|
{ |
|
"best_metric": 0.8333333333333334, |
|
"best_model_checkpoint": "swinv2-tiny-patch4-window8-256-dmae-va-U5-42/checkpoint-201", |
|
"epoch": 37.935483870967744, |
|
"eval_steps": 500, |
|
"global_step": 294, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.9, |
|
"eval_accuracy": 0.4, |
|
"eval_loss": 1.3298888206481934, |
|
"eval_runtime": 2.4022, |
|
"eval_samples_per_second": 24.977, |
|
"eval_steps_per_second": 0.833, |
|
"step": 7 |
|
}, |
|
{ |
|
"epoch": 1.55, |
|
"grad_norm": 3.2318508625030518, |
|
"learning_rate": 2e-05, |
|
"loss": 1.3678, |
|
"step": 12 |
|
}, |
|
{ |
|
"epoch": 1.94, |
|
"eval_accuracy": 0.45, |
|
"eval_loss": 1.2662369012832642, |
|
"eval_runtime": 1.2745, |
|
"eval_samples_per_second": 47.077, |
|
"eval_steps_per_second": 1.569, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 2.97, |
|
"eval_accuracy": 0.5166666666666667, |
|
"eval_loss": 1.09585440158844, |
|
"eval_runtime": 1.282, |
|
"eval_samples_per_second": 46.801, |
|
"eval_steps_per_second": 1.56, |
|
"step": 23 |
|
}, |
|
{ |
|
"epoch": 3.1, |
|
"grad_norm": 5.9810967445373535, |
|
"learning_rate": 4e-05, |
|
"loss": 1.2546, |
|
"step": 24 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.55, |
|
"eval_loss": 0.9759470820426941, |
|
"eval_runtime": 2.1999, |
|
"eval_samples_per_second": 27.274, |
|
"eval_steps_per_second": 0.909, |
|
"step": 31 |
|
}, |
|
{ |
|
"epoch": 4.65, |
|
"grad_norm": 8.929350852966309, |
|
"learning_rate": 4.886363636363637e-05, |
|
"loss": 1.0271, |
|
"step": 36 |
|
}, |
|
{ |
|
"epoch": 4.9, |
|
"eval_accuracy": 0.5666666666666667, |
|
"eval_loss": 0.9374793767929077, |
|
"eval_runtime": 1.3051, |
|
"eval_samples_per_second": 45.972, |
|
"eval_steps_per_second": 1.532, |
|
"step": 38 |
|
}, |
|
{ |
|
"epoch": 5.94, |
|
"eval_accuracy": 0.6, |
|
"eval_loss": 0.8728086948394775, |
|
"eval_runtime": 1.6572, |
|
"eval_samples_per_second": 36.206, |
|
"eval_steps_per_second": 1.207, |
|
"step": 46 |
|
}, |
|
{ |
|
"epoch": 6.19, |
|
"grad_norm": 12.113182067871094, |
|
"learning_rate": 4.659090909090909e-05, |
|
"loss": 0.8075, |
|
"step": 48 |
|
}, |
|
{ |
|
"epoch": 6.97, |
|
"eval_accuracy": 0.7166666666666667, |
|
"eval_loss": 0.7359870076179504, |
|
"eval_runtime": 1.8366, |
|
"eval_samples_per_second": 32.668, |
|
"eval_steps_per_second": 1.089, |
|
"step": 54 |
|
}, |
|
{ |
|
"epoch": 7.74, |
|
"grad_norm": 10.912686347961426, |
|
"learning_rate": 4.431818181818182e-05, |
|
"loss": 0.7026, |
|
"step": 60 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.6666666666666666, |
|
"eval_loss": 0.8097081780433655, |
|
"eval_runtime": 1.5216, |
|
"eval_samples_per_second": 39.432, |
|
"eval_steps_per_second": 1.314, |
|
"step": 62 |
|
}, |
|
{ |
|
"epoch": 8.9, |
|
"eval_accuracy": 0.7, |
|
"eval_loss": 0.7073581218719482, |
|
"eval_runtime": 1.2986, |
|
"eval_samples_per_second": 46.204, |
|
"eval_steps_per_second": 1.54, |
|
"step": 69 |
|
}, |
|
{ |
|
"epoch": 9.29, |
|
"grad_norm": 7.962848663330078, |
|
"learning_rate": 4.204545454545455e-05, |
|
"loss": 0.5711, |
|
"step": 72 |
|
}, |
|
{ |
|
"epoch": 9.94, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.6912961602210999, |
|
"eval_runtime": 1.3465, |
|
"eval_samples_per_second": 44.561, |
|
"eval_steps_per_second": 1.485, |
|
"step": 77 |
|
}, |
|
{ |
|
"epoch": 10.84, |
|
"grad_norm": 9.730597496032715, |
|
"learning_rate": 3.9772727272727275e-05, |
|
"loss": 0.5063, |
|
"step": 84 |
|
}, |
|
{ |
|
"epoch": 10.97, |
|
"eval_accuracy": 0.7166666666666667, |
|
"eval_loss": 0.7461908459663391, |
|
"eval_runtime": 1.3164, |
|
"eval_samples_per_second": 45.578, |
|
"eval_steps_per_second": 1.519, |
|
"step": 85 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.5833333333333334, |
|
"eval_loss": 0.8508856296539307, |
|
"eval_runtime": 1.2995, |
|
"eval_samples_per_second": 46.172, |
|
"eval_steps_per_second": 1.539, |
|
"step": 93 |
|
}, |
|
{ |
|
"epoch": 12.39, |
|
"grad_norm": 6.257162570953369, |
|
"learning_rate": 3.7500000000000003e-05, |
|
"loss": 0.4701, |
|
"step": 96 |
|
}, |
|
{ |
|
"epoch": 12.9, |
|
"eval_accuracy": 0.7333333333333333, |
|
"eval_loss": 0.6895251870155334, |
|
"eval_runtime": 1.2873, |
|
"eval_samples_per_second": 46.61, |
|
"eval_steps_per_second": 1.554, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 13.94, |
|
"grad_norm": 8.040353775024414, |
|
"learning_rate": 3.522727272727273e-05, |
|
"loss": 0.3708, |
|
"step": 108 |
|
}, |
|
{ |
|
"epoch": 13.94, |
|
"eval_accuracy": 0.6833333333333333, |
|
"eval_loss": 0.7592976689338684, |
|
"eval_runtime": 1.3064, |
|
"eval_samples_per_second": 45.927, |
|
"eval_steps_per_second": 1.531, |
|
"step": 108 |
|
}, |
|
{ |
|
"epoch": 14.97, |
|
"eval_accuracy": 0.7166666666666667, |
|
"eval_loss": 0.8621997833251953, |
|
"eval_runtime": 1.8003, |
|
"eval_samples_per_second": 33.327, |
|
"eval_steps_per_second": 1.111, |
|
"step": 116 |
|
}, |
|
{ |
|
"epoch": 15.48, |
|
"grad_norm": 9.414116859436035, |
|
"learning_rate": 3.295454545454545e-05, |
|
"loss": 0.3581, |
|
"step": 120 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy": 0.7666666666666667, |
|
"eval_loss": 0.750374972820282, |
|
"eval_runtime": 1.348, |
|
"eval_samples_per_second": 44.509, |
|
"eval_steps_per_second": 1.484, |
|
"step": 124 |
|
}, |
|
{ |
|
"epoch": 16.9, |
|
"eval_accuracy": 0.75, |
|
"eval_loss": 0.669380247592926, |
|
"eval_runtime": 1.2906, |
|
"eval_samples_per_second": 46.49, |
|
"eval_steps_per_second": 1.55, |
|
"step": 131 |
|
}, |
|
{ |
|
"epoch": 17.03, |
|
"grad_norm": 10.203311920166016, |
|
"learning_rate": 3.068181818181818e-05, |
|
"loss": 0.3342, |
|
"step": 132 |
|
}, |
|
{ |
|
"epoch": 17.94, |
|
"eval_accuracy": 0.7333333333333333, |
|
"eval_loss": 0.726170003414154, |
|
"eval_runtime": 1.2961, |
|
"eval_samples_per_second": 46.292, |
|
"eval_steps_per_second": 1.543, |
|
"step": 139 |
|
}, |
|
{ |
|
"epoch": 18.58, |
|
"grad_norm": 6.9751410484313965, |
|
"learning_rate": 2.8409090909090912e-05, |
|
"loss": 0.2979, |
|
"step": 144 |
|
}, |
|
{ |
|
"epoch": 18.97, |
|
"eval_accuracy": 0.7166666666666667, |
|
"eval_loss": 0.7233790159225464, |
|
"eval_runtime": 1.5486, |
|
"eval_samples_per_second": 38.744, |
|
"eval_steps_per_second": 1.291, |
|
"step": 147 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.6403045058250427, |
|
"eval_runtime": 1.8291, |
|
"eval_samples_per_second": 32.804, |
|
"eval_steps_per_second": 1.093, |
|
"step": 155 |
|
}, |
|
{ |
|
"epoch": 20.13, |
|
"grad_norm": 11.188940048217773, |
|
"learning_rate": 2.6136363636363637e-05, |
|
"loss": 0.2919, |
|
"step": 156 |
|
}, |
|
{ |
|
"epoch": 20.9, |
|
"eval_accuracy": 0.7666666666666667, |
|
"eval_loss": 0.6847358345985413, |
|
"eval_runtime": 1.8691, |
|
"eval_samples_per_second": 32.102, |
|
"eval_steps_per_second": 1.07, |
|
"step": 162 |
|
}, |
|
{ |
|
"epoch": 21.68, |
|
"grad_norm": 8.915830612182617, |
|
"learning_rate": 2.3863636363636365e-05, |
|
"loss": 0.274, |
|
"step": 168 |
|
}, |
|
{ |
|
"epoch": 21.94, |
|
"eval_accuracy": 0.75, |
|
"eval_loss": 0.6942654848098755, |
|
"eval_runtime": 1.6216, |
|
"eval_samples_per_second": 37.0, |
|
"eval_steps_per_second": 1.233, |
|
"step": 170 |
|
}, |
|
{ |
|
"epoch": 22.97, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.7235167026519775, |
|
"eval_runtime": 1.2987, |
|
"eval_samples_per_second": 46.199, |
|
"eval_steps_per_second": 1.54, |
|
"step": 178 |
|
}, |
|
{ |
|
"epoch": 23.23, |
|
"grad_norm": 5.901014804840088, |
|
"learning_rate": 2.1590909090909093e-05, |
|
"loss": 0.2434, |
|
"step": 180 |
|
}, |
|
{ |
|
"epoch": 24.0, |
|
"eval_accuracy": 0.75, |
|
"eval_loss": 0.7835569977760315, |
|
"eval_runtime": 1.3206, |
|
"eval_samples_per_second": 45.434, |
|
"eval_steps_per_second": 1.514, |
|
"step": 186 |
|
}, |
|
{ |
|
"epoch": 24.77, |
|
"grad_norm": 9.987520217895508, |
|
"learning_rate": 1.9318181818181818e-05, |
|
"loss": 0.239, |
|
"step": 192 |
|
}, |
|
{ |
|
"epoch": 24.9, |
|
"eval_accuracy": 0.8166666666666667, |
|
"eval_loss": 0.7199376225471497, |
|
"eval_runtime": 1.3056, |
|
"eval_samples_per_second": 45.956, |
|
"eval_steps_per_second": 1.532, |
|
"step": 193 |
|
}, |
|
{ |
|
"epoch": 25.94, |
|
"eval_accuracy": 0.8333333333333334, |
|
"eval_loss": 0.6805989146232605, |
|
"eval_runtime": 1.2869, |
|
"eval_samples_per_second": 46.625, |
|
"eval_steps_per_second": 1.554, |
|
"step": 201 |
|
}, |
|
{ |
|
"epoch": 26.32, |
|
"grad_norm": 9.460673332214355, |
|
"learning_rate": 1.7045454545454546e-05, |
|
"loss": 0.2184, |
|
"step": 204 |
|
}, |
|
{ |
|
"epoch": 26.97, |
|
"eval_accuracy": 0.8, |
|
"eval_loss": 0.6923350095748901, |
|
"eval_runtime": 1.3321, |
|
"eval_samples_per_second": 45.043, |
|
"eval_steps_per_second": 1.501, |
|
"step": 209 |
|
}, |
|
{ |
|
"epoch": 27.87, |
|
"grad_norm": 7.031528472900391, |
|
"learning_rate": 1.4772727272727274e-05, |
|
"loss": 0.2176, |
|
"step": 216 |
|
}, |
|
{ |
|
"epoch": 28.0, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.7069784998893738, |
|
"eval_runtime": 1.3346, |
|
"eval_samples_per_second": 44.957, |
|
"eval_steps_per_second": 1.499, |
|
"step": 217 |
|
}, |
|
{ |
|
"epoch": 28.9, |
|
"eval_accuracy": 0.7666666666666667, |
|
"eval_loss": 0.6990731954574585, |
|
"eval_runtime": 1.3399, |
|
"eval_samples_per_second": 44.779, |
|
"eval_steps_per_second": 1.493, |
|
"step": 224 |
|
}, |
|
{ |
|
"epoch": 29.42, |
|
"grad_norm": 6.380643367767334, |
|
"learning_rate": 1.25e-05, |
|
"loss": 0.231, |
|
"step": 228 |
|
}, |
|
{ |
|
"epoch": 29.94, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.7043098211288452, |
|
"eval_runtime": 1.3053, |
|
"eval_samples_per_second": 45.968, |
|
"eval_steps_per_second": 1.532, |
|
"step": 232 |
|
}, |
|
{ |
|
"epoch": 30.97, |
|
"grad_norm": 7.81527042388916, |
|
"learning_rate": 1.0227272727272729e-05, |
|
"loss": 0.1889, |
|
"step": 240 |
|
}, |
|
{ |
|
"epoch": 30.97, |
|
"eval_accuracy": 0.7666666666666667, |
|
"eval_loss": 0.6575068235397339, |
|
"eval_runtime": 1.3306, |
|
"eval_samples_per_second": 45.092, |
|
"eval_steps_per_second": 1.503, |
|
"step": 240 |
|
}, |
|
{ |
|
"epoch": 32.0, |
|
"eval_accuracy": 0.75, |
|
"eval_loss": 0.752100944519043, |
|
"eval_runtime": 1.3091, |
|
"eval_samples_per_second": 45.833, |
|
"eval_steps_per_second": 1.528, |
|
"step": 248 |
|
}, |
|
{ |
|
"epoch": 32.52, |
|
"grad_norm": 12.994545936584473, |
|
"learning_rate": 7.954545454545455e-06, |
|
"loss": 0.2033, |
|
"step": 252 |
|
}, |
|
{ |
|
"epoch": 32.9, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.7062455415725708, |
|
"eval_runtime": 1.3337, |
|
"eval_samples_per_second": 44.989, |
|
"eval_steps_per_second": 1.5, |
|
"step": 255 |
|
}, |
|
{ |
|
"epoch": 33.94, |
|
"eval_accuracy": 0.8, |
|
"eval_loss": 0.6957877278327942, |
|
"eval_runtime": 1.3005, |
|
"eval_samples_per_second": 46.136, |
|
"eval_steps_per_second": 1.538, |
|
"step": 263 |
|
}, |
|
{ |
|
"epoch": 34.06, |
|
"grad_norm": 6.9790143966674805, |
|
"learning_rate": 5.681818181818182e-06, |
|
"loss": 0.1891, |
|
"step": 264 |
|
}, |
|
{ |
|
"epoch": 34.97, |
|
"eval_accuracy": 0.8, |
|
"eval_loss": 0.7188975811004639, |
|
"eval_runtime": 1.2948, |
|
"eval_samples_per_second": 46.34, |
|
"eval_steps_per_second": 1.545, |
|
"step": 271 |
|
}, |
|
{ |
|
"epoch": 35.61, |
|
"grad_norm": 10.62868881225586, |
|
"learning_rate": 3.409090909090909e-06, |
|
"loss": 0.1739, |
|
"step": 276 |
|
}, |
|
{ |
|
"epoch": 36.0, |
|
"eval_accuracy": 0.8, |
|
"eval_loss": 0.7456963658332825, |
|
"eval_runtime": 1.3109, |
|
"eval_samples_per_second": 45.772, |
|
"eval_steps_per_second": 1.526, |
|
"step": 279 |
|
}, |
|
{ |
|
"epoch": 36.9, |
|
"eval_accuracy": 0.7833333333333333, |
|
"eval_loss": 0.7766147255897522, |
|
"eval_runtime": 1.3196, |
|
"eval_samples_per_second": 45.469, |
|
"eval_steps_per_second": 1.516, |
|
"step": 286 |
|
}, |
|
{ |
|
"epoch": 37.16, |
|
"grad_norm": 11.091560363769531, |
|
"learning_rate": 1.1363636363636364e-06, |
|
"loss": 0.1949, |
|
"step": 288 |
|
}, |
|
{ |
|
"epoch": 37.94, |
|
"eval_accuracy": 0.7666666666666667, |
|
"eval_loss": 0.7808470129966736, |
|
"eval_runtime": 1.3104, |
|
"eval_samples_per_second": 45.788, |
|
"eval_steps_per_second": 1.526, |
|
"step": 294 |
|
}, |
|
{ |
|
"epoch": 37.94, |
|
"step": 294, |
|
"total_flos": 1.2027310550050406e+18, |
|
"train_loss": 0.44187615152929915, |
|
"train_runtime": 903.9492, |
|
"train_samples_per_second": 45.255, |
|
"train_steps_per_second": 0.325 |
|
} |
|
], |
|
"logging_steps": 12, |
|
"max_steps": 294, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 42, |
|
"save_steps": 500, |
|
"total_flos": 1.2027310550050406e+18, |
|
"train_batch_size": 32, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|