| { | |
| "best_metric": 0.04180770367383957, | |
| "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": 1.4700136184692383, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7811, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.1012625694274902, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3056, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.9217509627342224, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2183, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 7.513153553009033, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.3319, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 5.002466201782227, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.3105, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 1.085290551185608, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2134, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.5019209980964661, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1534, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.18690814077854156, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.088, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.7312166094779968, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0624, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 1.807482361793518, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0731, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.06501816213130951, | |
| "eval_runtime": 1.3173, | |
| "eval_samples_per_second": 100.967, | |
| "eval_steps_per_second": 12.905, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 2.6930549144744873, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0564, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.09813439100980759, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0661, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 10.424873352050781, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0549, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 4.380281448364258, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0441, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 5.875505447387695, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0485, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.07221483439207077, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0145, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.09319509565830231, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0159, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.164328470826149, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0217, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06507332623004913, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0285, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.055057354271411896, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0128, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.04180770367383957, | |
| "eval_runtime": 1.271, | |
| "eval_samples_per_second": 104.641, | |
| "eval_steps_per_second": 13.375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.07171156257390976, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0106, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.05258367210626602, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0103, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.05462221801280975, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0101, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.05226493626832962, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0098, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.05405202507972717, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0097, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.04935304448008537, | |
| "learning_rate": 0.0, | |
| "loss": 0.0097, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11389572775134674, | |
| "train_runtime": 91.8479, | |
| "train_samples_per_second": 45.031, | |
| "train_steps_per_second": 2.831 | |
| } | |
| ], | |
| "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 | |
| } | |