| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.6382978723404256, | |
| "eval_steps": 8, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02127659574468085, | |
| "eval_loss": 2.25822377204895, | |
| "eval_runtime": 2.323, | |
| "eval_samples_per_second": 8.61, | |
| "eval_steps_per_second": 4.305, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06382978723404255, | |
| "grad_norm": 1.6993662118911743, | |
| "learning_rate": 6e-05, | |
| "loss": 2.6448, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.1276595744680851, | |
| "grad_norm": 2.3986430168151855, | |
| "learning_rate": 0.00012, | |
| "loss": 2.2149, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.1702127659574468, | |
| "eval_loss": 1.8613002300262451, | |
| "eval_runtime": 2.3205, | |
| "eval_samples_per_second": 8.619, | |
| "eval_steps_per_second": 4.31, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.19148936170212766, | |
| "grad_norm": 1.783658742904663, | |
| "learning_rate": 0.00018, | |
| "loss": 1.8729, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.2553191489361702, | |
| "grad_norm": 1.8582178354263306, | |
| "learning_rate": 0.00019510565162951537, | |
| "loss": 2.215, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.3191489361702128, | |
| "grad_norm": 2.771054267883301, | |
| "learning_rate": 0.00017071067811865476, | |
| "loss": 1.673, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.3404255319148936, | |
| "eval_loss": 1.5222913026809692, | |
| "eval_runtime": 2.3686, | |
| "eval_samples_per_second": 8.444, | |
| "eval_steps_per_second": 4.222, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.3829787234042553, | |
| "grad_norm": 2.1258344650268555, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 1.7386, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.44680851063829785, | |
| "grad_norm": 3.365093231201172, | |
| "learning_rate": 8.435655349597689e-05, | |
| "loss": 1.8971, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.5106382978723404, | |
| "grad_norm": 2.163831949234009, | |
| "learning_rate": 4.12214747707527e-05, | |
| "loss": 1.5753, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.5106382978723404, | |
| "eval_loss": 1.422104835510254, | |
| "eval_runtime": 2.3784, | |
| "eval_samples_per_second": 8.409, | |
| "eval_steps_per_second": 4.204, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.574468085106383, | |
| "grad_norm": 4.278740882873535, | |
| "learning_rate": 1.0899347581163221e-05, | |
| "loss": 2.1696, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.6382978723404256, | |
| "grad_norm": 2.4858059883117676, | |
| "learning_rate": 0.0, | |
| "loss": 2.2444, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 3, | |
| "max_steps": 30, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4980196662312960.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |