| { | |
| "best_metric": 0.004596503917127848, | |
| "best_model_checkpoint": "./vit-base-pets/checkpoint-200", | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 232, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 168186.984375, | |
| "learning_rate": 0.0002870689655172413, | |
| "loss": 2.1401, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 134643.921875, | |
| "learning_rate": 0.0002741379310344827, | |
| "loss": 0.5937, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 117247.1015625, | |
| "learning_rate": 0.00026120689655172413, | |
| "loss": 0.3283, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 148088.53125, | |
| "learning_rate": 0.0002482758620689655, | |
| "loss": 0.3447, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 118186.203125, | |
| "learning_rate": 0.00023534482758620685, | |
| "loss": 0.2972, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "grad_norm": 89050.390625, | |
| "learning_rate": 0.00022241379310344826, | |
| "loss": 0.2425, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 50770.4296875, | |
| "learning_rate": 0.00020948275862068963, | |
| "loss": 0.1025, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 102693.65625, | |
| "learning_rate": 0.000196551724137931, | |
| "loss": 0.0998, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 58018.40234375, | |
| "learning_rate": 0.0001836206896551724, | |
| "loss": 0.1048, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 51126.80078125, | |
| "learning_rate": 0.0001706896551724138, | |
| "loss": 0.0887, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "eval_accuracy": 0.9774018944519621, | |
| "eval_loss": 0.07652416825294495, | |
| "eval_runtime": 90.7155, | |
| "eval_samples_per_second": 81.463, | |
| "eval_steps_per_second": 5.093, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 86962.4921875, | |
| "learning_rate": 0.00015775862068965517, | |
| "loss": 0.0926, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "grad_norm": 72358.4609375, | |
| "learning_rate": 0.00014482758620689654, | |
| "loss": 0.0706, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 20616.7109375, | |
| "learning_rate": 0.00013189655172413792, | |
| "loss": 0.039, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "grad_norm": 70676.0625, | |
| "learning_rate": 0.00011896551724137931, | |
| "loss": 0.0351, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "grad_norm": 12601.1337890625, | |
| "learning_rate": 0.00010603448275862067, | |
| "loss": 0.0263, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 77678.21875, | |
| "learning_rate": 9.310344827586206e-05, | |
| "loss": 0.0296, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 36381.59765625, | |
| "learning_rate": 8.017241379310344e-05, | |
| "loss": 0.025, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 3.1, | |
| "grad_norm": 12375.634765625, | |
| "learning_rate": 6.724137931034483e-05, | |
| "loss": 0.0087, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 3.28, | |
| "grad_norm": 1895.7933349609375, | |
| "learning_rate": 5.4310344827586204e-05, | |
| "loss": 0.0043, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "grad_norm": 49621.76171875, | |
| "learning_rate": 4.137931034482758e-05, | |
| "loss": 0.0045, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "eval_accuracy": 0.9989174560216508, | |
| "eval_loss": 0.004596503917127848, | |
| "eval_runtime": 91.6398, | |
| "eval_samples_per_second": 80.642, | |
| "eval_steps_per_second": 5.041, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.62, | |
| "grad_norm": 798.2269287109375, | |
| "learning_rate": 2.8448275862068963e-05, | |
| "loss": 0.0045, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "grad_norm": 2413.81640625, | |
| "learning_rate": 1.5517241379310342e-05, | |
| "loss": 0.0034, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "grad_norm": 908.8674926757812, | |
| "learning_rate": 2.5862068965517237e-06, | |
| "loss": 0.0021, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 232, | |
| "total_flos": 2.2913817801515827e+18, | |
| "train_loss": 0.202088082896682, | |
| "train_runtime": 854.7685, | |
| "train_samples_per_second": 34.582, | |
| "train_steps_per_second": 0.271 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 232, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 2.2913817801515827e+18, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |