| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.9987187700192184, | |
| "eval_steps": 1000, | |
| "global_step": 9360, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0003205128205128205, | |
| "grad_norm": 5.640012741088867, | |
| "learning_rate": 0.0, | |
| "loss": 2.1128, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.32051282051282054, | |
| "grad_norm": 3.1269683837890625, | |
| "learning_rate": 2.9775641025641028e-05, | |
| "loss": 1.6035, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.32051282051282054, | |
| "eval_cosine_accuracy": 0.9652960300445557, | |
| "eval_loss": 0.7382573485374451, | |
| "eval_runtime": 29.5289, | |
| "eval_samples_per_second": 322.024, | |
| "eval_steps_per_second": 1.287, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6410256410256411, | |
| "grad_norm": 3.351823091506958, | |
| "learning_rate": 2.6214387464387465e-05, | |
| "loss": 1.0575, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.6410256410256411, | |
| "eval_cosine_accuracy": 0.9737091064453125, | |
| "eval_loss": 0.6786530017852783, | |
| "eval_runtime": 28.6588, | |
| "eval_samples_per_second": 331.801, | |
| "eval_steps_per_second": 1.326, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.9615384615384616, | |
| "grad_norm": 2.9926202297210693, | |
| "learning_rate": 2.2653133903133903e-05, | |
| "loss": 1.8049, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.9615384615384616, | |
| "eval_cosine_accuracy": 0.9739194512367249, | |
| "eval_loss": 0.659064531326294, | |
| "eval_runtime": 28.2842, | |
| "eval_samples_per_second": 336.195, | |
| "eval_steps_per_second": 1.344, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.2818705957719412, | |
| "grad_norm": 2.9942617416381836, | |
| "learning_rate": 1.9091880341880344e-05, | |
| "loss": 1.2967, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.2818705957719412, | |
| "eval_cosine_accuracy": 0.9779156446456909, | |
| "eval_loss": 0.6036034226417542, | |
| "eval_runtime": 28.3407, | |
| "eval_samples_per_second": 335.525, | |
| "eval_steps_per_second": 1.341, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.6021780909673287, | |
| "grad_norm": 2.967597723007202, | |
| "learning_rate": 1.553774928774929e-05, | |
| "loss": 1.1471, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.6021780909673287, | |
| "eval_cosine_accuracy": 0.9798086285591125, | |
| "eval_loss": 0.5827310681343079, | |
| "eval_runtime": 28.4402, | |
| "eval_samples_per_second": 334.35, | |
| "eval_steps_per_second": 1.336, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.9224855861627161, | |
| "grad_norm": 3.271042823791504, | |
| "learning_rate": 1.1976495726495726e-05, | |
| "loss": 1.0816, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.9224855861627161, | |
| "eval_cosine_accuracy": 0.9800189137458801, | |
| "eval_loss": 0.590634822845459, | |
| "eval_runtime": 28.2426, | |
| "eval_samples_per_second": 336.69, | |
| "eval_steps_per_second": 1.345, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.2427930813581036, | |
| "grad_norm": 2.800858497619629, | |
| "learning_rate": 8.415242165242166e-06, | |
| "loss": 1.0567, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.2427930813581036, | |
| "eval_cosine_accuracy": 0.9794931411743164, | |
| "eval_loss": 0.5923933386802673, | |
| "eval_runtime": 28.3839, | |
| "eval_samples_per_second": 335.014, | |
| "eval_steps_per_second": 1.339, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.5631005765534915, | |
| "grad_norm": 3.2554521560668945, | |
| "learning_rate": 4.853988603988604e-06, | |
| "loss": 1.0046, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.5631005765534915, | |
| "eval_cosine_accuracy": 0.9803344011306763, | |
| "eval_loss": 0.582694947719574, | |
| "eval_runtime": 28.4177, | |
| "eval_samples_per_second": 334.615, | |
| "eval_steps_per_second": 1.337, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 2.883408071748879, | |
| "grad_norm": 6.852128028869629, | |
| "learning_rate": 1.2998575498575498e-06, | |
| "loss": 1.0313, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.883408071748879, | |
| "eval_cosine_accuracy": 0.9803344011306763, | |
| "eval_loss": 0.5849770903587341, | |
| "eval_runtime": 28.7312, | |
| "eval_samples_per_second": 330.965, | |
| "eval_steps_per_second": 1.323, | |
| "step": 9000 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 9360, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "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": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |