| { |
| "best_metric": 5.178476810455322, |
| "best_model_checkpoint": "./combined/context/checkpoint-500", |
| "epoch": 3.0, |
| "eval_steps": 500, |
| "global_step": 1806, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.8305647840531561, |
| "grad_norm": 7.5, |
| "learning_rate": 4.1125145874283014e-05, |
| "loss": 1.276, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.8305647840531561, |
| "eval_accuracy": 0.24484439445069367, |
| "eval_loss": 5.178476810455322, |
| "eval_runtime": 3.9962, |
| "eval_samples_per_second": 73.569, |
| "eval_steps_per_second": 2.002, |
| "step": 500 |
| }, |
| { |
| "epoch": 1.6611295681063123, |
| "grad_norm": 3.765625, |
| "learning_rate": 2.080162635735251e-05, |
| "loss": 0.5321, |
| "step": 1000 |
| }, |
| { |
| "epoch": 1.6611295681063123, |
| "eval_accuracy": 0.2564947238738015, |
| "eval_loss": 5.263692855834961, |
| "eval_runtime": 2.4643, |
| "eval_samples_per_second": 119.305, |
| "eval_steps_per_second": 3.246, |
| "step": 1000 |
| }, |
| { |
| "epoch": 2.4916943521594686, |
| "grad_norm": 4.1875, |
| "learning_rate": 3.458903131922134e-06, |
| "loss": 0.3538, |
| "step": 1500 |
| }, |
| { |
| "epoch": 2.4916943521594686, |
| "eval_accuracy": 0.2552091702822861, |
| "eval_loss": 5.43364953994751, |
| "eval_runtime": 2.4877, |
| "eval_samples_per_second": 118.182, |
| "eval_steps_per_second": 3.216, |
| "step": 1500 |
| }, |
| { |
| "epoch": 3.0, |
| "step": 1806, |
| "total_flos": 4.300357860694426e+16, |
| "train_loss": 0.6495532799400761, |
| "train_runtime": 3174.7304, |
| "train_samples_per_second": 2.275, |
| "train_steps_per_second": 0.569 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 1806, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "EarlyStoppingCallback": { |
| "args": { |
| "early_stopping_patience": 3, |
| "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": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 4.300357860694426e+16, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
