| { | |
| "best_metric": 0.11481206119060516, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100", | |
| "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": 1.621594786643982, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7949, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.8978638052940369, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3326, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 4.217480659484863, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2518, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 3.5624806880950928, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.3291, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 3.0238661766052246, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2966, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 4.109220504760742, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.241, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.23048052191734314, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1185, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.15795736014842987, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1088, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.13900934159755707, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.1013, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.1411190629005432, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0851, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9548872180451128, | |
| "eval_loss": 0.11481206119060516, | |
| "eval_runtime": 1.8544, | |
| "eval_samples_per_second": 71.721, | |
| "eval_steps_per_second": 9.167, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 1.4182195663452148, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1494, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.4042862355709076, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1052, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.24075759947299957, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0793, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.09303337335586548, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0242, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 6.949472427368164, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0501, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.0855720266699791, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0155, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.07753770798444748, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0256, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.23120836913585663, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0425, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 3.7650270462036133, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0208, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06337559968233109, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0123, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9624060150375939, | |
| "eval_loss": 0.14262861013412476, | |
| "eval_runtime": 1.8782, | |
| "eval_samples_per_second": 70.811, | |
| "eval_steps_per_second": 9.051, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.10766848176717758, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0123, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.05954992398619652, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.04, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.07272188365459442, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0119, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.05848537012934685, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0146, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.05753796920180321, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0109, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05411148443818092, | |
| "learning_rate": 0.0, | |
| "loss": 0.011, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.12636138025957805, | |
| "train_runtime": 100.9353, | |
| "train_samples_per_second": 40.977, | |
| "train_steps_per_second": 2.576 | |
| } | |
| ], | |
| "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 | |
| } | |