": -384.0, | |
| "logps/rejected": -620.0, | |
| "loss": 0.2895, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.1796875, | |
| "rewards/margins": 2.40625, | |
| "rewards/rejected": -3.59375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2301255230125523, | |
| "grad_norm": 59.34338493619359, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.890625, | |
| "logits/rejected": -2.78125, | |
| "logps/chosen": -436.0, | |
| "logps/rejected": -656.0, | |
| "loss": 0.2514, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -1.3203125, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": -3.65625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2510460251046025, | |
| "grad_norm": 34.20651774561803, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.78125, | |
| "logits/rejected": -2.6875, | |
| "logps/chosen": -392.0, | |
| "logps/rejected": -628.0, | |
| "loss": 0.2079, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 2.46875, | |
| "rewards/rejected": -3.5625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2719665271966527, | |
| "grad_norm": 34.47099090322982, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.703125, | |
| "logps/chosen": -392.0, | |
| "logps/rejected": -664.0, | |
| "loss": 0.2106, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.1328125, | |
| "rewards/margins": 2.78125, | |
| "rewards/rejected": -3.90625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2928870292887029, | |
| "grad_norm": 25.816296928179703, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.671875, | |
| "logps/chosen": -450.0, | |
| "logps/rejected": -736.0, | |
| "loss": 0.2027, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.3828125, | |
| "rewards/margins": 3.140625, | |
| "rewards/rejected": -4.5, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3138075313807531, | |
| "grad_norm": 25.9983799849635, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.71875, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -432.0, | |
| "logps/rejected": -808.0, | |
| "loss": 0.2011, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.453125, | |
| "rewards/margins": 3.796875, | |
| "rewards/rejected": -5.25, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33472803347280333, | |
| "grad_norm": 25.6614814368812, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -744.0, | |
| "loss": 0.1733, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.2265625, | |
| "rewards/margins": 3.453125, | |
| "rewards/rejected": -4.6875, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35564853556485354, | |
| "grad_norm": 23.516456639083984, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.578125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -760.0, | |
| "loss": 0.1909, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.5, | |
| "rewards/margins": 3.5, | |
| "rewards/rejected": -5.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37656903765690375, | |
| "grad_norm": 40.42551024225754, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -812.0, | |
| "loss": 0.1967, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.4453125, | |
| "rewards/margins": 4.03125, | |
| "rewards/rejected": -5.5, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39748953974895396, | |
| "grad_norm": 19.01883870968105, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -406.0, | |
| "logps/rejected": -748.0, | |
| "loss": 0.1892, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -1.2265625, | |
| "rewards/margins": 3.453125, | |
| "rewards/rejected": -4.6875, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "grad_norm": 58.64663183468128, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -424.0, | |
| "logps/rejected": -836.0, | |
| "loss": 0.1791, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.5859375, | |
| "rewards/margins": 3.984375, | |
| "rewards/rejected": -5.5625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4393305439330544, | |
| "grad_norm": 35.29196540474136, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.5, | |
| "logps/chosen": -450.0, | |
| "logps/rejected": -860.0, | |
| "loss": 0.1883, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.5390625, | |
| "rewards/margins": 4.09375, | |
| "rewards/rejected": -5.65625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4602510460251046, | |
| "grad_norm": 42.58286407786604, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -400.0, | |
| "logps/rejected": -776.0, | |
| "loss": 0.1641, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.3671875, | |
| "rewards/margins": 3.65625, | |
| "rewards/rejected": -5.03125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4811715481171548, | |
| "grad_norm": 57.96355489388366, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.46875, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -408.0, | |
| "logps/rejected": -872.0, | |
| "loss": 0.1766, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -6.21875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.502092050209205, | |
| "grad_norm": 38.85695272700362, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.578125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -428.0, | |
| "logps/rejected": -880.0, | |
| "loss": 0.1776, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.6015625, | |
| "rewards/margins": 4.5625, | |
| "rewards/rejected": -6.15625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5230125523012552, | |
| "grad_norm": 32.765054711715315, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -436.0, | |
| "logps/rejected": -852.0, | |
| "loss": 0.1456, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -1.328125, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -5.78125, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5439330543933054, | |
| "grad_norm": 38.81475856875557, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -502.0, | |
| "logps/rejected": -1032.0, | |
| "loss": 0.1545, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -2.078125, | |
| "rewards/margins": 5.40625, | |
| "rewards/rejected": -7.5, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5648535564853556, | |
| "grad_norm": 49.7654758904445, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -500.0, | |
| "logps/rejected": -960.0, | |
| "loss": 0.1565, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -2.21875, | |
| "rewards/margins": 4.5625, | |
| "rewards/rejected": -6.8125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5857740585774058, | |
| "grad_norm": 31.170463152628752, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.9921875, | |
| "logps/chosen": -462.0, | |
| "logps/rejected": -968.0, | |
| "loss": 0.1641, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.984375, | |
| "rewards/margins": 4.96875, | |
| "rewards/rejected": -6.9375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.606694560669456, | |
| "grad_norm": 27.441348723076615, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -1.953125, | |
| "logps/chosen": -512.0, | |
| "logps/rejected": -956.0, | |
| "loss": 0.1675, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -2.53125, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -7.125, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "grad_norm": 20.30528868038641, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -1.890625, | |
| "logps/chosen": -540.0, | |
| "logps/rejected": -1088.0, | |
| "loss": 0.1308, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.40625, | |
| "rewards/margins": 5.75, | |
| "rewards/rejected": -8.125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6485355648535565, | |
| "grad_norm": 34.670577413518515, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -1.875, | |
| "logps/chosen": -592.0, | |
| "logps/rejected": -1152.0, | |
| "loss": 0.1346, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.953125, | |
| "rewards/margins": 5.90625, | |
| "rewards/rejected": -8.875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6694560669456067, | |
| "grad_norm": 29.612754262891425, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -1.8125, | |
| "logps/chosen": -604.0, | |
| "logps/rejected": -1248.0, | |
| "loss": 0.1433, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.09375, | |
| "rewards/margins": 6.53125, | |
| "rewards/rejected": -9.625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6903765690376569, | |
| "grad_norm": 25.080256733689154, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -1.9140625, | |
| "logps/chosen": -572.0, | |
| "logps/rejected": -1112.0, | |
| "loss": 0.1432, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.015625, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -8.5, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7112970711297071, | |
| "grad_norm": 35.91711456332533, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -1.9765625, | |
| "logps/chosen": -632.0, | |
| "logps/rejected": -1256.0, | |
| "loss": 0.1378, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.328125, | |
| "rewards/margins": 6.21875, | |
| "rewards/rejected": -9.5625, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7322175732217573, | |
| "grad_norm": 28.74300288909294, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -2.109375, | |
| "logits/rejected": -1.796875, | |
| "logps/chosen": -604.0, | |
| "logps/rejected": -1184.0, | |
| "loss": 0.1044, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.390625, | |
| "rewards/margins": 5.75, | |
| "rewards/rejected": -9.125, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7531380753138075, | |
| "grad_norm": 26.068963609585946, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -1.8828125, | |
| "logps/chosen": -680.0, | |
| "logps/rejected": -1272.0, | |
| "loss": 0.1278, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -3.75, | |
| "rewards/margins": 6.09375, | |
| "rewards/rejected": -9.8125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7740585774058577, | |
| "grad_norm": 88.91144605940063, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -1.7890625, | |
| "logps/chosen": -612.0, | |
| "logps/rejected": -1192.0, | |
| "loss": 0.1626, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.359375, | |
| "rewards/margins": 6.1875, | |
| "rewards/rejected": -9.5625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7949790794979079, | |
| "grad_norm": 16.463309523333024, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.09375, | |
| "logits/rejected": -1.796875, | |
| "logps/chosen": -552.0, | |
| "logps/rejected": -1136.0, | |
| "loss": 0.139, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.109375, | |
| "rewards/margins": 5.5625, | |
| "rewards/rejected": -8.6875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8158995815899581, | |
| "grad_norm": 37.449359063530714, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -1.9375, | |
| "logps/chosen": -644.0, | |
| "logps/rejected": -1240.0, | |
| "loss": 0.1281, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.34375, | |
| "rewards/margins": 6.03125, | |
| "rewards/rejected": -9.375, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 24.80612572089887, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.9375, | |
| "logps/chosen": -608.0, | |
| "logps/rejected": -1240.0, | |
| "loss": 0.1349, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.34375, | |
| "rewards/margins": 6.21875, | |
| "rewards/rejected": -9.5625, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8577405857740585, | |
| "grad_norm": 24.11144809301846, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -1.9921875, | |
| "logps/chosen": -620.0, | |
| "logps/rejected": -1208.0, | |
| "loss": 0.1372, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.265625, | |
| "rewards/margins": 6.21875, | |
| "rewards/rejected": -9.4375, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8786610878661087, | |
| "grad_norm": 20.712692773585854, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -2.0, | |
| "logps/chosen": -628.0, | |
| "logps/rejected": -1176.0, | |
| "loss": 0.1391, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.203125, | |
| "rewards/margins": 5.71875, | |
| "rewards/rejected": -8.9375, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.899581589958159, | |
| "grad_norm": 14.713063691289868, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.15625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -580.0, | |
| "logps/rejected": -1192.0, | |
| "loss": 0.1275, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.0, | |
| "rewards/margins": 6.53125, | |
| "rewards/rejected": -9.5, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9205020920502092, | |
| "grad_norm": 42.78914335876745, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.9296875, | |
| "logps/chosen": -604.0, | |
| "logps/rejected": -1272.0, | |
| "loss": 0.1322, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.140625, | |
| "rewards/margins": 6.65625, | |
| "rewards/rejected": -9.8125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9414225941422594, | |
| "grad_norm": 20.26822701097891, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -1.890625, | |
| "logps/chosen": -608.0, | |
| "logps/rejected": -1264.0, | |
| "loss": 0.1221, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.1875, | |
| "rewards/margins": 6.5, | |
| "rewards/rejected": -9.6875, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9623430962343096, | |
| "grad_norm": 18.31779303714104, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -2.1875, | |
| "logits/rejected": -1.8515625, | |
| "logps/chosen": -604.0, | |
| "logps/rejected": -1224.0, | |
| "loss": 0.1222, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.28125, | |
| "rewards/margins": 6.28125, | |
| "rewards/rejected": -9.5625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9832635983263598, | |
| "grad_norm": 31.833980309732567, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -1.84375, | |
| "logps/chosen": -632.0, | |
| "logps/rejected": -1168.0, | |
| "loss": 0.1265, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -3.5, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -9.0, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 0.0488231029969379, | |
| "train_runtime": 2602.953, | |
| "train_samples_per_second": 23.486, | |
| "train_steps_per_second": 0.184 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "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 | |
| } | |