| { | |
| "best_metric": 0.051282964646816254, | |
| "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": 1.4276326894760132, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7878, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.8709655404090881, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.2815, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 1.5181009769439697, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1882, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 8.139986038208008, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2897, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 7.810184001922607, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2317, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 3.843147039413452, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2651, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.6032278537750244, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1419, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.16680726408958435, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1682, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.18329288065433502, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.1014, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.6222877502441406, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0752, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.053070068359375, | |
| "eval_runtime": 1.4937, | |
| "eval_samples_per_second": 89.042, | |
| "eval_steps_per_second": 11.381, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 1.3583345413208008, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1348, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.5219836831092834, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.101, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.10262700170278549, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0645, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.7093880772590637, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0433, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.11172172427177429, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0205, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.08287137746810913, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0238, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.13058432936668396, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0236, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.15463727712631226, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.1165, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.44477012753486633, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0529, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.07537762075662613, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0261, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.051282964646816254, | |
| "eval_runtime": 1.0173, | |
| "eval_samples_per_second": 130.733, | |
| "eval_steps_per_second": 16.71, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.0681590884923935, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0264, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06600091606378555, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0131, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06940576434135437, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0126, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06425927579402924, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.07568276673555374, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0128, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05847496539354324, | |
| "learning_rate": 0.0, | |
| "loss": 0.0337, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.1249652250454976, | |
| "train_runtime": 82.8845, | |
| "train_samples_per_second": 49.901, | |
| "train_steps_per_second": 3.137 | |
| } | |
| ], | |
| "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 | |
| } | |