| { | |
| "best_metric": 0.033932678401470184, | |
| "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": 0.15, | |
| "grad_norm": 1.3803807497024536, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.7856, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 2.8112845420837402, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.2893, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 2.8095664978027344, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.2275, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 4.025726795196533, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.3265, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 4.194093704223633, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2264, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 1.8275835514068604, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2016, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 0.2430502474308014, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.1053, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 0.1496405005455017, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0802, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 0.12861108779907227, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0569, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.1920614093542099, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0532, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.033932678401470184, | |
| "eval_runtime": 2.105, | |
| "eval_samples_per_second": 63.183, | |
| "eval_steps_per_second": 8.076, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 2.8299429416656494, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0451, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 0.09644599258899689, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0485, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.07946658879518509, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0454, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 0.08122861385345459, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0343, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "grad_norm": 0.1984899640083313, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0452, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "grad_norm": 0.08809161186218262, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0386, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 5.108344554901123, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0411, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 11.191219329833984, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0344, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 0.062365446239709854, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0341, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "grad_norm": 0.9924643039703369, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0132, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.04654044657945633, | |
| "eval_runtime": 2.6877, | |
| "eval_samples_per_second": 49.484, | |
| "eval_steps_per_second": 6.325, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.05520316958427429, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0114, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "grad_norm": 0.055582791566848755, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0108, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "grad_norm": 0.060546137392520905, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0104, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "grad_norm": 0.0531916543841362, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0102, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.85, | |
| "grad_norm": 0.056302715092897415, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0103, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.0926387831568718, | |
| "learning_rate": 0.0, | |
| "loss": 0.0102, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.10752537130163266, | |
| "train_runtime": 152.8867, | |
| "train_samples_per_second": 27.053, | |
| "train_steps_per_second": 1.701 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |