| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 4395, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.3412969283276451, | |
| "grad_norm": 1.9755513668060303, | |
| "learning_rate": 4.431171786120592e-05, | |
| "loss": 7.0043, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6825938566552902, | |
| "grad_norm": 2.0670814514160156, | |
| "learning_rate": 3.862343572241183e-05, | |
| "loss": 6.3637, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.023890784982935, | |
| "grad_norm": 1.9433873891830444, | |
| "learning_rate": 3.293515358361775e-05, | |
| "loss": 6.1006, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.36518771331058, | |
| "grad_norm": 2.0810210704803467, | |
| "learning_rate": 2.7246871444823664e-05, | |
| "loss": 5.8529, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.7064846416382253, | |
| "grad_norm": 2.127770185470581, | |
| "learning_rate": 2.1558589306029582e-05, | |
| "loss": 5.7272, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.04778156996587, | |
| "grad_norm": 2.3847897052764893, | |
| "learning_rate": 1.5870307167235497e-05, | |
| "loss": 5.5975, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.3890784982935154, | |
| "grad_norm": 2.466097354888916, | |
| "learning_rate": 1.0182025028441412e-05, | |
| "loss": 5.4416, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.73037542662116, | |
| "grad_norm": 2.4412147998809814, | |
| "learning_rate": 4.493742889647327e-06, | |
| "loss": 5.3805, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 4395, | |
| "total_flos": 2.60103732553728e+16, | |
| "train_loss": 5.880197445605802, | |
| "train_runtime": 24831.4743, | |
| "train_samples_per_second": 2.831, | |
| "train_steps_per_second": 0.177 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 4395, | |
| "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": 2.60103732553728e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |