| { | |
| "best_metric": 0.018796762451529503, | |
| "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": 1.4108372926712036, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7924, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.044903039932251, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.301, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.7039980888366699, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1992, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.1257214546203613, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2634, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.34912949800491333, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1518, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 4.554572582244873, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.23, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.4593968093395233, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.0491, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.13646918535232544, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 5.695180416107178, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0552, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 1.3406460285186768, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0846, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.061424773186445236, | |
| "eval_runtime": 1.708, | |
| "eval_samples_per_second": 77.869, | |
| "eval_steps_per_second": 9.953, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 1.4606086015701294, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.1462, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.5630506873130798, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0893, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.09152624756097794, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0368, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.9141150116920471, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0192, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.7744202613830566, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0423, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.14094310998916626, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0894, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.2033337652683258, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0158, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 2.038670778274536, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0682, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.5425238609313965, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0148, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.12908649444580078, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0121, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.018796762451529503, | |
| "eval_runtime": 2.1434, | |
| "eval_samples_per_second": 62.05, | |
| "eval_steps_per_second": 7.931, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.07232276350259781, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0119, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.05829685181379318, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0131, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06218134984374046, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0306, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06792601197957993, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.011, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.060047898441553116, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0108, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05698742717504501, | |
| "learning_rate": 0.0, | |
| "loss": 0.0106, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.1095780831690018, | |
| "train_runtime": 90.7321, | |
| "train_samples_per_second": 45.585, | |
| "train_steps_per_second": 2.866 | |
| } | |
| ], | |
| "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 | |
| } | |