| { | |
| "best_metric": 0.014676429331302643, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-200", | |
| "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": 1.7548131942749023, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8278, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 3.528895378112793, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3232, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 4.442903518676758, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2161, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.861210286617279, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2552, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 8.467580795288086, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2294, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.317192316055298, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2226, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.34912338852882385, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.0866, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.17672394216060638, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0817, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.1419631987810135, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0779, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 2.7969260215759277, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0541, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.024188851937651634, | |
| "eval_runtime": 1.4458, | |
| "eval_samples_per_second": 91.994, | |
| "eval_steps_per_second": 11.759, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.6725256443023682, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.071, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 3.316457509994507, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1595, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.14894653856754303, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0907, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.10909508168697357, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0613, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.13060985505580902, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0203, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.08566407859325409, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0326, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.07223280519247055, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0182, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.0757821574807167, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0357, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 3.932257652282715, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0174, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06355487555265427, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.014, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.014676429331302643, | |
| "eval_runtime": 1.3175, | |
| "eval_samples_per_second": 100.95, | |
| "eval_steps_per_second": 12.903, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06115660443902016, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.028, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06409073621034622, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0116, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.05863337591290474, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0117, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.059554990381002426, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0113, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.05857035890221596, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.011, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.06412912160158157, | |
| "learning_rate": 0.0, | |
| "loss": 0.012, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11465750181904207, | |
| "train_runtime": 55.3297, | |
| "train_samples_per_second": 74.752, | |
| "train_steps_per_second": 4.699 | |
| } | |
| ], | |
| "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": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |