| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 34.8421052631579, | |
| "eval_steps": 500, | |
| "global_step": 140, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 2.4210526315789473, | |
| "grad_norm": 0.6825495362281799, | |
| "learning_rate": 0.00014285714285714287, | |
| "logits/chosen": -0.9514083862304688, | |
| "logits/rejected": -1.0533627271652222, | |
| "logps/chosen": -39.20041275024414, | |
| "logps/rejected": -21.37519073486328, | |
| "loss": 0.5664, | |
| "rewards/accuracies": 0.7743055820465088, | |
| "rewards/chosen": 0.15898703038692474, | |
| "rewards/margins": 0.5821112990379333, | |
| "rewards/rejected": -0.42312419414520264, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 4.842105263157895, | |
| "grad_norm": 0.08670935779809952, | |
| "learning_rate": 0.00019888308262251285, | |
| "logits/chosen": -0.13068915903568268, | |
| "logits/rejected": -0.16997402906417847, | |
| "logps/chosen": -74.40914916992188, | |
| "logps/rejected": -151.8779296875, | |
| "loss": 0.0471, | |
| "rewards/accuracies": 0.9635416865348816, | |
| "rewards/chosen": -3.3677978515625, | |
| "rewards/margins": 10.11207389831543, | |
| "rewards/rejected": -13.479872703552246, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 7.421052631578947, | |
| "grad_norm": 0.0034703314304351807, | |
| "learning_rate": 0.00019214762118704076, | |
| "logits/chosen": 0.19064301252365112, | |
| "logits/rejected": 0.24179647862911224, | |
| "logps/chosen": -95.65469360351562, | |
| "logps/rejected": -240.3935089111328, | |
| "loss": 0.0286, | |
| "rewards/accuracies": 0.9664474129676819, | |
| "rewards/chosen": -5.480339050292969, | |
| "rewards/margins": 16.84355354309082, | |
| "rewards/rejected": -22.323888778686523, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 9.842105263157894, | |
| "grad_norm": 0.0037726943846791983, | |
| "learning_rate": 0.00017971325072229226, | |
| "logits/chosen": 0.08421485126018524, | |
| "logits/rejected": 0.1458778977394104, | |
| "logps/chosen": -170.75747680664062, | |
| "logps/rejected": -437.0244140625, | |
| "loss": 0.0258, | |
| "rewards/accuracies": 0.9670138955116272, | |
| "rewards/chosen": -12.99338150024414, | |
| "rewards/margins": 28.99999237060547, | |
| "rewards/rejected": -41.993377685546875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 12.421052631578947, | |
| "grad_norm": 0.0033487407490611076, | |
| "learning_rate": 0.00016234898018587337, | |
| "logits/chosen": -0.04333849251270294, | |
| "logits/rejected": -0.01762447878718376, | |
| "logps/chosen": -186.37686157226562, | |
| "logps/rejected": -464.90765380859375, | |
| "loss": 0.0272, | |
| "rewards/accuracies": 0.9671053290367126, | |
| "rewards/chosen": -14.553586959838867, | |
| "rewards/margins": 30.223119735717773, | |
| "rewards/rejected": -44.77670669555664, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 14.842105263157894, | |
| "grad_norm": 0.003097180975601077, | |
| "learning_rate": 0.00014112871031306119, | |
| "logits/chosen": -0.13475045561790466, | |
| "logits/rejected": -0.1398223638534546, | |
| "logps/chosen": -184.91830444335938, | |
| "logps/rejected": -461.1513977050781, | |
| "loss": 0.0244, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -14.403005599975586, | |
| "rewards/margins": 29.99542999267578, | |
| "rewards/rejected": -44.398433685302734, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 17.42105263157895, | |
| "grad_norm": 0.0024468335323035717, | |
| "learning_rate": 0.00011736481776669306, | |
| "logits/chosen": -0.1959075629711151, | |
| "logits/rejected": -0.22540684044361115, | |
| "logps/chosen": -185.68663024902344, | |
| "logps/rejected": -458.96826171875, | |
| "loss": 0.0261, | |
| "rewards/accuracies": 0.9684211015701294, | |
| "rewards/chosen": -14.49423885345459, | |
| "rewards/margins": 29.691282272338867, | |
| "rewards/rejected": -44.185516357421875, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 19.842105263157894, | |
| "grad_norm": 0.002537067048251629, | |
| "learning_rate": 9.252699064135758e-05, | |
| "logits/chosen": -0.22896860539913177, | |
| "logits/rejected": -0.2724004089832306, | |
| "logps/chosen": -184.9608154296875, | |
| "logps/rejected": -460.61468505859375, | |
| "loss": 0.0266, | |
| "rewards/accuracies": 0.9659722447395325, | |
| "rewards/chosen": -14.424090385437012, | |
| "rewards/margins": 29.924333572387695, | |
| "rewards/rejected": -44.34842300415039, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 22.42105263157895, | |
| "grad_norm": 0.0032824031077325344, | |
| "learning_rate": 6.815133497483157e-05, | |
| "logits/chosen": -0.24728278815746307, | |
| "logits/rejected": -0.2936950922012329, | |
| "logps/chosen": -185.9248046875, | |
| "logps/rejected": -467.2992858886719, | |
| "loss": 0.0287, | |
| "rewards/accuracies": 0.9651316404342651, | |
| "rewards/chosen": -14.498348236083984, | |
| "rewards/margins": 30.518945693969727, | |
| "rewards/rejected": -45.01729202270508, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 24.842105263157894, | |
| "grad_norm": 0.003522921120747924, | |
| "learning_rate": 4.574537361342407e-05, | |
| "logits/chosen": -0.2618289887905121, | |
| "logits/rejected": -0.311085045337677, | |
| "logps/chosen": -187.0611572265625, | |
| "logps/rejected": -466.8361511230469, | |
| "loss": 0.0258, | |
| "rewards/accuracies": 0.9670138955116272, | |
| "rewards/chosen": -14.643547058105469, | |
| "rewards/margins": 30.329221725463867, | |
| "rewards/rejected": -44.97277069091797, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 27.42105263157895, | |
| "grad_norm": 0.0024658790789544582, | |
| "learning_rate": 2.669481281701739e-05, | |
| "logits/chosen": -0.27023741602897644, | |
| "logits/rejected": -0.3240560293197632, | |
| "logps/chosen": -186.94210815429688, | |
| "logps/rejected": -473.1488952636719, | |
| "loss": 0.0276, | |
| "rewards/accuracies": 0.9664474129676819, | |
| "rewards/chosen": -14.606731414794922, | |
| "rewards/margins": 30.998502731323242, | |
| "rewards/rejected": -45.6052360534668, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 29.842105263157894, | |
| "grad_norm": 0.0034348091576248407, | |
| "learning_rate": 1.2177842662977135e-05, | |
| "logits/chosen": -0.27230748534202576, | |
| "logits/rejected": -0.32499459385871887, | |
| "logps/chosen": -185.62405395507812, | |
| "logps/rejected": -470.46063232421875, | |
| "loss": 0.0252, | |
| "rewards/accuracies": 0.9677083492279053, | |
| "rewards/chosen": -14.492985725402832, | |
| "rewards/margins": 30.83690071105957, | |
| "rewards/rejected": -45.329891204833984, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 32.421052631578945, | |
| "grad_norm": 0.0032501835376024246, | |
| "learning_rate": 3.092271377092215e-06, | |
| "logits/chosen": -0.27767735719680786, | |
| "logits/rejected": -0.3315570652484894, | |
| "logps/chosen": -187.55230712890625, | |
| "logps/rejected": -470.9134521484375, | |
| "loss": 0.0287, | |
| "rewards/accuracies": 0.9651316404342651, | |
| "rewards/chosen": -14.68417739868164, | |
| "rewards/margins": 30.694440841674805, | |
| "rewards/rejected": -45.37861633300781, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 34.8421052631579, | |
| "grad_norm": 0.003052822547033429, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -0.27554523944854736, | |
| "logits/rejected": -0.32890552282333374, | |
| "logps/chosen": -185.5701904296875, | |
| "logps/rejected": -472.9700927734375, | |
| "loss": 0.0244, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -14.46578311920166, | |
| "rewards/margins": 31.117090225219727, | |
| "rewards/rejected": -45.58286666870117, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 34.8421052631579, | |
| "step": 140, | |
| "total_flos": 3.1072679368851456e+17, | |
| "train_loss": 0.06661632827350071, | |
| "train_runtime": 6674.5686, | |
| "train_samples_per_second": 6.319, | |
| "train_steps_per_second": 0.021 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 140, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 35, | |
| "save_steps": 70, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.1072679368851456e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |