Upload checkpoint-5040/trainer_state.json with huggingface_hub
Browse files
checkpoint-5040/trainer_state.json
ADDED
|
@@ -0,0 +1,347 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 5040,
|
| 3 |
+
"best_metric": 0.8438760541734693,
|
| 4 |
+
"best_model_checkpoint": "outputs/runs/vit5/checkpoint-5040",
|
| 5 |
+
"epoch": 16.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5040,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.9968253968253968,
|
| 14 |
+
"grad_norm": 48.960147857666016,
|
| 15 |
+
"learning_rate": 4.999432333543028e-06,
|
| 16 |
+
"loss": 4.2164,
|
| 17 |
+
"step": 314
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 1.0,
|
| 21 |
+
"eval_accuracy": 0.8120646345156866,
|
| 22 |
+
"eval_f1": 0.7880041194112739,
|
| 23 |
+
"eval_loss": 1.4588117599487305,
|
| 24 |
+
"eval_precision": 0.8369479599718508,
|
| 25 |
+
"eval_recall": 0.7755522799379979,
|
| 26 |
+
"eval_runtime": 1.628,
|
| 27 |
+
"eval_samples_per_second": 799.123,
|
| 28 |
+
"eval_steps_per_second": 25.184,
|
| 29 |
+
"step": 315
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.9936507936507937,
|
| 33 |
+
"grad_norm": 15.666463851928711,
|
| 34 |
+
"learning_rate": 4.996525669931999e-06,
|
| 35 |
+
"loss": 1.5636,
|
| 36 |
+
"step": 628
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 2.0,
|
| 40 |
+
"eval_accuracy": 0.8272790436657106,
|
| 41 |
+
"eval_f1": 0.8200906148872792,
|
| 42 |
+
"eval_loss": 0.8009101748466492,
|
| 43 |
+
"eval_precision": 0.819129062168523,
|
| 44 |
+
"eval_recall": 0.8211574663523349,
|
| 45 |
+
"eval_runtime": 1.6386,
|
| 46 |
+
"eval_samples_per_second": 793.961,
|
| 47 |
+
"eval_steps_per_second": 25.021,
|
| 48 |
+
"step": 630
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 2.9904761904761905,
|
| 52 |
+
"grad_norm": 19.285402297973633,
|
| 53 |
+
"learning_rate": 4.991155236893945e-06,
|
| 54 |
+
"loss": 1.1621,
|
| 55 |
+
"step": 942
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 3.0,
|
| 59 |
+
"eval_accuracy": 0.8486576045322549,
|
| 60 |
+
"eval_f1": 0.8387492829457615,
|
| 61 |
+
"eval_loss": 0.6665186285972595,
|
| 62 |
+
"eval_precision": 0.8476389137544842,
|
| 63 |
+
"eval_recall": 0.833091047050099,
|
| 64 |
+
"eval_runtime": 1.6403,
|
| 65 |
+
"eval_samples_per_second": 793.151,
|
| 66 |
+
"eval_steps_per_second": 24.996,
|
| 67 |
+
"step": 945
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 3.9873015873015873,
|
| 71 |
+
"grad_norm": 28.716716766357422,
|
| 72 |
+
"learning_rate": 4.983326334397891e-06,
|
| 73 |
+
"loss": 0.9673,
|
| 74 |
+
"step": 1256
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.0,
|
| 78 |
+
"eval_accuracy": 0.8452548502609966,
|
| 79 |
+
"eval_f1": 0.8306632859859202,
|
| 80 |
+
"eval_loss": 0.7017992734909058,
|
| 81 |
+
"eval_precision": 0.8573660674461923,
|
| 82 |
+
"eval_recall": 0.8196344002399345,
|
| 83 |
+
"eval_runtime": 1.6405,
|
| 84 |
+
"eval_samples_per_second": 793.027,
|
| 85 |
+
"eval_steps_per_second": 24.992,
|
| 86 |
+
"step": 1260
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.984126984126984,
|
| 90 |
+
"grad_norm": 62.80952072143555,
|
| 91 |
+
"learning_rate": 4.97304668862541e-06,
|
| 92 |
+
"loss": 0.8578,
|
| 93 |
+
"step": 1570
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 5.0,
|
| 97 |
+
"eval_accuracy": 0.8483012951844792,
|
| 98 |
+
"eval_f1": 0.8374094515073132,
|
| 99 |
+
"eval_loss": 0.5731419920921326,
|
| 100 |
+
"eval_precision": 0.8497690158969053,
|
| 101 |
+
"eval_recall": 0.8303468421765257,
|
| 102 |
+
"eval_runtime": 1.6424,
|
| 103 |
+
"eval_samples_per_second": 792.155,
|
| 104 |
+
"eval_steps_per_second": 24.964,
|
| 105 |
+
"step": 1575
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 5.980952380952381,
|
| 109 |
+
"grad_norm": 28.359893798828125,
|
| 110 |
+
"learning_rate": 4.9603264443458e-06,
|
| 111 |
+
"loss": 0.7528,
|
| 112 |
+
"step": 1884
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 6.0,
|
| 116 |
+
"eval_accuracy": 0.8424043754787907,
|
| 117 |
+
"eval_f1": 0.8368542522225871,
|
| 118 |
+
"eval_loss": 0.5212520360946655,
|
| 119 |
+
"eval_precision": 0.8343696637314615,
|
| 120 |
+
"eval_recall": 0.8403125137303477,
|
| 121 |
+
"eval_runtime": 1.6431,
|
| 122 |
+
"eval_samples_per_second": 791.777,
|
| 123 |
+
"eval_steps_per_second": 24.952,
|
| 124 |
+
"step": 1890
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 6.977777777777778,
|
| 128 |
+
"grad_norm": 13.184873580932617,
|
| 129 |
+
"learning_rate": 4.945178154904432e-06,
|
| 130 |
+
"loss": 0.6746,
|
| 131 |
+
"step": 2198
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 7.0,
|
| 135 |
+
"eval_accuracy": 0.8452726657283853,
|
| 136 |
+
"eval_f1": 0.8388474476297365,
|
| 137 |
+
"eval_loss": 0.5077288746833801,
|
| 138 |
+
"eval_precision": 0.8378019445498359,
|
| 139 |
+
"eval_recall": 0.8400070136197805,
|
| 140 |
+
"eval_runtime": 1.6424,
|
| 141 |
+
"eval_samples_per_second": 792.11,
|
| 142 |
+
"eval_steps_per_second": 24.963,
|
| 143 |
+
"step": 2205
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 7.974603174603175,
|
| 147 |
+
"grad_norm": 9.050921440124512,
|
| 148 |
+
"learning_rate": 4.92761676983411e-06,
|
| 149 |
+
"loss": 0.6276,
|
| 150 |
+
"step": 2512
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 8.0,
|
| 154 |
+
"eval_accuracy": 0.8498868717820812,
|
| 155 |
+
"eval_f1": 0.8402758437579932,
|
| 156 |
+
"eval_loss": 0.48333004117012024,
|
| 157 |
+
"eval_precision": 0.8484423552268903,
|
| 158 |
+
"eval_recall": 0.8349357523808886,
|
| 159 |
+
"eval_runtime": 1.6444,
|
| 160 |
+
"eval_samples_per_second": 791.192,
|
| 161 |
+
"eval_steps_per_second": 24.934,
|
| 162 |
+
"step": 2520
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 8.971428571428572,
|
| 166 |
+
"grad_norm": 14.836031913757324,
|
| 167 |
+
"learning_rate": 4.9076596201016856e-06,
|
| 168 |
+
"loss": 0.591,
|
| 169 |
+
"step": 2826
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 9.0,
|
| 173 |
+
"eval_accuracy": 0.8510092462275748,
|
| 174 |
+
"eval_f1": 0.8380468647887628,
|
| 175 |
+
"eval_loss": 0.5170930624008179,
|
| 176 |
+
"eval_precision": 0.8600643338742522,
|
| 177 |
+
"eval_recall": 0.8279538920826988,
|
| 178 |
+
"eval_runtime": 1.6449,
|
| 179 |
+
"eval_samples_per_second": 790.913,
|
| 180 |
+
"eval_steps_per_second": 24.925,
|
| 181 |
+
"step": 2835
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 9.968253968253968,
|
| 185 |
+
"grad_norm": 17.490633010864258,
|
| 186 |
+
"learning_rate": 4.88532640100449e-06,
|
| 187 |
+
"loss": 0.5411,
|
| 188 |
+
"step": 3140
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 10.0,
|
| 192 |
+
"eval_accuracy": 0.840872245283355,
|
| 193 |
+
"eval_f1": 0.8357643191911723,
|
| 194 |
+
"eval_loss": 0.4666968882083893,
|
| 195 |
+
"eval_precision": 0.8328248424035201,
|
| 196 |
+
"eval_recall": 0.8404652389103109,
|
| 197 |
+
"eval_runtime": 1.6447,
|
| 198 |
+
"eval_samples_per_second": 791.007,
|
| 199 |
+
"eval_steps_per_second": 24.928,
|
| 200 |
+
"step": 3150
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 10.965079365079365,
|
| 204 |
+
"grad_norm": 27.608245849609375,
|
| 205 |
+
"learning_rate": 4.860639152733449e-06,
|
| 206 |
+
"loss": 0.5193,
|
| 207 |
+
"step": 3454
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 11.0,
|
| 211 |
+
"eval_accuracy": 0.8485863426626997,
|
| 212 |
+
"eval_f1": 0.8406162312164841,
|
| 213 |
+
"eval_loss": 0.4396151602268219,
|
| 214 |
+
"eval_precision": 0.8434539842909232,
|
| 215 |
+
"eval_recall": 0.8382779460247154,
|
| 216 |
+
"eval_runtime": 1.6435,
|
| 217 |
+
"eval_samples_per_second": 791.592,
|
| 218 |
+
"eval_steps_per_second": 24.946,
|
| 219 |
+
"step": 3465
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 11.961904761904762,
|
| 223 |
+
"grad_norm": 7.845660209655762,
|
| 224 |
+
"learning_rate": 4.833622238622079e-06,
|
| 225 |
+
"loss": 0.4943,
|
| 226 |
+
"step": 3768
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 12.0,
|
| 230 |
+
"eval_accuracy": 0.8496374552386382,
|
| 231 |
+
"eval_f1": 0.842659244060233,
|
| 232 |
+
"eval_loss": 0.43954724073410034,
|
| 233 |
+
"eval_precision": 0.843136531563339,
|
| 234 |
+
"eval_recall": 0.8422001823877059,
|
| 235 |
+
"eval_runtime": 1.6428,
|
| 236 |
+
"eval_samples_per_second": 791.964,
|
| 237 |
+
"eval_steps_per_second": 24.958,
|
| 238 |
+
"step": 3780
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 12.958730158730159,
|
| 242 |
+
"grad_norm": 9.281864166259766,
|
| 243 |
+
"learning_rate": 4.804302321102816e-06,
|
| 244 |
+
"loss": 0.4715,
|
| 245 |
+
"step": 4082
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 13.0,
|
| 249 |
+
"eval_accuracy": 0.8486576045322549,
|
| 250 |
+
"eval_f1": 0.8384831543048918,
|
| 251 |
+
"eval_loss": 0.478777676820755,
|
| 252 |
+
"eval_precision": 0.8483098559263628,
|
| 253 |
+
"eval_recall": 0.8324188900247866,
|
| 254 |
+
"eval_runtime": 1.6428,
|
| 255 |
+
"eval_samples_per_second": 791.931,
|
| 256 |
+
"eval_steps_per_second": 24.957,
|
| 257 |
+
"step": 4095
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 13.955555555555556,
|
| 261 |
+
"grad_norm": 21.69331169128418,
|
| 262 |
+
"learning_rate": 4.772708335394416e-06,
|
| 263 |
+
"loss": 0.4482,
|
| 264 |
+
"step": 4396
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 14.0,
|
| 268 |
+
"eval_accuracy": 0.8356344978710516,
|
| 269 |
+
"eval_f1": 0.8313378359536021,
|
| 270 |
+
"eval_loss": 0.49439841508865356,
|
| 271 |
+
"eval_precision": 0.8281393298498705,
|
| 272 |
+
"eval_recall": 0.8386230277449718,
|
| 273 |
+
"eval_runtime": 1.6423,
|
| 274 |
+
"eval_samples_per_second": 792.192,
|
| 275 |
+
"eval_steps_per_second": 24.965,
|
| 276 |
+
"step": 4410
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 14.952380952380953,
|
| 280 |
+
"grad_norm": 11.402544021606445,
|
| 281 |
+
"learning_rate": 4.738871460946384e-06,
|
| 282 |
+
"loss": 0.4305,
|
| 283 |
+
"step": 4710
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 15.0,
|
| 287 |
+
"eval_accuracy": 0.8504569667385223,
|
| 288 |
+
"eval_f1": 0.8397878915405506,
|
| 289 |
+
"eval_loss": 0.4550122618675232,
|
| 290 |
+
"eval_precision": 0.8519608079867367,
|
| 291 |
+
"eval_recall": 0.8327656095361328,
|
| 292 |
+
"eval_runtime": 1.6443,
|
| 293 |
+
"eval_samples_per_second": 791.219,
|
| 294 |
+
"eval_steps_per_second": 24.935,
|
| 295 |
+
"step": 4725
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 15.94920634920635,
|
| 299 |
+
"grad_norm": 9.356096267700195,
|
| 300 |
+
"learning_rate": 4.702825090668624e-06,
|
| 301 |
+
"loss": 0.4115,
|
| 302 |
+
"step": 5024
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 16.0,
|
| 306 |
+
"eval_accuracy": 0.8531114713794516,
|
| 307 |
+
"eval_f1": 0.8438760541734693,
|
| 308 |
+
"eval_loss": 0.4762667417526245,
|
| 309 |
+
"eval_precision": 0.8515227136302977,
|
| 310 |
+
"eval_recall": 0.8387574009086176,
|
| 311 |
+
"eval_runtime": 1.6403,
|
| 312 |
+
"eval_samples_per_second": 793.152,
|
| 313 |
+
"eval_steps_per_second": 24.996,
|
| 314 |
+
"step": 5040
|
| 315 |
+
}
|
| 316 |
+
],
|
| 317 |
+
"logging_steps": 314,
|
| 318 |
+
"max_steps": 31500,
|
| 319 |
+
"num_input_tokens_seen": 0,
|
| 320 |
+
"num_train_epochs": 100,
|
| 321 |
+
"save_steps": 500,
|
| 322 |
+
"stateful_callbacks": {
|
| 323 |
+
"EarlyStoppingCallback": {
|
| 324 |
+
"args": {
|
| 325 |
+
"early_stopping_patience": 5,
|
| 326 |
+
"early_stopping_threshold": 0.0
|
| 327 |
+
},
|
| 328 |
+
"attributes": {
|
| 329 |
+
"early_stopping_patience_counter": 0
|
| 330 |
+
}
|
| 331 |
+
},
|
| 332 |
+
"TrainerControl": {
|
| 333 |
+
"args": {
|
| 334 |
+
"should_epoch_stop": false,
|
| 335 |
+
"should_evaluate": false,
|
| 336 |
+
"should_log": false,
|
| 337 |
+
"should_save": true,
|
| 338 |
+
"should_training_stop": false
|
| 339 |
+
},
|
| 340 |
+
"attributes": {}
|
| 341 |
+
}
|
| 342 |
+
},
|
| 343 |
+
"total_flos": 5258284886974464.0,
|
| 344 |
+
"train_batch_size": 32,
|
| 345 |
+
"trial_name": null,
|
| 346 |
+
"trial_params": null
|
| 347 |
+
}
|