| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 10730, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.9319664492078286, | |
| "grad_norm": 14656.0576171875, | |
| "learning_rate": 0.0005993999999999999, | |
| "loss": 1.127, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.3059451842323442, | |
| "eval_loss": 4.31203031539917, | |
| "eval_runtime": 619.7282, | |
| "eval_samples_per_second": 57.07, | |
| "eval_steps_per_second": 1.785, | |
| "step": 1073 | |
| }, | |
| { | |
| "epoch": 1.8639328984156571, | |
| "grad_norm": 52568.05859375, | |
| "learning_rate": 0.0005383967112024666, | |
| "loss": 1.0263, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.30483972598313247, | |
| "eval_loss": 4.31310510635376, | |
| "eval_runtime": 619.4529, | |
| "eval_samples_per_second": 57.096, | |
| "eval_steps_per_second": 1.785, | |
| "step": 2146 | |
| }, | |
| { | |
| "epoch": 2.7958993476234855, | |
| "grad_norm": 31445.708984375, | |
| "learning_rate": 0.00047673175745118184, | |
| "loss": 1.0232, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.30701318338777167, | |
| "eval_loss": 4.279686450958252, | |
| "eval_runtime": 619.2893, | |
| "eval_samples_per_second": 57.111, | |
| "eval_steps_per_second": 1.786, | |
| "step": 3219 | |
| }, | |
| { | |
| "epoch": 3.7278657968313142, | |
| "grad_norm": 49623.2265625, | |
| "learning_rate": 0.00041506680369989723, | |
| "loss": 1.0192, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.3116210945713197, | |
| "eval_loss": 4.23567533493042, | |
| "eval_runtime": 619.8333, | |
| "eval_samples_per_second": 57.061, | |
| "eval_steps_per_second": 1.784, | |
| "step": 4292 | |
| }, | |
| { | |
| "epoch": 4.659832246039143, | |
| "grad_norm": 35489.96875, | |
| "learning_rate": 0.0003534018499486125, | |
| "loss": 1.0035, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.3179641308981197, | |
| "eval_loss": 4.164117813110352, | |
| "eval_runtime": 620.4234, | |
| "eval_samples_per_second": 57.006, | |
| "eval_steps_per_second": 1.783, | |
| "step": 5365 | |
| }, | |
| { | |
| "epoch": 5.591798695246971, | |
| "grad_norm": 31029.140625, | |
| "learning_rate": 0.00029173689619732784, | |
| "loss": 0.9839, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.32672640497607264, | |
| "eval_loss": 4.077977657318115, | |
| "eval_runtime": 620.2339, | |
| "eval_samples_per_second": 57.024, | |
| "eval_steps_per_second": 1.783, | |
| "step": 6438 | |
| }, | |
| { | |
| "epoch": 6.523765144454799, | |
| "grad_norm": 17874.236328125, | |
| "learning_rate": 0.00023007194244604312, | |
| "loss": 0.9631, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.3359358089460062, | |
| "eval_loss": 3.9749491214752197, | |
| "eval_runtime": 621.0925, | |
| "eval_samples_per_second": 56.945, | |
| "eval_steps_per_second": 1.781, | |
| "step": 7511 | |
| }, | |
| { | |
| "epoch": 7.4557315936626285, | |
| "grad_norm": 9566.0390625, | |
| "learning_rate": 0.00016840698869475846, | |
| "loss": 0.9374, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.34521553862967663, | |
| "eval_loss": 3.8823366165161133, | |
| "eval_runtime": 625.3406, | |
| "eval_samples_per_second": 56.558, | |
| "eval_steps_per_second": 1.769, | |
| "step": 8584 | |
| }, | |
| { | |
| "epoch": 8.387698042870456, | |
| "grad_norm": 8717.26953125, | |
| "learning_rate": 0.00010674203494347379, | |
| "loss": 0.9131, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.35117496506400026, | |
| "eval_loss": 3.818809986114502, | |
| "eval_runtime": 636.379, | |
| "eval_samples_per_second": 55.577, | |
| "eval_steps_per_second": 1.738, | |
| "step": 9657 | |
| }, | |
| { | |
| "epoch": 9.319664492078285, | |
| "grad_norm": 7765.57373046875, | |
| "learning_rate": 4.50770811921891e-05, | |
| "loss": 0.8966, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.35374796769200195, | |
| "eval_loss": 3.7910778522491455, | |
| "eval_runtime": 625.9975, | |
| "eval_samples_per_second": 56.499, | |
| "eval_steps_per_second": 1.767, | |
| "step": 10730 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 10730, | |
| "total_flos": 8.970939334656e+16, | |
| "train_loss": 0.9823163575583935, | |
| "train_runtime": 16060.4728, | |
| "train_samples_per_second": 21.377, | |
| "train_steps_per_second": 0.668 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 10730, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 8.970939334656e+16, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |