| { | |
| "best_metric": 0.01571890525519848, | |
| "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": 1.55097234249115, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7977, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.663919687271118, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3046, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.4333198368549347, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2134, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.0663870573043823, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2098, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.7898901700973511, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1677, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 4.087493419647217, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2881, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.7810757160186768, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1081, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 1.1969906091690063, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0889, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 7.20994758605957, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0653, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.1867348849773407, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0936, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9548872180451128, | |
| "eval_loss": 0.1083720475435257, | |
| "eval_runtime": 1.2758, | |
| "eval_samples_per_second": 104.252, | |
| "eval_steps_per_second": 13.325, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.36014097929000854, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1049, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 3.73897123336792, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1034, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.10135021805763245, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0846, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.21503755450248718, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0318, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.08955585211515427, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0168, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.10519933700561523, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0466, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.15952926874160767, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.016, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 2.9184176921844482, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0371, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06390605121850967, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.026, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06250115483999252, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0119, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.01571890525519848, | |
| "eval_runtime": 1.742, | |
| "eval_samples_per_second": 76.349, | |
| "eval_steps_per_second": 9.759, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.059098560363054276, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0117, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.05702250450849533, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0116, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06359326094388962, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0109, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.05622411519289017, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0106, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.057671088725328445, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0107, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05275500565767288, | |
| "learning_rate": 0.0, | |
| "loss": 0.0111, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11088088481472089, | |
| "train_runtime": 85.6554, | |
| "train_samples_per_second": 48.287, | |
| "train_steps_per_second": 3.035 | |
| } | |
| ], | |
| "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 | |
| } | |