| { | |
| "best_metric": 0.017012804746627808, | |
| "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": 1.7155200242996216, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8423, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.7316235303878784, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3193, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.5514141321182251, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1967, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 2.297652006149292, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2321, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 1.2160619497299194, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1504, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 0.6102649569511414, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2133, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.2573701739311218, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.181, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.16006311774253845, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0758, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 4.630112171173096, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0691, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 6.319807052612305, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0921, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.06150772050023079, | |
| "eval_runtime": 1.2973, | |
| "eval_samples_per_second": 102.519, | |
| "eval_steps_per_second": 13.104, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.8589983582496643, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1085, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 1.9701812267303467, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1495, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.1063208058476448, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0509, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.5021787285804749, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0599, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 4.338947296142578, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0273, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.1511262208223343, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0178, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.07180149853229523, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0261, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.07007061690092087, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0199, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06695059686899185, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0139, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06239445134997368, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0693, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.017012804746627808, | |
| "eval_runtime": 1.2818, | |
| "eval_samples_per_second": 103.759, | |
| "eval_steps_per_second": 13.262, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06727304309606552, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0121, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.061900340020656586, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0117, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06652108579874039, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0114, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.058176226913928986, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0111, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.058512184768915176, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0113, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05545031651854515, | |
| "learning_rate": 0.0, | |
| "loss": 0.0244, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11528049953855, | |
| "train_runtime": 92.6738, | |
| "train_samples_per_second": 44.63, | |
| "train_steps_per_second": 2.806 | |
| } | |
| ], | |
| "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 | |
| } | |