| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9893390191897654, | |
| "eval_steps": 100, | |
| "global_step": 58, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 594.6719055175781, | |
| "epoch": 0.017057569296375266, | |
| "grad_norm": 0.3902476131916046, | |
| "kl": 0.0, | |
| "learning_rate": 5e-07, | |
| "loss": -0.0015, | |
| "reward": 0.5937500223517418, | |
| "reward_std": 0.36258383840322495, | |
| "rewards/accuracy_reward": 0.5937500223517418, | |
| "rewards/format_reward": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 604.8621883392334, | |
| "epoch": 0.08528784648187633, | |
| "grad_norm": 0.5581268072128296, | |
| "kl": 0.000360637903213501, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.0195, | |
| "reward": 0.5867745792493224, | |
| "reward_std": 0.3615266978740692, | |
| "rewards/accuracy_reward": 0.5867745792493224, | |
| "rewards/format_reward": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 607.3701171875, | |
| "epoch": 0.17057569296375266, | |
| "grad_norm": 0.44140002131462097, | |
| "kl": 0.0072880744934082035, | |
| "learning_rate": 2.956412726139078e-06, | |
| "loss": 0.0586, | |
| "reward": 0.6774553880095482, | |
| "reward_std": 0.3047178775072098, | |
| "rewards/accuracy_reward": 0.6772321730852127, | |
| "rewards/format_reward": 0.00022321429569274187, | |
| "step": 10 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 613.5143089294434, | |
| "epoch": 0.255863539445629, | |
| "grad_norm": 0.2380095273256302, | |
| "kl": 0.009201431274414062, | |
| "learning_rate": 2.7836719084521715e-06, | |
| "loss": 0.0776, | |
| "reward": 0.7488839656114579, | |
| "reward_std": 0.24192895889282226, | |
| "rewards/accuracy_reward": 0.7488839656114579, | |
| "rewards/format_reward": 0.0, | |
| "step": 15 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 588.3607383728028, | |
| "epoch": 0.3411513859275053, | |
| "grad_norm": 0.23240025341510773, | |
| "kl": 0.004347991943359375, | |
| "learning_rate": 2.4946839873611927e-06, | |
| "loss": 0.0627, | |
| "reward": 0.7662946805357933, | |
| "reward_std": 0.2006999494507909, | |
| "rewards/accuracy_reward": 0.7662946805357933, | |
| "rewards/format_reward": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 596.7750228881836, | |
| "epoch": 0.42643923240938164, | |
| "grad_norm": 0.34883901476860046, | |
| "kl": 428.8045944213867, | |
| "learning_rate": 2.1156192081791355e-06, | |
| "loss": 9.7863, | |
| "reward": 0.7622768193483352, | |
| "reward_std": 0.18778416123241187, | |
| "rewards/accuracy_reward": 0.7622768193483352, | |
| "rewards/format_reward": 0.0, | |
| "step": 25 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 598.0908752441406, | |
| "epoch": 0.511727078891258, | |
| "grad_norm": 0.09987140446901321, | |
| "kl": 0.003299903869628906, | |
| "learning_rate": 1.6808050203829845e-06, | |
| "loss": 0.0314, | |
| "reward": 0.7569196805357933, | |
| "reward_std": 0.17923235325142742, | |
| "rewards/accuracy_reward": 0.7569196805357933, | |
| "rewards/format_reward": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 589.8533729553222, | |
| "epoch": 0.5970149253731343, | |
| "grad_norm": 0.15311484038829803, | |
| "kl": 0.0029478073120117188, | |
| "learning_rate": 1.2296174432791415e-06, | |
| "loss": 0.0343, | |
| "reward": 0.7401786059141159, | |
| "reward_std": 0.18878106027841568, | |
| "rewards/accuracy_reward": 0.7401786059141159, | |
| "rewards/format_reward": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 577.8973472595214, | |
| "epoch": 0.6823027718550106, | |
| "grad_norm": 0.4649916887283325, | |
| "kl": 0.004101181030273437, | |
| "learning_rate": 8.029152419343472e-07, | |
| "loss": 0.0392, | |
| "reward": 0.7642857506871223, | |
| "reward_std": 0.17380300173535942, | |
| "rewards/accuracy_reward": 0.7642857506871223, | |
| "rewards/format_reward": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 599.5591804504395, | |
| "epoch": 0.767590618336887, | |
| "grad_norm": 0.09541165828704834, | |
| "kl": 0.0041015625, | |
| "learning_rate": 4.3933982822017883e-07, | |
| "loss": 0.0381, | |
| "reward": 0.7560268223285675, | |
| "reward_std": 0.19011783925816417, | |
| "rewards/accuracy_reward": 0.7560268223285675, | |
| "rewards/format_reward": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 597.773908996582, | |
| "epoch": 0.8528784648187633, | |
| "grad_norm": 0.18816858530044556, | |
| "kl": 0.006811904907226563, | |
| "learning_rate": 1.718159615201853e-07, | |
| "loss": 0.0341, | |
| "reward": 0.7439732447266578, | |
| "reward_std": 0.18638971280306577, | |
| "rewards/accuracy_reward": 0.7439732447266578, | |
| "rewards/format_reward": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 590.9884216308594, | |
| "epoch": 0.9381663113006397, | |
| "grad_norm": 0.16142897307872772, | |
| "kl": 0.00367431640625, | |
| "learning_rate": 2.4570139579284723e-08, | |
| "loss": 0.0348, | |
| "reward": 0.784821467101574, | |
| "reward_std": 0.1836383467540145, | |
| "rewards/accuracy_reward": 0.784821467101574, | |
| "rewards/format_reward": 0.0, | |
| "step": 55 | |
| }, | |
| { | |
| "clip_ratio": 0.0, | |
| "completion_length": 581.6608149210612, | |
| "epoch": 0.9893390191897654, | |
| "kl": 0.003367741902669271, | |
| "reward": 0.7723214651147524, | |
| "reward_std": 0.17806254141032696, | |
| "rewards/accuracy_reward": 0.7723214651147524, | |
| "rewards/format_reward": 0.0, | |
| "step": 58, | |
| "total_flos": 0.0, | |
| "train_loss": 0.8815625811654046, | |
| "train_runtime": 25970.4468, | |
| "train_samples_per_second": 0.289, | |
| "train_steps_per_second": 0.002 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 58, | |
| "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": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |