| { | |
| "best_metric": 0.07900336384773254, | |
| "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": 1.744454026222229, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7907, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.931182861328125, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.307, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 1.9947254657745361, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2054, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.5009201169013977, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.1954, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 4.2814106941223145, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2091, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.5581812858581543, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.3254, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 2.9047513008117676, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.0949, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.151845782995224, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0854, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.550007700920105, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0679, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.20410487055778503, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.1199, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9624060150375939, | |
| "eval_loss": 0.18091289699077606, | |
| "eval_runtime": 1.2837, | |
| "eval_samples_per_second": 103.609, | |
| "eval_steps_per_second": 13.243, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 4.168725490570068, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.149, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 2.4681456089019775, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.089, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 3.3101487159729004, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0753, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.7124854326248169, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0535, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.5007983446121216, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0649, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.1649583876132965, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0423, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.6068108677864075, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0161, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.7045071721076965, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0483, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.10050582140684128, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0175, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.061487264931201935, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9699248120300752, | |
| "eval_loss": 0.07900336384773254, | |
| "eval_runtime": 1.2415, | |
| "eval_samples_per_second": 107.13, | |
| "eval_steps_per_second": 13.693, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06725984066724777, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0135, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.059695739299058914, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0114, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06050793081521988, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.011, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06765594333410263, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0109, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.062375880777835846, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.011, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05423872172832489, | |
| "learning_rate": 0.0, | |
| "loss": 0.0105, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11683348210958334, | |
| "train_runtime": 86.5175, | |
| "train_samples_per_second": 47.805, | |
| "train_steps_per_second": 3.005 | |
| } | |
| ], | |
| "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 | |
| } | |