| { | |
| "best_metric": 0.31034482758620696, | |
| "best_model_checkpoint": "./models/checkpoint/checkpoint-68", | |
| "epoch": 9.275862068965518, | |
| "eval_steps": 500, | |
| "global_step": 80, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.9195402298850575, | |
| "eval_f1": 0.2727272727272727, | |
| "eval_loss": 9.31263256072998, | |
| "eval_runtime": 0.1958, | |
| "eval_samples_per_second": 51.067, | |
| "eval_steps_per_second": 25.534, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9195402298850575, | |
| "step": 8, | |
| "total_flos": 20939936811486.0, | |
| "train_loss": 9.918171882629395, | |
| "train_runtime": 21.4307, | |
| "train_samples_per_second": 4.06, | |
| "train_steps_per_second": 0.373 | |
| }, | |
| { | |
| "epoch": 0.9195402298850575, | |
| "step": 8, | |
| "total_flos": 20939936811486.0, | |
| "train_loss": 0.0, | |
| "train_runtime": 0.5294, | |
| "train_samples_per_second": 164.352, | |
| "train_steps_per_second": 15.113 | |
| }, | |
| { | |
| "epoch": 1.9195402298850575, | |
| "eval_f1": 0.2530120481927711, | |
| "eval_loss": 9.20499038696289, | |
| "eval_runtime": 0.1794, | |
| "eval_samples_per_second": 55.739, | |
| "eval_steps_per_second": 27.87, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.954022988505747, | |
| "eval_f1": 0.25842696629213485, | |
| "eval_loss": 8.635601997375488, | |
| "eval_runtime": 0.1743, | |
| "eval_samples_per_second": 57.382, | |
| "eval_steps_per_second": 28.691, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.9885057471264367, | |
| "eval_f1": 0.30409356725146197, | |
| "eval_loss": 8.560417175292969, | |
| "eval_runtime": 0.1737, | |
| "eval_samples_per_second": 57.572, | |
| "eval_steps_per_second": 28.786, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.908045977011494, | |
| "eval_f1": 0.3076923076923077, | |
| "eval_loss": 8.47392463684082, | |
| "eval_runtime": 0.1856, | |
| "eval_samples_per_second": 53.876, | |
| "eval_steps_per_second": 26.938, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.942528735632184, | |
| "eval_f1": 0.30857142857142855, | |
| "eval_loss": 8.22740364074707, | |
| "eval_runtime": 0.1773, | |
| "eval_samples_per_second": 56.411, | |
| "eval_steps_per_second": 28.205, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.977011494252873, | |
| "eval_f1": 0.3, | |
| "eval_loss": 8.102139472961426, | |
| "eval_runtime": 0.1819, | |
| "eval_samples_per_second": 54.979, | |
| "eval_steps_per_second": 27.49, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.896551724137931, | |
| "eval_f1": 0.31034482758620696, | |
| "eval_loss": 7.982325553894043, | |
| "eval_runtime": 0.1721, | |
| "eval_samples_per_second": 58.115, | |
| "eval_steps_per_second": 29.057, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.931034482758621, | |
| "eval_f1": 0.30681818181818177, | |
| "eval_loss": 7.987342834472656, | |
| "eval_runtime": 0.1737, | |
| "eval_samples_per_second": 57.583, | |
| "eval_steps_per_second": 28.792, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 9.275862068965518, | |
| "eval_f1": 0.30681818181818177, | |
| "eval_loss": 7.902113437652588, | |
| "eval_runtime": 0.1822, | |
| "eval_samples_per_second": 54.889, | |
| "eval_steps_per_second": 27.444, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.275862068965518, | |
| "step": 80, | |
| "total_flos": 202609940734866.0, | |
| "train_loss": 7.418544769287109, | |
| "train_runtime": 132.4614, | |
| "train_samples_per_second": 6.568, | |
| "train_steps_per_second": 0.604 | |
| }, | |
| { | |
| "epoch": 9.275862068965518, | |
| "step": 80, | |
| "total_flos": 202609940734866.0, | |
| "train_loss": 0.0, | |
| "train_runtime": 0.245, | |
| "train_samples_per_second": 355.061, | |
| "train_steps_per_second": 32.649 | |
| }, | |
| { | |
| "epoch": 9.275862068965518, | |
| "step": 80, | |
| "total_flos": 202609940734866.0, | |
| "train_loss": 0.0, | |
| "train_runtime": 0.244, | |
| "train_samples_per_second": 356.519, | |
| "train_steps_per_second": 32.783 | |
| }, | |
| { | |
| "epoch": 9.275862068965518, | |
| "step": 80, | |
| "total_flos": 202609940734866.0, | |
| "train_loss": 0.0, | |
| "train_runtime": 0.2425, | |
| "train_samples_per_second": 358.716, | |
| "train_steps_per_second": 32.985 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 8, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 202609940734866.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |