| { | |
| "best_metric": 0.38773971796035767, | |
| "best_model_checkpoint": "./vit-base-brain-tumor-detection2/checkpoint-1500", | |
| "epoch": 30.0, | |
| "eval_steps": 500, | |
| "global_step": 1920, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.78125, | |
| "grad_norm": 1.7783442735671997, | |
| "learning_rate": 2.5000000000000004e-07, | |
| "loss": 1.3686, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "grad_norm": 1.8243844509124756, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 1.3468, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.34375, | |
| "grad_norm": 2.272806167602539, | |
| "learning_rate": 7.5e-07, | |
| "loss": 1.2931, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "grad_norm": 1.0346544981002808, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 1.1861, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.90625, | |
| "grad_norm": 0.897363007068634, | |
| "learning_rate": 1.25e-06, | |
| "loss": 1.092, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.6875, | |
| "grad_norm": 1.0376484394073486, | |
| "learning_rate": 1.5e-06, | |
| "loss": 1.0504, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 5.46875, | |
| "grad_norm": 1.2069748640060425, | |
| "learning_rate": 1.75e-06, | |
| "loss": 0.993, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "grad_norm": 1.41449773311615, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.9428, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 7.03125, | |
| "grad_norm": 3.5506372451782227, | |
| "learning_rate": 2.25e-06, | |
| "loss": 0.9065, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 7.8125, | |
| "grad_norm": 2.408358573913574, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.8758, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 7.8125, | |
| "eval_accuracy": 0.6181640625, | |
| "eval_loss": 0.891155481338501, | |
| "eval_runtime": 5.9944, | |
| "eval_samples_per_second": 170.826, | |
| "eval_steps_per_second": 21.353, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 8.59375, | |
| "grad_norm": 4.127248764038086, | |
| "learning_rate": 2.7500000000000004e-06, | |
| "loss": 0.8371, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 9.375, | |
| "grad_norm": 4.548064231872559, | |
| "learning_rate": 3e-06, | |
| "loss": 0.7951, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 10.15625, | |
| "grad_norm": 7.226322174072266, | |
| "learning_rate": 3.2500000000000002e-06, | |
| "loss": 0.7666, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 10.9375, | |
| "grad_norm": 4.653890132904053, | |
| "learning_rate": 3.5e-06, | |
| "loss": 0.7219, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 11.71875, | |
| "grad_norm": 3.7779295444488525, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.6926, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "grad_norm": 3.6519362926483154, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.646, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 13.28125, | |
| "grad_norm": 24.227773666381836, | |
| "learning_rate": 4.25e-06, | |
| "loss": 0.6269, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 14.0625, | |
| "grad_norm": 8.960350036621094, | |
| "learning_rate": 4.5e-06, | |
| "loss": 0.5882, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 14.84375, | |
| "grad_norm": 6.60162878036499, | |
| "learning_rate": 4.75e-06, | |
| "loss": 0.544, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 15.625, | |
| "grad_norm": 10.583850860595703, | |
| "learning_rate": 5e-06, | |
| "loss": 0.5256, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 15.625, | |
| "eval_accuracy": 0.7607421875, | |
| "eval_loss": 0.6331567764282227, | |
| "eval_runtime": 5.4782, | |
| "eval_samples_per_second": 186.924, | |
| "eval_steps_per_second": 23.365, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 16.40625, | |
| "grad_norm": 3.9983010292053223, | |
| "learning_rate": 4.999850432733413e-06, | |
| "loss": 0.4768, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 17.1875, | |
| "grad_norm": 12.768200874328613, | |
| "learning_rate": 4.999401748829942e-06, | |
| "loss": 0.4863, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 17.96875, | |
| "grad_norm": 14.71330451965332, | |
| "learning_rate": 4.9986540019763296e-06, | |
| "loss": 0.4181, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 18.75, | |
| "grad_norm": 9.602484703063965, | |
| "learning_rate": 4.997607281643338e-06, | |
| "loss": 0.3901, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 19.53125, | |
| "grad_norm": 27.293039321899414, | |
| "learning_rate": 4.996261713075046e-06, | |
| "loss": 0.3496, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 20.3125, | |
| "grad_norm": 3.6233201026916504, | |
| "learning_rate": 4.994617457273862e-06, | |
| "loss": 0.3312, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 21.09375, | |
| "grad_norm": 11.52304458618164, | |
| "learning_rate": 4.992674710981266e-06, | |
| "loss": 0.2929, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 21.875, | |
| "grad_norm": 6.805290222167969, | |
| "learning_rate": 4.990433706654258e-06, | |
| "loss": 0.2653, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 22.65625, | |
| "grad_norm": 15.234956741333008, | |
| "learning_rate": 4.987894712437552e-06, | |
| "loss": 0.2635, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 23.4375, | |
| "grad_norm": 3.9582715034484863, | |
| "learning_rate": 4.985058032131488e-06, | |
| "loss": 0.2247, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 23.4375, | |
| "eval_accuracy": 0.8935546875, | |
| "eval_loss": 0.38773971796035767, | |
| "eval_runtime": 5.9931, | |
| "eval_samples_per_second": 170.863, | |
| "eval_steps_per_second": 21.358, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 24.21875, | |
| "grad_norm": 3.0134124755859375, | |
| "learning_rate": 4.98192400515568e-06, | |
| "loss": 0.1965, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 5.4471540451049805, | |
| "learning_rate": 4.978564534206973e-06, | |
| "loss": 0.193, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 25.78125, | |
| "grad_norm": 11.847530364990234, | |
| "learning_rate": 4.974842901390282e-06, | |
| "loss": 0.1709, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 26.5625, | |
| "grad_norm": 4.015501022338867, | |
| "learning_rate": 4.970825144183195e-06, | |
| "loss": 0.158, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 27.34375, | |
| "grad_norm": 4.216943264007568, | |
| "learning_rate": 4.966511743325682e-06, | |
| "loss": 0.1452, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 28.125, | |
| "grad_norm": 23.560443878173828, | |
| "learning_rate": 4.961903214932604e-06, | |
| "loss": 0.1288, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 28.90625, | |
| "grad_norm": 0.9162831902503967, | |
| "learning_rate": 4.957000110431956e-06, | |
| "loss": 0.1195, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 29.6875, | |
| "grad_norm": 2.046557664871216, | |
| "learning_rate": 4.95180301649889e-06, | |
| "loss": 0.1002, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "step": 1920, | |
| "total_flos": 9.52238637305561e+18, | |
| "train_loss": 0.5872499863306682, | |
| "train_runtime": 1494.2494, | |
| "train_samples_per_second": 82.235, | |
| "train_steps_per_second": 1.285 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 1920, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "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": 9.52238637305561e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |