| { | |
| "best_metric": 0.020068835467100143, | |
| "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": 0.026720570400357246, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.0057, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.024325191974639893, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.0048, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.02510487474501133, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.0044, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 5.864213943481445, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.0082, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.2858871519565582, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.0225, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 0.10690926015377045, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.0706, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.036041658371686935, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.0422, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.029629569500684738, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0341, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.036782171577215195, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0288, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.021833963692188263, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.038, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.09355173259973526, | |
| "eval_runtime": 1.2113, | |
| "eval_samples_per_second": 109.796, | |
| "eval_steps_per_second": 14.034, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 4.63957405090332, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1793, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.08258672058582306, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1784, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.04464903846383095, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0927, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 9.66904067993164, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0598, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.23427046835422516, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0462, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.02267806977033615, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0198, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.5346509218215942, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.005, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.021317703649401665, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0435, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.0290632713586092, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0037, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.0228396225720644, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0037, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.020068835467100143, | |
| "eval_runtime": 1.3217, | |
| "eval_samples_per_second": 100.628, | |
| "eval_steps_per_second": 12.862, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.016700690612196922, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0036, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.017470838502049446, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0031, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.1028652936220169, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0031, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.01951315440237522, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.003, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.016699790954589844, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0029, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.015097931027412415, | |
| "learning_rate": 0.0, | |
| "loss": 0.003, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.03500677141575859, | |
| "train_runtime": 87.5927, | |
| "train_samples_per_second": 47.219, | |
| "train_steps_per_second": 2.968 | |
| } | |
| ], | |
| "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 | |
| } | |