| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.4981684981684982, | |
| "eval_steps": 500, | |
| "global_step": 238, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020931449502878076, | |
| "grad_norm": 26.601607092770625, | |
| "learning_rate": 2.083333333333333e-08, | |
| "logits/chosen": -2.9139022827148438, | |
| "logits/rejected": -2.8787596225738525, | |
| "logps/chosen": -325.8533020019531, | |
| "logps/rejected": -372.9187927246094, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 16.144681025948177, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.730161428451538, | |
| "logits/rejected": -2.7064931392669678, | |
| "logps/chosen": -331.04364013671875, | |
| "logps/rejected": -309.8411865234375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.4861111044883728, | |
| "rewards/chosen": 0.001100256573408842, | |
| "rewards/margins": 0.0007432710262946784, | |
| "rewards/rejected": 0.0003569853724911809, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 14.575580508073863, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.798081874847412, | |
| "logits/rejected": -2.7473387718200684, | |
| "logps/chosen": -321.3909912109375, | |
| "logps/rejected": -299.4921569824219, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": 0.006536015775054693, | |
| "rewards/margins": 0.0068586282432079315, | |
| "rewards/rejected": -0.0003226128756068647, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 8.070909879894359, | |
| "learning_rate": 4.99030821197584e-07, | |
| "logits/chosen": -2.7842161655426025, | |
| "logits/rejected": -2.728454113006592, | |
| "logps/chosen": -304.2439880371094, | |
| "logps/rejected": -275.6659851074219, | |
| "loss": 0.6863, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.04330765828490257, | |
| "rewards/margins": 0.05840452387928963, | |
| "rewards/rejected": -0.015096860937774181, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 8.530980057951403, | |
| "learning_rate": 4.931352528237397e-07, | |
| "logits/chosen": -2.7036166191101074, | |
| "logits/rejected": -2.671600103378296, | |
| "logps/chosen": -289.6343078613281, | |
| "logps/rejected": -277.79425048828125, | |
| "loss": 0.6767, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.02837621606886387, | |
| "rewards/margins": 0.037141989916563034, | |
| "rewards/rejected": -0.008765773847699165, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 11.721866187252868, | |
| "learning_rate": 4.820092227512735e-07, | |
| "logits/chosen": -2.740696430206299, | |
| "logits/rejected": -2.6848952770233154, | |
| "logps/chosen": -290.8323059082031, | |
| "logps/rejected": -310.22576904296875, | |
| "loss": 0.6607, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.014719474129378796, | |
| "rewards/margins": 0.10523072630167007, | |
| "rewards/rejected": -0.11995019763708115, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 10.915648900872057, | |
| "learning_rate": 4.658920803689553e-07, | |
| "logits/chosen": -2.785388469696045, | |
| "logits/rejected": -2.744814395904541, | |
| "logps/chosen": -268.2928466796875, | |
| "logps/rejected": -272.32122802734375, | |
| "loss": 0.6461, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.02511899545788765, | |
| "rewards/margins": 0.14482316374778748, | |
| "rewards/rejected": -0.16994217038154602, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 11.152537838170248, | |
| "learning_rate": 4.4513054666826144e-07, | |
| "logits/chosen": -2.745926856994629, | |
| "logits/rejected": -2.7201263904571533, | |
| "logps/chosen": -332.7641296386719, | |
| "logps/rejected": -326.02862548828125, | |
| "loss": 0.6338, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.12678995728492737, | |
| "rewards/margins": 0.15600599348545074, | |
| "rewards/rejected": -0.2827959656715393, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 20.747423551686875, | |
| "learning_rate": 4.201712553872657e-07, | |
| "logits/chosen": -2.8221614360809326, | |
| "logits/rejected": -2.75636887550354, | |
| "logps/chosen": -364.9082336425781, | |
| "logps/rejected": -309.4717102050781, | |
| "loss": 0.6129, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.07667700946331024, | |
| "rewards/margins": 0.23987922072410583, | |
| "rewards/rejected": -0.3165562152862549, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 12.643373189764793, | |
| "learning_rate": 3.9155114477557926e-07, | |
| "logits/chosen": -2.7849178314208984, | |
| "logits/rejected": -2.748291492462158, | |
| "logps/chosen": -287.956787109375, | |
| "logps/rejected": -340.713134765625, | |
| "loss": 0.5971, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.24098113179206848, | |
| "rewards/margins": 0.3181685507297516, | |
| "rewards/rejected": -0.5591496825218201, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 13.623537820936637, | |
| "learning_rate": 3.598859066780754e-07, | |
| "logits/chosen": -2.76792311668396, | |
| "logits/rejected": -2.7459118366241455, | |
| "logps/chosen": -346.0640563964844, | |
| "logps/rejected": -348.5198059082031, | |
| "loss": 0.6098, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.2652357220649719, | |
| "rewards/margins": 0.34126365184783936, | |
| "rewards/rejected": -0.6064993739128113, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 15.41884704737969, | |
| "learning_rate": 3.2585674142717477e-07, | |
| "logits/chosen": -2.7856485843658447, | |
| "logits/rejected": -2.7417104244232178, | |
| "logps/chosen": -335.01116943359375, | |
| "logps/rejected": -349.71392822265625, | |
| "loss": 0.6003, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.32472461462020874, | |
| "rewards/margins": 0.3165324330329895, | |
| "rewards/rejected": -0.6412570476531982, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 15.439444003212843, | |
| "learning_rate": 2.9019570347986706e-07, | |
| "logits/chosen": -2.710737705230713, | |
| "logits/rejected": -2.718116283416748, | |
| "logps/chosen": -330.0270690917969, | |
| "logps/rejected": -354.4695739746094, | |
| "loss": 0.5832, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.3180462718009949, | |
| "rewards/margins": 0.3863913416862488, | |
| "rewards/rejected": -0.7044375538825989, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 42.02069053016682, | |
| "learning_rate": 2.536699530523291e-07, | |
| "logits/chosen": -2.7691874504089355, | |
| "logits/rejected": -2.752965211868286, | |
| "logps/chosen": -347.0602722167969, | |
| "logps/rejected": -357.85504150390625, | |
| "loss": 0.5817, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.2763240337371826, | |
| "rewards/margins": 0.27957138419151306, | |
| "rewards/rejected": -0.5558954477310181, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 24.79876226607902, | |
| "learning_rate": 2.1706525253979534e-07, | |
| "logits/chosen": -2.76993989944458, | |
| "logits/rejected": -2.750406265258789, | |
| "logps/chosen": -360.31005859375, | |
| "logps/rejected": -362.851806640625, | |
| "loss": 0.588, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.26679104566574097, | |
| "rewards/margins": 0.46773427724838257, | |
| "rewards/rejected": -0.7345253229141235, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 22.26447349378322, | |
| "learning_rate": 1.8116906275593507e-07, | |
| "logits/chosen": -2.7364563941955566, | |
| "logits/rejected": -2.7079169750213623, | |
| "logps/chosen": -351.19189453125, | |
| "logps/rejected": -349.318603515625, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.43332210183143616, | |
| "rewards/margins": 0.3800078332424164, | |
| "rewards/rejected": -0.8133300542831421, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 26.67231794642688, | |
| "learning_rate": 1.4675360263490295e-07, | |
| "logits/chosen": -2.730668544769287, | |
| "logits/rejected": -2.683061122894287, | |
| "logps/chosen": -329.1626281738281, | |
| "logps/rejected": -350.14288330078125, | |
| "loss": 0.5765, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.4029548764228821, | |
| "rewards/margins": 0.5021006464958191, | |
| "rewards/rejected": -0.9050555229187012, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 20.223571694588983, | |
| "learning_rate": 1.1455923682523475e-07, | |
| "logits/chosen": -2.532707691192627, | |
| "logits/rejected": -2.512026309967041, | |
| "logps/chosen": -324.9989929199219, | |
| "logps/rejected": -334.66888427734375, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.36262771487236023, | |
| "rewards/margins": 0.39737457036972046, | |
| "rewards/rejected": -0.7600023150444031, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 18.84156651008044, | |
| "learning_rate": 8.527854855097224e-08, | |
| "logits/chosen": -2.5160155296325684, | |
| "logits/rejected": -2.4737634658813477, | |
| "logps/chosen": -328.61224365234375, | |
| "logps/rejected": -344.9612121582031, | |
| "loss": 0.5499, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.454254150390625, | |
| "rewards/margins": 0.40306931734085083, | |
| "rewards/rejected": -0.8573234677314758, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 32.06520543659375, | |
| "learning_rate": 5.9541440373546445e-08, | |
| "logits/chosen": -2.3760242462158203, | |
| "logits/rejected": -2.352184772491455, | |
| "logps/chosen": -323.0282897949219, | |
| "logps/rejected": -362.2243957519531, | |
| "loss": 0.5613, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.5423271059989929, | |
| "rewards/margins": 0.3743038773536682, | |
| "rewards/rejected": -0.9166310429573059, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 23.15044859855663, | |
| "learning_rate": 3.790158337517127e-08, | |
| "logits/chosen": -2.311206817626953, | |
| "logits/rejected": -2.2949588298797607, | |
| "logps/chosen": -342.80731201171875, | |
| "logps/rejected": -394.5885314941406, | |
| "loss": 0.5711, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.5416313409805298, | |
| "rewards/margins": 0.41420310735702515, | |
| "rewards/rejected": -0.9558345079421997, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 19.06533234677666, | |
| "learning_rate": 2.0824506276503894e-08, | |
| "logits/chosen": -2.3925139904022217, | |
| "logits/rejected": -2.268831491470337, | |
| "logps/chosen": -395.4692077636719, | |
| "logps/rejected": -360.0453186035156, | |
| "loss": 0.5805, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.5327693223953247, | |
| "rewards/margins": 0.4689061641693115, | |
| "rewards/rejected": -1.0016754865646362, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 24.03298915942009, | |
| "learning_rate": 8.677580722139671e-09, | |
| "logits/chosen": -2.316939353942871, | |
| "logits/rejected": -2.2927820682525635, | |
| "logps/chosen": -342.931640625, | |
| "logps/rejected": -364.62982177734375, | |
| "loss": 0.5674, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.501736581325531, | |
| "rewards/margins": 0.49570217728614807, | |
| "rewards/rejected": -0.9974387884140015, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 19.987346780573713, | |
| "learning_rate": 1.722118176089915e-09, | |
| "logits/chosen": -2.4182724952697754, | |
| "logits/rejected": -2.337088108062744, | |
| "logps/chosen": -337.9970703125, | |
| "logps/rejected": -381.84320068359375, | |
| "loss": 0.5667, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.39490121603012085, | |
| "rewards/margins": 0.7370277643203735, | |
| "rewards/rejected": -1.1319290399551392, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.4981684981684982, | |
| "eval_logits/chosen": -2.3480114936828613, | |
| "eval_logits/rejected": -2.3013908863067627, | |
| "eval_logps/chosen": -315.5382385253906, | |
| "eval_logps/rejected": -372.7966003417969, | |
| "eval_loss": 0.5668980479240417, | |
| "eval_rewards/accuracies": 0.74609375, | |
| "eval_rewards/chosen": -0.4034212827682495, | |
| "eval_rewards/margins": 0.5594114065170288, | |
| "eval_rewards/rejected": -0.9628326892852783, | |
| "eval_runtime": 171.8973, | |
| "eval_samples_per_second": 11.635, | |
| "eval_steps_per_second": 0.186, | |
| "step": 238 | |
| }, | |
| { | |
| "epoch": 0.4981684981684982, | |
| "step": 238, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6061525304778284, | |
| "train_runtime": 7204.502, | |
| "train_samples_per_second": 4.228, | |
| "train_steps_per_second": 0.033 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 238, | |
| "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": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |