| { | |
| "best_metric": 0.05026606097817421, | |
| "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": 1.2843873500823975, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8434, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 3.7151455879211426, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3085, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.8530910611152649, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.248, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.1804994344711304, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2334, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.5084353089332581, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1626, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 6.173972129821777, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2912, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 2.3577582836151123, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1226, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.14449569582939148, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0796, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 4.846245288848877, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0813, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 3.6284821033477783, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0645, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9699248120300752, | |
| "eval_loss": 0.09509244561195374, | |
| "eval_runtime": 1.3146, | |
| "eval_samples_per_second": 101.169, | |
| "eval_steps_per_second": 12.931, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 2.209476947784424, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0932, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.10745280981063843, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1311, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.1176445409655571, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0584, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.11741472780704498, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0464, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.5773084163665771, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0727, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.5665645003318787, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0669, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.20813684165477753, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0246, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.12770630419254303, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0285, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06980925053358078, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0185, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.07084397971630096, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0316, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.05026606097817421, | |
| "eval_runtime": 2.2826, | |
| "eval_samples_per_second": 58.267, | |
| "eval_steps_per_second": 7.448, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.061542343348264694, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0128, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.4158130884170532, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0173, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.07470071315765381, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0119, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.0632583349943161, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0391, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.46634915471076965, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.012, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.056492287665605545, | |
| "learning_rate": 0.0, | |
| "loss": 0.0116, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11968801841139794, | |
| "train_runtime": 105.0874, | |
| "train_samples_per_second": 39.358, | |
| "train_steps_per_second": 2.474 | |
| } | |
| ], | |
| "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 | |
| } | |