| { | |
| "best_metric": 0.07552934437990189, | |
| "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": 1.4994887113571167, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8034, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.9837390184402466, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3034, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 3.1711220741271973, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2491, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.84904944896698, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2425, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 1.688535213470459, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2365, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.4718668460845947, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2497, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 1.9351365566253662, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1205, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.3785780370235443, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0672, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 11.389372825622559, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0761, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 4.747922897338867, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0745, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.0811515748500824, | |
| "eval_runtime": 1.37, | |
| "eval_samples_per_second": 97.077, | |
| "eval_steps_per_second": 12.408, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.1738795042037964, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1314, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 6.780663967132568, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.097, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.32149961590766907, | |
| "learning_rate": 0.0001, | |
| "loss": 0.139, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.28285670280456543, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0961, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 2.1620373725891113, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0271, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 2.2063798904418945, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0334, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 7.840737819671631, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.1151, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.16310855746269226, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0399, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.08073098212480545, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.039, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06902778148651123, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0288, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9699248120300752, | |
| "eval_loss": 0.07552934437990189, | |
| "eval_runtime": 1.2571, | |
| "eval_samples_per_second": 105.797, | |
| "eval_steps_per_second": 13.523, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.08519002795219421, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.015, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.0743965357542038, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0275, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06629299372434616, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0121, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06420692056417465, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0123, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.06406693905591965, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0119, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.062306832522153854, | |
| "learning_rate": 0.0, | |
| "loss": 0.0122, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.1254174125309174, | |
| "train_runtime": 97.1379, | |
| "train_samples_per_second": 42.579, | |
| "train_steps_per_second": 2.677 | |
| } | |
| ], | |
| "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 | |
| } | |