| { | |
| "best_metric": 0.03209061175584793, | |
| "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": 2.2777774333953857, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.852, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.2117365598678589, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3376, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 2.5329957008361816, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.267, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 3.198817014694214, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.3698, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 4.2858476638793945, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.3999, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.978250741958618, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2925, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.26028183102607727, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1744, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.4725353419780731, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1176, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 6.265408992767334, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0748, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.2950698137283325, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0581, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.09987714886665344, | |
| "eval_runtime": 1.2871, | |
| "eval_samples_per_second": 103.331, | |
| "eval_steps_per_second": 13.208, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 5.468327045440674, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1112, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 2.1302363872528076, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0505, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.09734752774238586, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0658, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 3.8837532997131348, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0604, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.12428873777389526, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0435, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.16831739246845245, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0324, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.0807306244969368, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0156, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.4099869430065155, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0154, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.08725185692310333, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0141, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06723841279745102, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0129, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.03209061175584793, | |
| "eval_runtime": 1.3501, | |
| "eval_samples_per_second": 98.508, | |
| "eval_steps_per_second": 12.591, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06230653077363968, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0126, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.08260240405797958, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0126, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06481257826089859, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0122, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.07664787024259567, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0116, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.06441407650709152, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0118, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.07199066877365112, | |
| "learning_rate": 0.0, | |
| "loss": 0.0119, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.13223712255175296, | |
| "train_runtime": 92.9237, | |
| "train_samples_per_second": 44.51, | |
| "train_steps_per_second": 2.798 | |
| } | |
| ], | |
| "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 | |
| } | |