| { | |
| "best_metric": 0.03705849498510361, | |
| "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": 2.0439679622650146, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8763, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.124077796936035, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3952, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.6128961443901062, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2365, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.7809900045394897, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2401, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 3.320781707763672, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2405, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 3.576935291290283, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.269, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 2.830223321914673, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1713, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.6125717163085938, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0729, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 1.0431435108184814, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0601, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 8.226062774658203, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0928, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.062415093183517456, | |
| "eval_runtime": 1.2568, | |
| "eval_samples_per_second": 105.823, | |
| "eval_steps_per_second": 13.526, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.6293368339538574, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0773, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 1.033178448677063, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1894, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.12555956840515137, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0654, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 2.768048048019409, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.1294, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.1214001402258873, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0226, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.08666225522756577, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0177, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.15021829307079315, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0414, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 1.330576777458191, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0871, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.08042296022176743, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.032, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.07944784313440323, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0221, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.03705849498510361, | |
| "eval_runtime": 1.2806, | |
| "eval_samples_per_second": 103.855, | |
| "eval_steps_per_second": 13.275, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06947614997625351, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0141, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06896186619997025, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.016, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06684739887714386, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0131, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.0635317862033844, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0127, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.06400393694639206, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0123, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.06617986410856247, | |
| "learning_rate": 0.0, | |
| "loss": 0.0132, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.13155142492972888, | |
| "train_runtime": 88.2657, | |
| "train_samples_per_second": 46.859, | |
| "train_steps_per_second": 2.946 | |
| } | |
| ], | |
| "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 | |
| } | |