| { | |
| "best_metric": 0.023933812975883484, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100", | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 260, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15384615384615385, | |
| "grad_norm": 1.524735450744629, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8025, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.1621341705322266, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3034, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 3.7148265838623047, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1962, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.666541337966919, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2723, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.3225570619106293, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1386, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.0221970081329346, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2213, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.27449536323547363, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1468, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.14676600694656372, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1299, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.13736337423324585, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0451, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 1.207114815711975, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.043, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.023933812975883484, | |
| "eval_runtime": 1.306, | |
| "eval_samples_per_second": 101.837, | |
| "eval_steps_per_second": 13.017, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.6053730249404907, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0338, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.09518508613109589, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0606, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.0819147452712059, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1016, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 2.0847203731536865, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0605, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.346297025680542, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0411, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.10980492830276489, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0619, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.5999782085418701, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0417, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 2.092048406600952, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0867, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 5.020918369293213, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0466, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.0641295537352562, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0141, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.04076213389635086, | |
| "eval_runtime": 1.2335, | |
| "eval_samples_per_second": 107.82, | |
| "eval_steps_per_second": 13.782, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.060642510652542114, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0121, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.05819069966673851, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0116, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06333259493112564, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0118, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.05809474363923073, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0111, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.16743071377277374, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0112, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.052998460829257965, | |
| "learning_rate": 0.0, | |
| "loss": 0.0119, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11221663410273883, | |
| "train_runtime": 83.3793, | |
| "train_samples_per_second": 49.605, | |
| "train_steps_per_second": 3.118 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |