| { | |
| "best_metric": 0.09767168015241623, | |
| "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": 0.0549064539372921, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.011, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.053657062351703644, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.0099, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.04586618021130562, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.0085, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.03834086284041405, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.0074, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.0455704927444458, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.0081, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 4.641096591949463, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.0547, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 10.7581205368042, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.028, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.04066062346100807, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0387, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.042889270931482315, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0808, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.039594896137714386, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0316, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.09767168015241623, | |
| "eval_runtime": 1.2787, | |
| "eval_samples_per_second": 104.015, | |
| "eval_steps_per_second": 13.295, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.4228585660457611, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0357, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.03368207439780235, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0721, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.041465017944574356, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0286, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 4.102177619934082, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0239, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.726514458656311, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0085, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.025889277458190918, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0052, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.025063877925276756, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0046, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.024265997111797333, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0046, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 15.569941520690918, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0232, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.023741118609905243, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0106, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9624060150375939, | |
| "eval_loss": 0.19794000685214996, | |
| "eval_runtime": 1.6975, | |
| "eval_samples_per_second": 78.351, | |
| "eval_steps_per_second": 10.015, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.03866928815841675, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0377, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.023310711607336998, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0221, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.02209044247865677, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0057, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.020965544506907463, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.004, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.02124304138123989, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0039, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.0217234306037426, | |
| "learning_rate": 0.0, | |
| "loss": 0.0039, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.02204044573009014, | |
| "train_runtime": 91.5843, | |
| "train_samples_per_second": 45.161, | |
| "train_steps_per_second": 2.839 | |
| } | |
| ], | |
| "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 | |
| } | |