| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.8368200836820083, | |
| "eval_steps": 500, | |
| "global_step": 100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.008368200836820083, | |
| "grad_norm": 9.75007616796795, | |
| "learning_rate": 4.166666666666666e-08, | |
| "logits/chosen": -2.761155366897583, | |
| "logits/rejected": -2.7110397815704346, | |
| "logps/chosen": -188.2120819091797, | |
| "logps/rejected": -227.97329711914062, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.08368200836820083, | |
| "grad_norm": 8.857631148542579, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.752246856689453, | |
| "logits/rejected": -2.7512197494506836, | |
| "logps/chosen": -286.0818786621094, | |
| "logps/rejected": -270.5452880859375, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.4791666567325592, | |
| "rewards/chosen": 0.0016119987703859806, | |
| "rewards/margins": 0.0004296954721212387, | |
| "rewards/rejected": 0.001182303298264742, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16736401673640167, | |
| "grad_norm": 8.648350893662677, | |
| "learning_rate": 4.931352528237397e-07, | |
| "logits/chosen": -2.816281795501709, | |
| "logits/rejected": -2.7909739017486572, | |
| "logps/chosen": -281.4281311035156, | |
| "logps/rejected": -259.23931884765625, | |
| "loss": 0.6846, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.02481575682759285, | |
| "rewards/margins": 0.019779205322265625, | |
| "rewards/rejected": 0.00503655057400465, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2510460251046025, | |
| "grad_norm": 9.462527888383683, | |
| "learning_rate": 4.658920803689553e-07, | |
| "logits/chosen": -2.7952637672424316, | |
| "logits/rejected": -2.7442078590393066, | |
| "logps/chosen": -272.9776306152344, | |
| "logps/rejected": -227.1405792236328, | |
| "loss": 0.6647, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.021379830315709114, | |
| "rewards/margins": 0.0722198635339737, | |
| "rewards/rejected": -0.05084002763032913, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.33472803347280333, | |
| "grad_norm": 10.802670702282992, | |
| "learning_rate": 4.201712553872657e-07, | |
| "logits/chosen": -2.808018445968628, | |
| "logits/rejected": -2.7901508808135986, | |
| "logps/chosen": -269.8119812011719, | |
| "logps/rejected": -276.64312744140625, | |
| "loss": 0.6451, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.11026054620742798, | |
| "rewards/margins": 0.09342513978481293, | |
| "rewards/rejected": -0.2036857157945633, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "grad_norm": 14.0383641925761, | |
| "learning_rate": 3.598859066780754e-07, | |
| "logits/chosen": -2.801910877227783, | |
| "logits/rejected": -2.7805628776550293, | |
| "logps/chosen": -308.90960693359375, | |
| "logps/rejected": -296.39837646484375, | |
| "loss": 0.6148, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.1705964356660843, | |
| "rewards/margins": 0.27841097116470337, | |
| "rewards/rejected": -0.44900742173194885, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.502092050209205, | |
| "grad_norm": 17.441329707174045, | |
| "learning_rate": 2.9019570347986706e-07, | |
| "logits/chosen": -2.8074169158935547, | |
| "logits/rejected": -2.7729382514953613, | |
| "logps/chosen": -314.9671325683594, | |
| "logps/rejected": -307.9980773925781, | |
| "loss": 0.6091, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.3278293013572693, | |
| "rewards/margins": 0.2875938415527344, | |
| "rewards/rejected": -0.6154230833053589, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.5857740585774058, | |
| "grad_norm": 13.990328041846443, | |
| "learning_rate": 2.1706525253979534e-07, | |
| "logits/chosen": -2.808964252471924, | |
| "logits/rejected": -2.7739078998565674, | |
| "logps/chosen": -318.0167541503906, | |
| "logps/rejected": -318.5818786621094, | |
| "loss": 0.6029, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.400399774312973, | |
| "rewards/margins": 0.22247812151908875, | |
| "rewards/rejected": -0.6228778958320618, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6694560669456067, | |
| "grad_norm": 17.12917679172734, | |
| "learning_rate": 1.4675360263490295e-07, | |
| "logits/chosen": -2.7654800415039062, | |
| "logits/rejected": -2.7729649543762207, | |
| "logps/chosen": -259.6241149902344, | |
| "logps/rejected": -305.7326354980469, | |
| "loss": 0.6021, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.29664868116378784, | |
| "rewards/margins": 0.25704866647720337, | |
| "rewards/rejected": -0.5536972880363464, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.7531380753138075, | |
| "grad_norm": 19.058433635316273, | |
| "learning_rate": 8.527854855097224e-08, | |
| "logits/chosen": -2.727175712585449, | |
| "logits/rejected": -2.7393229007720947, | |
| "logps/chosen": -297.5341491699219, | |
| "logps/rejected": -314.9481201171875, | |
| "loss": 0.5793, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.36374929547309875, | |
| "rewards/margins": 0.3506723940372467, | |
| "rewards/rejected": -0.7144217491149902, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 15.477637607113444, | |
| "learning_rate": 3.790158337517127e-08, | |
| "logits/chosen": -2.7408394813537598, | |
| "logits/rejected": -2.671611785888672, | |
| "logps/chosen": -293.8873596191406, | |
| "logps/rejected": -300.93572998046875, | |
| "loss": 0.5918, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.41290682554244995, | |
| "rewards/margins": 0.2459905445575714, | |
| "rewards/rejected": -0.6588973999023438, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "step": 100, | |
| "total_flos": 0.0, | |
| "train_loss": 0.0, | |
| "train_runtime": 11.541, | |
| "train_samples_per_second": 1324.231, | |
| "train_steps_per_second": 5.112 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 59, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |