| { | |
| "best_metric": 0.05104082077741623, | |
| "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": 2.294018268585205, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8957, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 0.8602861762046814, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3831, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 2.684948682785034, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2334, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.219984531402588, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2021, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.8389067053794861, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1938, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 1.1451736688613892, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.1996, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.6764400005340576, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1578, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.15344880521297455, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.051, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.12597502768039703, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0659, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 2.088905096054077, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.1132, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9473684210526315, | |
| "eval_loss": 0.16588357090950012, | |
| "eval_runtime": 1.2545, | |
| "eval_samples_per_second": 106.017, | |
| "eval_steps_per_second": 13.551, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.3840338885784149, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1396, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.1193251684308052, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1276, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.09910660982131958, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0583, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.0939779207110405, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0388, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 6.65282678604126, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0397, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 5.5235981941223145, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0363, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.10575386881828308, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0227, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 2.821331262588501, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0729, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06803790479898453, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0398, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.14886723458766937, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0193, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.05104082077741623, | |
| "eval_runtime": 1.289, | |
| "eval_samples_per_second": 103.178, | |
| "eval_steps_per_second": 13.188, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.10110070556402206, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0134, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06876010447740555, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0123, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06473541259765625, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0122, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06058332324028015, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0116, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.062133342027664185, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0113, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05672670528292656, | |
| "learning_rate": 0.0, | |
| "loss": 0.0113, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.12164702398272662, | |
| "train_runtime": 85.5488, | |
| "train_samples_per_second": 48.347, | |
| "train_steps_per_second": 3.039 | |
| } | |
| ], | |
| "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 | |
| } | |