| { | |
| "best_metric": 0.008718447759747505, | |
| "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": 0.021665610373020172, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.0042, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.03046106919646263, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.0602, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.4583040475845337, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.0375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.023877069354057312, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.043, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 7.474393844604492, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.0716, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 0.031120434403419495, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2261, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 7.339480400085449, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.2382, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.03332442045211792, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0961, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.3657234013080597, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0183, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 2.6633493900299072, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0429, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.008718447759747505, | |
| "eval_runtime": 1.2803, | |
| "eval_samples_per_second": 103.878, | |
| "eval_steps_per_second": 13.278, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.25266826152801514, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0249, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.0192479956895113, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0216, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.04104306921362877, | |
| "learning_rate": 0.0001, | |
| "loss": 0.009, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 10.378777503967285, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0192, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.017906097695231438, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0219, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.02769538387656212, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.01, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.02914060465991497, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0043, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.014721982181072235, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.003, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.014472324401140213, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0028, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.014757129363715649, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0027, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.038653451949357986, | |
| "eval_runtime": 1.2802, | |
| "eval_samples_per_second": 103.889, | |
| "eval_steps_per_second": 13.279, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.014285999350249767, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0026, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.01450322289019823, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0026, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.014540684409439564, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0025, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.014169692993164062, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0024, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.013526924885809422, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0027, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.017109964042901993, | |
| "learning_rate": 0.0, | |
| "loss": 0.0027, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.217496245940552e+17, | |
| "train_loss": 0.03741855882108212, | |
| "train_runtime": 83.1006, | |
| "train_samples_per_second": 49.771, | |
| "train_steps_per_second": 3.129 | |
| } | |
| ], | |
| "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": 3.217496245940552e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |