| { | |
| "best_metric": 1.2398513555526733, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-10", | |
| "epoch": 0.0006903574325607083, | |
| "eval_steps": 5, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 6.903574325607083e-05, | |
| "eval_loss": 1.6781178712844849, | |
| "eval_runtime": 838.6739, | |
| "eval_samples_per_second": 7.272, | |
| "eval_steps_per_second": 3.637, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.00020710722976821248, | |
| "grad_norm": 0.9852650761604309, | |
| "learning_rate": 6e-05, | |
| "loss": 1.4438, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00034517871628035417, | |
| "eval_loss": 1.6058242321014404, | |
| "eval_runtime": 842.8088, | |
| "eval_samples_per_second": 7.237, | |
| "eval_steps_per_second": 3.619, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.00041421445953642496, | |
| "grad_norm": 1.22545325756073, | |
| "learning_rate": 0.00012, | |
| "loss": 1.5041, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0006213216893046374, | |
| "grad_norm": 1.2531858682632446, | |
| "learning_rate": 0.00018, | |
| "loss": 1.5687, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0006903574325607083, | |
| "eval_loss": 1.2398513555526733, | |
| "eval_runtime": 842.4208, | |
| "eval_samples_per_second": 7.24, | |
| "eval_steps_per_second": 3.621, | |
| "step": 10 | |
| } | |
| ], | |
| "logging_steps": 3, | |
| "max_steps": 30, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 1, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3987713405485056.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |