| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.03088803088803089, | |
| "eval_steps": 8, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010296010296010295, | |
| "eval_loss": 8.052611351013184, | |
| "eval_runtime": 32.5782, | |
| "eval_samples_per_second": 12.554, | |
| "eval_steps_per_second": 6.293, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.003088803088803089, | |
| "grad_norm": 2.1777353286743164, | |
| "learning_rate": 6e-05, | |
| "loss": 8.1384, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.006177606177606178, | |
| "grad_norm": 2.114084482192993, | |
| "learning_rate": 0.00012, | |
| "loss": 7.8177, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.008236808236808236, | |
| "eval_loss": 6.961575031280518, | |
| "eval_runtime": 31.2926, | |
| "eval_samples_per_second": 13.07, | |
| "eval_steps_per_second": 6.551, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.009266409266409266, | |
| "grad_norm": 1.4109786748886108, | |
| "learning_rate": 0.00018, | |
| "loss": 7.3149, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.012355212355212355, | |
| "grad_norm": 1.8765102624893188, | |
| "learning_rate": 0.00019510565162951537, | |
| "loss": 6.4621, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.015444015444015444, | |
| "grad_norm": 2.23150634765625, | |
| "learning_rate": 0.00017071067811865476, | |
| "loss": 6.2479, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.016473616473616472, | |
| "eval_loss": 5.598282337188721, | |
| "eval_runtime": 31.4187, | |
| "eval_samples_per_second": 13.018, | |
| "eval_steps_per_second": 6.525, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.018532818532818532, | |
| "grad_norm": 1.7026886940002441, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 5.5097, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.021621621621621623, | |
| "grad_norm": 1.913030743598938, | |
| "learning_rate": 8.435655349597689e-05, | |
| "loss": 4.9087, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.02471042471042471, | |
| "grad_norm": 1.5813980102539062, | |
| "learning_rate": 4.12214747707527e-05, | |
| "loss": 4.9872, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.02471042471042471, | |
| "eval_loss": 4.9767022132873535, | |
| "eval_runtime": 31.5052, | |
| "eval_samples_per_second": 12.982, | |
| "eval_steps_per_second": 6.507, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.027799227799227798, | |
| "grad_norm": 1.6165424585342407, | |
| "learning_rate": 1.0899347581163221e-05, | |
| "loss": 5.1278, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.03088803088803089, | |
| "grad_norm": 1.4993691444396973, | |
| "learning_rate": 0.0, | |
| "loss": 4.9896, | |
| "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": 1.0747719909900288e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |