End of training
Browse files- README.md +1 -1
- all_results.json +11 -11
- eval_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +564 -894
README.md
CHANGED
|
@@ -33,7 +33,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the imagefolder dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
-
- Loss: 0.
|
| 37 |
- Accuracy: 0.9338
|
| 38 |
|
| 39 |
## Model description
|
|
|
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the imagefolder dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
+
- Loss: 0.3311
|
| 37 |
- Accuracy: 0.9338
|
| 38 |
|
| 39 |
## Model description
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second": 0.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 46.15384615384615,
|
| 3 |
+
"eval_accuracy": 0.9338235294117647,
|
| 4 |
+
"eval_loss": 0.33108076453208923,
|
| 5 |
+
"eval_runtime": 29.2681,
|
| 6 |
+
"eval_samples_per_second": 4.647,
|
| 7 |
+
"eval_steps_per_second": 0.171,
|
| 8 |
+
"total_flos": 4.3781443993328026e+18,
|
| 9 |
+
"train_loss": 0.41361336602105037,
|
| 10 |
+
"train_runtime": 14457.8356,
|
| 11 |
+
"train_samples_per_second": 4.233,
|
| 12 |
+
"train_steps_per_second": 0.031
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 46.15384615384615,
|
| 3 |
+
"eval_accuracy": 0.9338235294117647,
|
| 4 |
+
"eval_loss": 0.33108076453208923,
|
| 5 |
+
"eval_runtime": 29.2681,
|
| 6 |
+
"eval_samples_per_second": 4.647,
|
| 7 |
+
"eval_steps_per_second": 0.171
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 46.15384615384615,
|
| 3 |
+
"total_flos": 4.3781443993328026e+18,
|
| 4 |
+
"train_loss": 0.41361336602105037,
|
| 5 |
+
"train_runtime": 14457.8356,
|
| 6 |
+
"train_samples_per_second": 4.233,
|
| 7 |
+
"train_steps_per_second": 0.031
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1095 +1,765 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-papsmear/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"
|
| 14 |
-
"
|
| 15 |
-
"
|
| 16 |
-
"
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
"epoch": 0.5228758169934641,
|
| 20 |
-
"grad_norm": 1.3496487140655518,
|
| 21 |
-
"learning_rate": 8.771929824561403e-06,
|
| 22 |
-
"loss": 1.7834,
|
| 23 |
-
"step": 20
|
| 24 |
-
},
|
| 25 |
-
{
|
| 26 |
-
"epoch": 0.7843137254901961,
|
| 27 |
-
"grad_norm": 1.857043981552124,
|
| 28 |
-
"learning_rate": 1.3157894736842106e-05,
|
| 29 |
-
"loss": 1.6954,
|
| 30 |
-
"step": 30
|
| 31 |
-
},
|
| 32 |
-
{
|
| 33 |
-
"epoch": 0.9934640522875817,
|
| 34 |
-
"eval_accuracy": 0.34558823529411764,
|
| 35 |
-
"eval_loss": 1.610603928565979,
|
| 36 |
-
"eval_runtime": 33.3958,
|
| 37 |
-
"eval_samples_per_second": 4.072,
|
| 38 |
-
"eval_steps_per_second": 0.509,
|
| 39 |
-
"step": 38
|
| 40 |
-
},
|
| 41 |
-
{
|
| 42 |
-
"epoch": 1.0457516339869282,
|
| 43 |
-
"grad_norm": 1.5567628145217896,
|
| 44 |
-
"learning_rate": 1.7543859649122806e-05,
|
| 45 |
-
"loss": 1.5872,
|
| 46 |
-
"step": 40
|
| 47 |
},
|
| 48 |
{
|
| 49 |
-
"epoch": 1.
|
| 50 |
-
"grad_norm": 1.
|
| 51 |
-
"learning_rate":
|
| 52 |
-
"loss": 1.
|
| 53 |
-
"step":
|
| 54 |
-
},
|
| 55 |
-
{
|
| 56 |
-
"epoch": 1.5686274509803921,
|
| 57 |
-
"grad_norm": 1.6773402690887451,
|
| 58 |
-
"learning_rate": 2.6315789473684212e-05,
|
| 59 |
-
"loss": 1.3823,
|
| 60 |
-
"step": 60
|
| 61 |
-
},
|
| 62 |
-
{
|
| 63 |
-
"epoch": 1.8300653594771243,
|
| 64 |
-
"grad_norm": 1.4024033546447754,
|
| 65 |
-
"learning_rate": 3.0701754385964913e-05,
|
| 66 |
-
"loss": 1.2818,
|
| 67 |
-
"step": 70
|
| 68 |
-
},
|
| 69 |
-
{
|
| 70 |
-
"epoch": 1.9869281045751634,
|
| 71 |
-
"eval_accuracy": 0.5735294117647058,
|
| 72 |
-
"eval_loss": 1.2412389516830444,
|
| 73 |
-
"eval_runtime": 32.7103,
|
| 74 |
-
"eval_samples_per_second": 4.158,
|
| 75 |
-
"eval_steps_per_second": 0.52,
|
| 76 |
-
"step": 76
|
| 77 |
-
},
|
| 78 |
-
{
|
| 79 |
-
"epoch": 2.0915032679738563,
|
| 80 |
-
"grad_norm": 1.6061407327651978,
|
| 81 |
-
"learning_rate": 3.508771929824561e-05,
|
| 82 |
-
"loss": 1.2088,
|
| 83 |
-
"step": 80
|
| 84 |
-
},
|
| 85 |
-
{
|
| 86 |
-
"epoch": 2.3529411764705883,
|
| 87 |
-
"grad_norm": 3.8964905738830566,
|
| 88 |
-
"learning_rate": 3.9473684210526316e-05,
|
| 89 |
-
"loss": 1.0983,
|
| 90 |
-
"step": 90
|
| 91 |
-
},
|
| 92 |
-
{
|
| 93 |
-
"epoch": 2.6143790849673203,
|
| 94 |
-
"grad_norm": 3.24979305267334,
|
| 95 |
-
"learning_rate": 4.3859649122807014e-05,
|
| 96 |
-
"loss": 1.0632,
|
| 97 |
-
"step": 100
|
| 98 |
-
},
|
| 99 |
-
{
|
| 100 |
-
"epoch": 2.8758169934640523,
|
| 101 |
-
"grad_norm": 2.1930079460144043,
|
| 102 |
-
"learning_rate": 4.824561403508772e-05,
|
| 103 |
-
"loss": 1.0023,
|
| 104 |
-
"step": 110
|
| 105 |
},
|
| 106 |
{
|
| 107 |
-
"epoch":
|
| 108 |
-
"eval_accuracy": 0.
|
| 109 |
-
"eval_loss":
|
| 110 |
-
"eval_runtime": 32.
|
| 111 |
-
"eval_samples_per_second": 4.
|
| 112 |
-
"eval_steps_per_second": 0.
|
| 113 |
-
"step":
|
| 114 |
},
|
| 115 |
{
|
| 116 |
-
"epoch":
|
| 117 |
-
"grad_norm":
|
| 118 |
-
"learning_rate":
|
| 119 |
-
"loss":
|
| 120 |
-
"step":
|
| 121 |
},
|
| 122 |
{
|
| 123 |
-
"epoch":
|
| 124 |
-
"
|
| 125 |
-
"
|
| 126 |
-
"
|
| 127 |
-
"
|
|
|
|
|
|
|
| 128 |
},
|
| 129 |
{
|
| 130 |
-
"epoch": 3.
|
| 131 |
-
"grad_norm":
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss":
|
| 134 |
-
"step":
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 3.9215686274509802,
|
| 138 |
-
"grad_norm": 2.1056406497955322,
|
| 139 |
-
"learning_rate": 4.824561403508772e-05,
|
| 140 |
-
"loss": 0.7163,
|
| 141 |
-
"step": 150
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 4.0,
|
| 145 |
-
"eval_accuracy": 0.
|
| 146 |
-
"eval_loss":
|
| 147 |
-
"eval_runtime":
|
| 148 |
-
"eval_samples_per_second": 4.
|
| 149 |
-
"eval_steps_per_second": 0.
|
| 150 |
-
"step":
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 4.
|
| 154 |
-
"grad_norm":
|
| 155 |
-
"learning_rate": 4.
|
| 156 |
-
"loss":
|
| 157 |
-
"step":
|
| 158 |
-
},
|
| 159 |
-
{
|
| 160 |
-
"epoch": 4.444444444444445,
|
| 161 |
-
"grad_norm": 2.1595349311828613,
|
| 162 |
-
"learning_rate": 4.727095516569201e-05,
|
| 163 |
-
"loss": 0.6366,
|
| 164 |
-
"step": 170
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 4.705882352941177,
|
| 168 |
-
"grad_norm": 2.2182297706604004,
|
| 169 |
-
"learning_rate": 4.678362573099415e-05,
|
| 170 |
-
"loss": 0.5876,
|
| 171 |
-
"step": 180
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 4.967320261437909,
|
| 175 |
-
"grad_norm": 5.272632122039795,
|
| 176 |
-
"learning_rate": 4.62962962962963e-05,
|
| 177 |
-
"loss": 0.5173,
|
| 178 |
-
"step": 190
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 4.993464052287582,
|
| 182 |
-
"eval_accuracy": 0.8161764705882353,
|
| 183 |
-
"eval_loss": 0.6545882821083069,
|
| 184 |
-
"eval_runtime": 31.9706,
|
| 185 |
-
"eval_samples_per_second": 4.254,
|
| 186 |
-
"eval_steps_per_second": 0.532,
|
| 187 |
-
"step": 191
|
| 188 |
-
},
|
| 189 |
-
{
|
| 190 |
-
"epoch": 5.228758169934641,
|
| 191 |
-
"grad_norm": 5.355340003967285,
|
| 192 |
-
"learning_rate": 4.580896686159844e-05,
|
| 193 |
-
"loss": 0.5057,
|
| 194 |
-
"step": 200
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 5.490196078431373,
|
| 198 |
-
"grad_norm": 2.3517274856567383,
|
| 199 |
-
"learning_rate": 4.5321637426900585e-05,
|
| 200 |
-
"loss": 0.4577,
|
| 201 |
-
"step": 210
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 5.751633986928105,
|
| 205 |
-
"grad_norm": 7.760730266571045,
|
| 206 |
-
"learning_rate": 4.483430799220273e-05,
|
| 207 |
-
"loss": 0.5057,
|
| 208 |
-
"step": 220
|
| 209 |
-
},
|
| 210 |
-
{
|
| 211 |
-
"epoch": 5.9869281045751634,
|
| 212 |
-
"eval_accuracy": 0.8308823529411765,
|
| 213 |
-
"eval_loss": 0.6250559091567993,
|
| 214 |
-
"eval_runtime": 32.3916,
|
| 215 |
-
"eval_samples_per_second": 4.199,
|
| 216 |
-
"eval_steps_per_second": 0.525,
|
| 217 |
-
"step": 229
|
| 218 |
-
},
|
| 219 |
-
{
|
| 220 |
-
"epoch": 6.0130718954248366,
|
| 221 |
-
"grad_norm": 3.0986075401306152,
|
| 222 |
-
"learning_rate": 4.4346978557504876e-05,
|
| 223 |
-
"loss": 0.4856,
|
| 224 |
-
"step": 230
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 6.2745098039215685,
|
| 228 |
-
"grad_norm": 7.31836462020874,
|
| 229 |
-
"learning_rate": 4.3859649122807014e-05,
|
| 230 |
-
"loss": 0.5036,
|
| 231 |
-
"step": 240
|
| 232 |
-
},
|
| 233 |
-
{
|
| 234 |
-
"epoch": 6.5359477124183005,
|
| 235 |
-
"grad_norm": 2.5667953491210938,
|
| 236 |
-
"learning_rate": 4.3372319688109166e-05,
|
| 237 |
-
"loss": 0.4063,
|
| 238 |
-
"step": 250
|
| 239 |
},
|
| 240 |
{
|
| 241 |
-
"epoch":
|
| 242 |
-
"
|
| 243 |
-
"
|
| 244 |
-
"
|
| 245 |
-
"
|
|
|
|
|
|
|
| 246 |
},
|
| 247 |
{
|
| 248 |
-
"epoch":
|
| 249 |
-
"
|
| 250 |
-
"
|
| 251 |
-
"
|
| 252 |
-
"
|
| 253 |
-
"eval_steps_per_second": 0.518,
|
| 254 |
-
"step": 267
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch":
|
| 258 |
-
"
|
| 259 |
-
"
|
| 260 |
-
"
|
| 261 |
-
"
|
|
|
|
|
|
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch":
|
| 265 |
-
"grad_norm":
|
| 266 |
-
"learning_rate": 4.
|
| 267 |
-
"loss": 0.
|
| 268 |
-
"step":
|
| 269 |
},
|
| 270 |
{
|
| 271 |
-
"epoch":
|
| 272 |
-
"
|
| 273 |
-
"
|
| 274 |
-
"
|
| 275 |
-
"
|
|
|
|
|
|
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 7.
|
| 279 |
-
"grad_norm":
|
| 280 |
-
"learning_rate": 4.
|
| 281 |
-
"loss": 0.
|
| 282 |
-
"step":
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 8.0,
|
| 286 |
-
"eval_accuracy": 0.
|
| 287 |
-
"eval_loss": 0.
|
| 288 |
-
"eval_runtime":
|
| 289 |
-
"eval_samples_per_second": 4.
|
| 290 |
-
"eval_steps_per_second": 0.
|
| 291 |
-
"step":
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 8.
|
| 295 |
-
"grad_norm":
|
| 296 |
-
"learning_rate": 4.
|
| 297 |
-
"loss": 0.
|
| 298 |
-
"step":
|
| 299 |
-
},
|
| 300 |
-
{
|
| 301 |
-
"epoch": 8.366013071895425,
|
| 302 |
-
"grad_norm": 4.172213554382324,
|
| 303 |
-
"learning_rate": 3.996101364522417e-05,
|
| 304 |
-
"loss": 0.3194,
|
| 305 |
-
"step": 320
|
| 306 |
-
},
|
| 307 |
-
{
|
| 308 |
-
"epoch": 8.627450980392156,
|
| 309 |
-
"grad_norm": 3.775691270828247,
|
| 310 |
-
"learning_rate": 3.9473684210526316e-05,
|
| 311 |
-
"loss": 0.3115,
|
| 312 |
-
"step": 330
|
| 313 |
-
},
|
| 314 |
-
{
|
| 315 |
-
"epoch": 8.88888888888889,
|
| 316 |
-
"grad_norm": 3.285763740539551,
|
| 317 |
-
"learning_rate": 3.898635477582846e-05,
|
| 318 |
-
"loss": 0.3811,
|
| 319 |
-
"step": 340
|
| 320 |
-
},
|
| 321 |
-
{
|
| 322 |
-
"epoch": 8.993464052287582,
|
| 323 |
-
"eval_accuracy": 0.8676470588235294,
|
| 324 |
-
"eval_loss": 0.4429008662700653,
|
| 325 |
-
"eval_runtime": 32.2721,
|
| 326 |
-
"eval_samples_per_second": 4.214,
|
| 327 |
-
"eval_steps_per_second": 0.527,
|
| 328 |
-
"step": 344
|
| 329 |
-
},
|
| 330 |
-
{
|
| 331 |
-
"epoch": 9.15032679738562,
|
| 332 |
-
"grad_norm": 2.8171193599700928,
|
| 333 |
-
"learning_rate": 3.849902534113061e-05,
|
| 334 |
-
"loss": 0.3617,
|
| 335 |
-
"step": 350
|
| 336 |
-
},
|
| 337 |
-
{
|
| 338 |
-
"epoch": 9.411764705882353,
|
| 339 |
-
"grad_norm": 2.766669273376465,
|
| 340 |
-
"learning_rate": 3.8011695906432746e-05,
|
| 341 |
-
"loss": 0.2679,
|
| 342 |
-
"step": 360
|
| 343 |
-
},
|
| 344 |
-
{
|
| 345 |
-
"epoch": 9.673202614379084,
|
| 346 |
-
"grad_norm": 5.8598833084106445,
|
| 347 |
-
"learning_rate": 3.75243664717349e-05,
|
| 348 |
-
"loss": 0.2909,
|
| 349 |
-
"step": 370
|
| 350 |
-
},
|
| 351 |
-
{
|
| 352 |
-
"epoch": 9.934640522875817,
|
| 353 |
-
"grad_norm": 4.5529608726501465,
|
| 354 |
-
"learning_rate": 3.7037037037037037e-05,
|
| 355 |
-
"loss": 0.2341,
|
| 356 |
-
"step": 380
|
| 357 |
-
},
|
| 358 |
-
{
|
| 359 |
-
"epoch": 9.986928104575163,
|
| 360 |
-
"eval_accuracy": 0.875,
|
| 361 |
-
"eval_loss": 0.42220985889434814,
|
| 362 |
-
"eval_runtime": 32.0969,
|
| 363 |
-
"eval_samples_per_second": 4.237,
|
| 364 |
-
"eval_steps_per_second": 0.53,
|
| 365 |
-
"step": 382
|
| 366 |
-
},
|
| 367 |
-
{
|
| 368 |
-
"epoch": 10.196078431372548,
|
| 369 |
-
"grad_norm": 3.98348069190979,
|
| 370 |
-
"learning_rate": 3.654970760233918e-05,
|
| 371 |
-
"loss": 0.2612,
|
| 372 |
-
"step": 390
|
| 373 |
},
|
| 374 |
{
|
| 375 |
-
"epoch":
|
| 376 |
-
"
|
| 377 |
-
"
|
| 378 |
-
"
|
| 379 |
-
"
|
|
|
|
|
|
|
| 380 |
},
|
| 381 |
{
|
| 382 |
-
"epoch":
|
| 383 |
-
"grad_norm": 3.
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss": 0.
|
| 386 |
-
"step":
|
| 387 |
-
},
|
| 388 |
-
{
|
| 389 |
-
"epoch": 10.980392156862745,
|
| 390 |
-
"grad_norm": 6.444194793701172,
|
| 391 |
-
"learning_rate": 3.508771929824561e-05,
|
| 392 |
-
"loss": 0.3082,
|
| 393 |
-
"step": 420
|
| 394 |
},
|
| 395 |
{
|
| 396 |
-
"epoch":
|
| 397 |
-
"eval_accuracy": 0.
|
| 398 |
-
"eval_loss": 0.
|
| 399 |
-
"eval_runtime":
|
| 400 |
-
"eval_samples_per_second": 4.
|
| 401 |
-
"eval_steps_per_second": 0.
|
| 402 |
-
"step":
|
| 403 |
},
|
| 404 |
{
|
| 405 |
-
"epoch":
|
| 406 |
-
"grad_norm":
|
| 407 |
-
"learning_rate":
|
| 408 |
-
"loss": 0.
|
| 409 |
-
"step":
|
| 410 |
},
|
| 411 |
{
|
| 412 |
-
"epoch":
|
| 413 |
-
"
|
| 414 |
-
"
|
| 415 |
-
"
|
| 416 |
-
"
|
|
|
|
|
|
|
| 417 |
},
|
| 418 |
{
|
| 419 |
-
"epoch": 11.
|
| 420 |
-
"grad_norm":
|
| 421 |
-
"learning_rate":
|
| 422 |
-
"loss": 0.
|
| 423 |
-
"step":
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 12.0,
|
| 427 |
-
"eval_accuracy": 0.
|
| 428 |
-
"eval_loss": 0.
|
| 429 |
-
"eval_runtime":
|
| 430 |
-
"eval_samples_per_second": 4.
|
| 431 |
-
"eval_steps_per_second": 0.
|
| 432 |
-
"step":
|
| 433 |
-
},
|
| 434 |
-
{
|
| 435 |
-
"epoch": 12.
|
| 436 |
-
"grad_norm":
|
| 437 |
-
"learning_rate":
|
| 438 |
-
"loss": 0.
|
| 439 |
-
"step":
|
| 440 |
-
},
|
| 441 |
-
{
|
| 442 |
-
"epoch": 12.287581699346406,
|
| 443 |
-
"grad_norm": 3.5303471088409424,
|
| 444 |
-
"learning_rate": 3.265107212475634e-05,
|
| 445 |
-
"loss": 0.23,
|
| 446 |
-
"step": 470
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 12.549019607843137,
|
| 450 |
-
"grad_norm": 4.927082538604736,
|
| 451 |
-
"learning_rate": 3.216374269005848e-05,
|
| 452 |
-
"loss": 0.1693,
|
| 453 |
-
"step": 480
|
| 454 |
-
},
|
| 455 |
-
{
|
| 456 |
-
"epoch": 12.81045751633987,
|
| 457 |
-
"grad_norm": 1.5047080516815186,
|
| 458 |
-
"learning_rate": 3.167641325536063e-05,
|
| 459 |
-
"loss": 0.2374,
|
| 460 |
-
"step": 490
|
| 461 |
-
},
|
| 462 |
-
{
|
| 463 |
-
"epoch": 12.993464052287582,
|
| 464 |
-
"eval_accuracy": 0.875,
|
| 465 |
-
"eval_loss": 0.42333418130874634,
|
| 466 |
-
"eval_runtime": 32.071,
|
| 467 |
-
"eval_samples_per_second": 4.241,
|
| 468 |
-
"eval_steps_per_second": 0.53,
|
| 469 |
-
"step": 497
|
| 470 |
},
|
| 471 |
{
|
| 472 |
-
"epoch":
|
| 473 |
-
"
|
| 474 |
-
"
|
| 475 |
-
"
|
| 476 |
-
"
|
|
|
|
|
|
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 13.333333333333334,
|
| 480 |
-
"grad_norm":
|
| 481 |
-
"learning_rate": 3.
|
| 482 |
-
"loss": 0.
|
| 483 |
-
"step":
|
| 484 |
-
},
|
| 485 |
-
{
|
| 486 |
-
"epoch": 13.594771241830065,
|
| 487 |
-
"grad_norm": 0.8992928266525269,
|
| 488 |
-
"learning_rate": 3.0214424951267055e-05,
|
| 489 |
-
"loss": 0.1784,
|
| 490 |
-
"step": 520
|
| 491 |
-
},
|
| 492 |
-
{
|
| 493 |
-
"epoch": 13.856209150326798,
|
| 494 |
-
"grad_norm": 1.4664764404296875,
|
| 495 |
-
"learning_rate": 2.9727095516569204e-05,
|
| 496 |
-
"loss": 0.128,
|
| 497 |
-
"step": 530
|
| 498 |
-
},
|
| 499 |
-
{
|
| 500 |
-
"epoch": 13.986928104575163,
|
| 501 |
-
"eval_accuracy": 0.8970588235294118,
|
| 502 |
-
"eval_loss": 0.367097944021225,
|
| 503 |
-
"eval_runtime": 31.5031,
|
| 504 |
-
"eval_samples_per_second": 4.317,
|
| 505 |
-
"eval_steps_per_second": 0.54,
|
| 506 |
-
"step": 535
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch":
|
| 510 |
-
"
|
| 511 |
-
"
|
| 512 |
-
"
|
| 513 |
-
"
|
|
|
|
|
|
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch": 14.
|
| 517 |
-
"grad_norm":
|
| 518 |
-
"learning_rate":
|
| 519 |
-
"loss": 0.
|
| 520 |
-
"step":
|
| 521 |
},
|
| 522 |
{
|
| 523 |
-
"epoch": 14.
|
| 524 |
-
"
|
| 525 |
-
"
|
| 526 |
-
"
|
| 527 |
-
"
|
|
|
|
|
|
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch":
|
| 531 |
-
"grad_norm":
|
| 532 |
-
"learning_rate":
|
| 533 |
-
"loss": 0.
|
| 534 |
-
"step":
|
| 535 |
},
|
| 536 |
{
|
| 537 |
-
"epoch":
|
| 538 |
"eval_accuracy": 0.8970588235294118,
|
| 539 |
-
"eval_loss": 0.
|
| 540 |
-
"eval_runtime":
|
| 541 |
-
"eval_samples_per_second": 4.
|
| 542 |
-
"eval_steps_per_second": 0.
|
| 543 |
-
"step":
|
| 544 |
-
},
|
| 545 |
-
{
|
| 546 |
-
"epoch": 15.163398692810457,
|
| 547 |
-
"grad_norm": 4.025688171386719,
|
| 548 |
-
"learning_rate": 2.729044834307992e-05,
|
| 549 |
-
"loss": 0.1681,
|
| 550 |
-
"step": 580
|
| 551 |
-
},
|
| 552 |
-
{
|
| 553 |
-
"epoch": 15.42483660130719,
|
| 554 |
-
"grad_norm": 0.5761296153068542,
|
| 555 |
-
"learning_rate": 2.680311890838207e-05,
|
| 556 |
-
"loss": 0.1558,
|
| 557 |
-
"step": 590
|
| 558 |
},
|
| 559 |
{
|
| 560 |
-
"epoch":
|
| 561 |
-
"grad_norm": 3.
|
| 562 |
-
"learning_rate":
|
| 563 |
-
"loss": 0.
|
| 564 |
-
"step":
|
| 565 |
},
|
| 566 |
{
|
| 567 |
-
"epoch":
|
| 568 |
-
"
|
| 569 |
-
"
|
| 570 |
-
"
|
| 571 |
-
"
|
|
|
|
|
|
|
| 572 |
},
|
| 573 |
{
|
| 574 |
-
"epoch":
|
| 575 |
-
"
|
| 576 |
-
"
|
| 577 |
-
"
|
| 578 |
-
"
|
| 579 |
-
"eval_steps_per_second": 0.531,
|
| 580 |
-
"step": 612
|
| 581 |
},
|
| 582 |
{
|
| 583 |
-
"epoch":
|
| 584 |
-
"
|
| 585 |
-
"
|
| 586 |
-
"
|
| 587 |
-
"
|
|
|
|
|
|
|
| 588 |
},
|
| 589 |
{
|
| 590 |
-
"epoch":
|
| 591 |
-
"grad_norm":
|
| 592 |
-
"learning_rate":
|
| 593 |
-
"loss": 0.
|
| 594 |
-
"step":
|
| 595 |
},
|
| 596 |
{
|
| 597 |
-
"epoch":
|
| 598 |
-
"
|
| 599 |
-
"
|
| 600 |
-
"
|
| 601 |
-
"
|
|
|
|
|
|
|
| 602 |
},
|
| 603 |
{
|
| 604 |
-
"epoch":
|
| 605 |
-
"grad_norm":
|
| 606 |
-
"learning_rate":
|
| 607 |
-
"loss": 0.
|
| 608 |
-
"step":
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch":
|
| 612 |
"eval_accuracy": 0.9117647058823529,
|
| 613 |
-
"eval_loss": 0.
|
| 614 |
-
"eval_runtime":
|
| 615 |
-
"eval_samples_per_second": 4.
|
| 616 |
-
"eval_steps_per_second": 0.
|
| 617 |
-
"step":
|
| 618 |
},
|
| 619 |
{
|
| 620 |
-
"epoch":
|
| 621 |
-
"grad_norm":
|
| 622 |
-
"learning_rate":
|
| 623 |
-
"loss": 0.
|
| 624 |
-
"step":
|
| 625 |
-
},
|
| 626 |
-
{
|
| 627 |
-
"epoch": 17.516339869281047,
|
| 628 |
-
"grad_norm": 0.38019096851348877,
|
| 629 |
-
"learning_rate": 2.290448343079922e-05,
|
| 630 |
-
"loss": 0.1058,
|
| 631 |
-
"step": 670
|
| 632 |
-
},
|
| 633 |
-
{
|
| 634 |
-
"epoch": 17.77777777777778,
|
| 635 |
-
"grad_norm": 9.7896089553833,
|
| 636 |
-
"learning_rate": 2.2417153996101365e-05,
|
| 637 |
-
"loss": 0.1408,
|
| 638 |
-
"step": 680
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch":
|
| 642 |
"eval_accuracy": 0.9044117647058824,
|
| 643 |
-
"eval_loss": 0.
|
| 644 |
-
"eval_runtime":
|
| 645 |
-
"eval_samples_per_second": 4.
|
| 646 |
-
"eval_steps_per_second": 0.
|
| 647 |
-
"step":
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 18.03921568627451,
|
| 651 |
-
"grad_norm": 1.3320598602294922,
|
| 652 |
-
"learning_rate": 2.1929824561403507e-05,
|
| 653 |
-
"loss": 0.1352,
|
| 654 |
-
"step": 690
|
| 655 |
},
|
| 656 |
{
|
| 657 |
-
"epoch":
|
| 658 |
-
"grad_norm":
|
| 659 |
-
"learning_rate": 2.
|
| 660 |
-
"loss": 0.
|
| 661 |
-
"step":
|
| 662 |
},
|
| 663 |
{
|
| 664 |
-
"epoch":
|
| 665 |
-
"
|
| 666 |
-
"
|
| 667 |
-
"
|
| 668 |
-
"
|
|
|
|
|
|
|
| 669 |
},
|
| 670 |
{
|
| 671 |
-
"epoch":
|
| 672 |
-
"grad_norm":
|
| 673 |
-
"learning_rate": 2.
|
| 674 |
-
"loss": 0.
|
| 675 |
-
"step":
|
| 676 |
},
|
| 677 |
{
|
| 678 |
-
"epoch":
|
| 679 |
-
"eval_accuracy": 0.
|
| 680 |
-
"eval_loss": 0.
|
| 681 |
-
"eval_runtime":
|
| 682 |
-
"eval_samples_per_second": 4.
|
| 683 |
-
"eval_steps_per_second": 0.
|
| 684 |
-
"step":
|
| 685 |
},
|
| 686 |
{
|
| 687 |
-
"epoch":
|
| 688 |
-
"grad_norm":
|
| 689 |
-
"learning_rate":
|
| 690 |
-
"loss": 0.
|
| 691 |
-
"step":
|
| 692 |
},
|
| 693 |
{
|
| 694 |
-
"epoch":
|
| 695 |
-
"
|
| 696 |
-
"
|
| 697 |
-
"
|
| 698 |
-
"
|
|
|
|
|
|
|
| 699 |
},
|
| 700 |
{
|
| 701 |
-
"epoch":
|
| 702 |
-
"grad_norm":
|
| 703 |
-
"learning_rate":
|
| 704 |
-
"loss": 0.
|
| 705 |
-
"step":
|
| 706 |
},
|
| 707 |
{
|
| 708 |
-
"epoch":
|
| 709 |
-
"
|
| 710 |
-
"
|
| 711 |
-
"
|
| 712 |
-
"
|
|
|
|
|
|
|
| 713 |
},
|
| 714 |
{
|
| 715 |
-
"epoch":
|
| 716 |
-
"
|
| 717 |
-
"
|
| 718 |
-
"
|
| 719 |
-
"
|
| 720 |
-
"eval_steps_per_second": 0.535,
|
| 721 |
-
"step": 765
|
| 722 |
},
|
| 723 |
{
|
| 724 |
-
"epoch":
|
| 725 |
-
"
|
| 726 |
-
"
|
| 727 |
-
"
|
| 728 |
-
"
|
|
|
|
|
|
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch":
|
| 732 |
-
"grad_norm": 3.
|
| 733 |
-
"learning_rate":
|
| 734 |
-
"loss": 0.
|
| 735 |
-
"step":
|
| 736 |
},
|
| 737 |
{
|
| 738 |
-
"epoch":
|
| 739 |
-
"
|
| 740 |
-
"
|
| 741 |
-
"
|
| 742 |
-
"
|
|
|
|
|
|
|
| 743 |
},
|
| 744 |
{
|
| 745 |
-
"epoch":
|
| 746 |
-
"grad_norm":
|
| 747 |
-
"learning_rate":
|
| 748 |
-
"loss": 0.
|
| 749 |
-
"step":
|
| 750 |
},
|
| 751 |
{
|
| 752 |
-
"epoch":
|
| 753 |
"eval_accuracy": 0.8823529411764706,
|
| 754 |
-
"eval_loss": 0.
|
| 755 |
-
"eval_runtime":
|
| 756 |
-
"eval_samples_per_second": 4.
|
| 757 |
-
"eval_steps_per_second": 0.
|
| 758 |
-
"step":
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 21.176470588235293,
|
| 762 |
-
"grad_norm": 2.389382839202881,
|
| 763 |
-
"learning_rate": 1.608187134502924e-05,
|
| 764 |
-
"loss": 0.1316,
|
| 765 |
-
"step": 810
|
| 766 |
},
|
| 767 |
{
|
| 768 |
-
"epoch":
|
| 769 |
-
"grad_norm":
|
| 770 |
-
"learning_rate":
|
| 771 |
-
"loss": 0.
|
| 772 |
-
"step":
|
| 773 |
},
|
| 774 |
{
|
| 775 |
-
"epoch":
|
| 776 |
-
"
|
| 777 |
-
"
|
| 778 |
-
"
|
| 779 |
-
"
|
|
|
|
|
|
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch":
|
| 783 |
-
"grad_norm":
|
| 784 |
-
"learning_rate": 1.
|
| 785 |
-
"loss": 0.
|
| 786 |
-
"step":
|
| 787 |
},
|
| 788 |
{
|
| 789 |
-
"epoch":
|
| 790 |
-
"eval_accuracy": 0.
|
| 791 |
-
"eval_loss": 0.
|
| 792 |
-
"eval_runtime":
|
| 793 |
-
"eval_samples_per_second": 4.
|
| 794 |
-
"eval_steps_per_second": 0.
|
| 795 |
-
"step":
|
| 796 |
},
|
| 797 |
{
|
| 798 |
-
"epoch":
|
| 799 |
-
"grad_norm":
|
| 800 |
-
"learning_rate": 1.
|
| 801 |
-
"loss": 0.
|
| 802 |
-
"step":
|
| 803 |
},
|
| 804 |
{
|
| 805 |
-
"epoch":
|
| 806 |
-
"
|
| 807 |
-
"
|
| 808 |
-
"
|
| 809 |
-
"
|
|
|
|
|
|
|
| 810 |
},
|
| 811 |
{
|
| 812 |
-
"epoch":
|
| 813 |
-
"grad_norm":
|
| 814 |
-
"learning_rate": 1.
|
| 815 |
-
"loss": 0.
|
| 816 |
-
"step":
|
| 817 |
},
|
| 818 |
{
|
| 819 |
-
"epoch":
|
| 820 |
"eval_accuracy": 0.9191176470588235,
|
| 821 |
-
"eval_loss": 0.
|
| 822 |
-
"eval_runtime":
|
| 823 |
-
"eval_samples_per_second": 4.
|
| 824 |
-
"eval_steps_per_second": 0.
|
| 825 |
-
"step":
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 23.00653594771242,
|
| 829 |
-
"grad_norm": 1.2901531457901,
|
| 830 |
-
"learning_rate": 1.267056530214425e-05,
|
| 831 |
-
"loss": 0.0937,
|
| 832 |
-
"step": 880
|
| 833 |
},
|
| 834 |
{
|
| 835 |
-
"epoch":
|
| 836 |
-
"grad_norm":
|
| 837 |
-
"learning_rate": 1.
|
| 838 |
-
"loss": 0.
|
| 839 |
-
"step":
|
| 840 |
-
},
|
| 841 |
-
{
|
| 842 |
-
"epoch": 23.529411764705884,
|
| 843 |
-
"grad_norm": 0.1317284107208252,
|
| 844 |
-
"learning_rate": 1.1695906432748537e-05,
|
| 845 |
-
"loss": 0.0742,
|
| 846 |
-
"step": 900
|
| 847 |
-
},
|
| 848 |
-
{
|
| 849 |
-
"epoch": 23.790849673202615,
|
| 850 |
-
"grad_norm": 1.982415795326233,
|
| 851 |
-
"learning_rate": 1.1208576998050683e-05,
|
| 852 |
-
"loss": 0.0769,
|
| 853 |
-
"step": 910
|
| 854 |
},
|
| 855 |
{
|
| 856 |
-
"epoch":
|
| 857 |
"eval_accuracy": 0.9264705882352942,
|
| 858 |
-
"eval_loss": 0.
|
| 859 |
-
"eval_runtime":
|
| 860 |
-
"eval_samples_per_second": 4.
|
| 861 |
-
"eval_steps_per_second": 0.
|
| 862 |
-
"step":
|
| 863 |
},
|
| 864 |
{
|
| 865 |
-
"epoch":
|
| 866 |
-
"grad_norm":
|
| 867 |
-
"learning_rate": 1.
|
| 868 |
-
"loss": 0.
|
| 869 |
-
"step":
|
| 870 |
},
|
| 871 |
{
|
| 872 |
-
"epoch":
|
| 873 |
-
"
|
| 874 |
-
"
|
| 875 |
-
"
|
| 876 |
-
"
|
|
|
|
|
|
|
| 877 |
},
|
| 878 |
{
|
| 879 |
-
"epoch":
|
| 880 |
-
"grad_norm": 1.
|
| 881 |
-
"learning_rate":
|
| 882 |
-
"loss": 0.
|
| 883 |
-
"step":
|
| 884 |
},
|
| 885 |
{
|
| 886 |
-
"epoch":
|
| 887 |
-
"
|
| 888 |
-
"
|
| 889 |
-
"
|
| 890 |
-
"
|
|
|
|
|
|
|
| 891 |
},
|
| 892 |
{
|
| 893 |
-
"epoch":
|
| 894 |
-
"
|
| 895 |
-
"
|
| 896 |
-
"
|
| 897 |
-
"
|
| 898 |
-
"eval_steps_per_second": 0.529,
|
| 899 |
-
"step": 956
|
| 900 |
},
|
| 901 |
{
|
| 902 |
-
"epoch":
|
| 903 |
-
"
|
| 904 |
-
"
|
| 905 |
-
"
|
| 906 |
-
"
|
|
|
|
|
|
|
| 907 |
},
|
| 908 |
{
|
| 909 |
-
"epoch":
|
| 910 |
-
"grad_norm":
|
| 911 |
-
"learning_rate":
|
| 912 |
-
"loss": 0.
|
| 913 |
-
"step":
|
| 914 |
},
|
| 915 |
{
|
| 916 |
-
"epoch":
|
| 917 |
-
"
|
| 918 |
-
"
|
| 919 |
-
"
|
| 920 |
-
"
|
|
|
|
|
|
|
| 921 |
},
|
| 922 |
{
|
| 923 |
-
"epoch":
|
| 924 |
-
"grad_norm":
|
| 925 |
-
"learning_rate":
|
| 926 |
-
"loss": 0.
|
| 927 |
-
"step":
|
| 928 |
},
|
| 929 |
{
|
| 930 |
-
"epoch":
|
| 931 |
"eval_accuracy": 0.9264705882352942,
|
| 932 |
-
"eval_loss": 0.
|
| 933 |
-
"eval_runtime":
|
| 934 |
-
"eval_samples_per_second": 4.
|
| 935 |
-
"eval_steps_per_second": 0.
|
| 936 |
-
"step":
|
| 937 |
-
},
|
| 938 |
-
{
|
| 939 |
-
"epoch": 26.143790849673202,
|
| 940 |
-
"grad_norm": 1.0143533945083618,
|
| 941 |
-
"learning_rate": 6.82261208576998e-06,
|
| 942 |
-
"loss": 0.0604,
|
| 943 |
-
"step": 1000
|
| 944 |
},
|
| 945 |
{
|
| 946 |
-
"epoch":
|
| 947 |
-
"grad_norm": 0.
|
| 948 |
-
"learning_rate":
|
| 949 |
-
"loss": 0.
|
| 950 |
-
"step":
|
| 951 |
},
|
| 952 |
{
|
| 953 |
-
"epoch":
|
| 954 |
-
"
|
| 955 |
-
"
|
| 956 |
-
"
|
| 957 |
-
"
|
|
|
|
|
|
|
| 958 |
},
|
| 959 |
{
|
| 960 |
-
"epoch":
|
| 961 |
-
"grad_norm":
|
| 962 |
-
"learning_rate":
|
| 963 |
-
"loss": 0.
|
| 964 |
-
"step":
|
| 965 |
},
|
| 966 |
{
|
| 967 |
-
"epoch":
|
| 968 |
-
"eval_accuracy": 0.
|
| 969 |
-
"eval_loss": 0.
|
| 970 |
-
"eval_runtime":
|
| 971 |
-
"eval_samples_per_second": 4.
|
| 972 |
-
"eval_steps_per_second": 0.
|
| 973 |
-
"step":
|
| 974 |
},
|
| 975 |
{
|
| 976 |
-
"epoch":
|
| 977 |
-
"
|
| 978 |
-
"
|
| 979 |
-
"
|
| 980 |
-
"
|
|
|
|
|
|
|
| 981 |
},
|
| 982 |
{
|
| 983 |
-
"epoch":
|
| 984 |
-
"grad_norm": 0.
|
| 985 |
-
"learning_rate":
|
| 986 |
-
"loss": 0.
|
| 987 |
-
"step":
|
| 988 |
},
|
| 989 |
{
|
| 990 |
-
"epoch":
|
| 991 |
-
"
|
| 992 |
-
"
|
| 993 |
-
"
|
| 994 |
-
"
|
|
|
|
|
|
|
| 995 |
},
|
| 996 |
{
|
| 997 |
-
"epoch":
|
| 998 |
-
"grad_norm":
|
| 999 |
-
"learning_rate":
|
| 1000 |
-
"loss": 0.
|
| 1001 |
-
"step":
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
-
"epoch":
|
| 1005 |
-
"eval_accuracy": 0.
|
| 1006 |
-
"eval_loss": 0.
|
| 1007 |
-
"eval_runtime":
|
| 1008 |
-
"eval_samples_per_second": 4.
|
| 1009 |
-
"eval_steps_per_second": 0.
|
| 1010 |
-
"step":
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
-
"epoch":
|
| 1014 |
-
"grad_norm":
|
| 1015 |
-
"learning_rate":
|
| 1016 |
-
"loss": 0.
|
| 1017 |
-
"step":
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
-
"epoch":
|
| 1021 |
-
"
|
| 1022 |
-
"
|
| 1023 |
-
"
|
| 1024 |
-
"
|
|
|
|
|
|
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
-
"epoch":
|
| 1028 |
-
"grad_norm":
|
| 1029 |
-
"learning_rate":
|
| 1030 |
-
"loss": 0.
|
| 1031 |
-
"step":
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
-
"epoch":
|
| 1035 |
-
"eval_accuracy": 0.
|
| 1036 |
-
"eval_loss": 0.
|
| 1037 |
-
"eval_runtime":
|
| 1038 |
-
"eval_samples_per_second": 4.
|
| 1039 |
-
"eval_steps_per_second": 0.
|
| 1040 |
-
"step":
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
-
"epoch":
|
| 1044 |
-
"grad_norm":
|
| 1045 |
-
"learning_rate": 1.
|
| 1046 |
-
"loss": 0.
|
| 1047 |
-
"step":
|
| 1048 |
},
|
| 1049 |
{
|
| 1050 |
-
"epoch":
|
| 1051 |
-
"
|
| 1052 |
-
"
|
| 1053 |
-
"
|
| 1054 |
-
"
|
|
|
|
|
|
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
-
"epoch":
|
| 1058 |
-
"grad_norm":
|
| 1059 |
-
"learning_rate":
|
| 1060 |
-
"loss": 0.
|
| 1061 |
-
"step":
|
| 1062 |
},
|
| 1063 |
{
|
| 1064 |
-
"epoch":
|
| 1065 |
-
"
|
| 1066 |
-
"
|
| 1067 |
-
"
|
| 1068 |
-
"
|
|
|
|
|
|
|
| 1069 |
},
|
| 1070 |
{
|
| 1071 |
-
"epoch":
|
| 1072 |
-
"
|
| 1073 |
-
"
|
| 1074 |
-
"
|
| 1075 |
-
"
|
| 1076 |
-
"
|
| 1077 |
-
"
|
| 1078 |
-
},
|
| 1079 |
-
{
|
| 1080 |
-
"epoch": 29.80392156862745,
|
| 1081 |
-
"step": 1140,
|
| 1082 |
-
"total_flos": 2.827009127861453e+18,
|
| 1083 |
-
"train_loss": 0.3311853929046999,
|
| 1084 |
-
"train_runtime": 10131.822,
|
| 1085 |
-
"train_samples_per_second": 3.624,
|
| 1086 |
-
"train_steps_per_second": 0.113
|
| 1087 |
}
|
| 1088 |
],
|
| 1089 |
"logging_steps": 10,
|
| 1090 |
-
"max_steps":
|
| 1091 |
"num_input_tokens_seen": 0,
|
| 1092 |
-
"num_train_epochs":
|
| 1093 |
"save_steps": 500,
|
| 1094 |
"stateful_callbacks": {
|
| 1095 |
"TrainerControl": {
|
|
@@ -1103,8 +773,8 @@
|
|
| 1103 |
"attributes": {}
|
| 1104 |
}
|
| 1105 |
},
|
| 1106 |
-
"total_flos":
|
| 1107 |
-
"train_batch_size":
|
| 1108 |
"trial_name": null,
|
| 1109 |
"trial_params": null
|
| 1110 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.9338235294117647,
|
| 3 |
+
"best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-papsmear/checkpoint-341",
|
| 4 |
+
"epoch": 46.15384615384615,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 450,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.9230769230769231,
|
| 13 |
+
"eval_accuracy": 0.20588235294117646,
|
| 14 |
+
"eval_loss": 1.7879393100738525,
|
| 15 |
+
"eval_runtime": 31.0661,
|
| 16 |
+
"eval_samples_per_second": 4.378,
|
| 17 |
+
"eval_steps_per_second": 0.161,
|
| 18 |
+
"step": 9
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
},
|
| 20 |
{
|
| 21 |
+
"epoch": 1.0256410256410255,
|
| 22 |
+
"grad_norm": 1.2670034170150757,
|
| 23 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 24 |
+
"loss": 1.8037,
|
| 25 |
+
"step": 10
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
},
|
| 27 |
{
|
| 28 |
+
"epoch": 1.9487179487179487,
|
| 29 |
+
"eval_accuracy": 0.40441176470588236,
|
| 30 |
+
"eval_loss": 1.648475170135498,
|
| 31 |
+
"eval_runtime": 32.5892,
|
| 32 |
+
"eval_samples_per_second": 4.173,
|
| 33 |
+
"eval_steps_per_second": 0.153,
|
| 34 |
+
"step": 19
|
| 35 |
},
|
| 36 |
{
|
| 37 |
+
"epoch": 2.051282051282051,
|
| 38 |
+
"grad_norm": 1.1165878772735596,
|
| 39 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 40 |
+
"loss": 1.6961,
|
| 41 |
+
"step": 20
|
| 42 |
},
|
| 43 |
{
|
| 44 |
+
"epoch": 2.9743589743589745,
|
| 45 |
+
"eval_accuracy": 0.39705882352941174,
|
| 46 |
+
"eval_loss": 1.4881964921951294,
|
| 47 |
+
"eval_runtime": 31.9851,
|
| 48 |
+
"eval_samples_per_second": 4.252,
|
| 49 |
+
"eval_steps_per_second": 0.156,
|
| 50 |
+
"step": 29
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 3.076923076923077,
|
| 54 |
+
"grad_norm": 1.1624480485916138,
|
| 55 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 56 |
+
"loss": 1.5407,
|
| 57 |
+
"step": 30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 4.0,
|
| 61 |
+
"eval_accuracy": 0.5220588235294118,
|
| 62 |
+
"eval_loss": 1.3068838119506836,
|
| 63 |
+
"eval_runtime": 31.423,
|
| 64 |
+
"eval_samples_per_second": 4.328,
|
| 65 |
+
"eval_steps_per_second": 0.159,
|
| 66 |
+
"step": 39
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 4.102564102564102,
|
| 70 |
+
"grad_norm": 0.9998506307601929,
|
| 71 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 72 |
+
"loss": 1.3308,
|
| 73 |
+
"step": 40
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
},
|
| 75 |
{
|
| 76 |
+
"epoch": 4.923076923076923,
|
| 77 |
+
"eval_accuracy": 0.6029411764705882,
|
| 78 |
+
"eval_loss": 1.1338790655136108,
|
| 79 |
+
"eval_runtime": 30.9165,
|
| 80 |
+
"eval_samples_per_second": 4.399,
|
| 81 |
+
"eval_steps_per_second": 0.162,
|
| 82 |
+
"step": 48
|
| 83 |
},
|
| 84 |
{
|
| 85 |
+
"epoch": 5.128205128205128,
|
| 86 |
+
"grad_norm": 1.2114465236663818,
|
| 87 |
+
"learning_rate": 4.938271604938271e-05,
|
| 88 |
+
"loss": 1.1074,
|
| 89 |
+
"step": 50
|
|
|
|
|
|
|
| 90 |
},
|
| 91 |
{
|
| 92 |
+
"epoch": 5.948717948717949,
|
| 93 |
+
"eval_accuracy": 0.75,
|
| 94 |
+
"eval_loss": 0.9395625591278076,
|
| 95 |
+
"eval_runtime": 30.6345,
|
| 96 |
+
"eval_samples_per_second": 4.439,
|
| 97 |
+
"eval_steps_per_second": 0.163,
|
| 98 |
+
"step": 58
|
| 99 |
},
|
| 100 |
{
|
| 101 |
+
"epoch": 6.153846153846154,
|
| 102 |
+
"grad_norm": 1.676944375038147,
|
| 103 |
+
"learning_rate": 4.814814814814815e-05,
|
| 104 |
+
"loss": 0.9162,
|
| 105 |
+
"step": 60
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 6.9743589743589745,
|
| 109 |
+
"eval_accuracy": 0.7647058823529411,
|
| 110 |
+
"eval_loss": 0.8551253080368042,
|
| 111 |
+
"eval_runtime": 30.343,
|
| 112 |
+
"eval_samples_per_second": 4.482,
|
| 113 |
+
"eval_steps_per_second": 0.165,
|
| 114 |
+
"step": 68
|
| 115 |
},
|
| 116 |
{
|
| 117 |
+
"epoch": 7.17948717948718,
|
| 118 |
+
"grad_norm": 1.8491935729980469,
|
| 119 |
+
"learning_rate": 4.691358024691358e-05,
|
| 120 |
+
"loss": 0.8174,
|
| 121 |
+
"step": 70
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 8.0,
|
| 125 |
+
"eval_accuracy": 0.7573529411764706,
|
| 126 |
+
"eval_loss": 0.8290878534317017,
|
| 127 |
+
"eval_runtime": 30.401,
|
| 128 |
+
"eval_samples_per_second": 4.474,
|
| 129 |
+
"eval_steps_per_second": 0.164,
|
| 130 |
+
"step": 78
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 8.205128205128204,
|
| 134 |
+
"grad_norm": 1.4983110427856445,
|
| 135 |
+
"learning_rate": 4.567901234567901e-05,
|
| 136 |
+
"loss": 0.7135,
|
| 137 |
+
"step": 80
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 138 |
},
|
| 139 |
{
|
| 140 |
+
"epoch": 8.923076923076923,
|
| 141 |
+
"eval_accuracy": 0.7941176470588235,
|
| 142 |
+
"eval_loss": 0.7505079507827759,
|
| 143 |
+
"eval_runtime": 31.6046,
|
| 144 |
+
"eval_samples_per_second": 4.303,
|
| 145 |
+
"eval_steps_per_second": 0.158,
|
| 146 |
+
"step": 87
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 9.23076923076923,
|
| 150 |
+
"grad_norm": 3.118570566177368,
|
| 151 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 152 |
+
"loss": 0.6222,
|
| 153 |
+
"step": 90
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 9.948717948717949,
|
| 157 |
+
"eval_accuracy": 0.8455882352941176,
|
| 158 |
+
"eval_loss": 0.6433916687965393,
|
| 159 |
+
"eval_runtime": 30.9566,
|
| 160 |
+
"eval_samples_per_second": 4.393,
|
| 161 |
+
"eval_steps_per_second": 0.162,
|
| 162 |
+
"step": 97
|
| 163 |
},
|
| 164 |
{
|
| 165 |
+
"epoch": 10.256410256410255,
|
| 166 |
+
"grad_norm": 1.8012874126434326,
|
| 167 |
+
"learning_rate": 4.3209876543209875e-05,
|
| 168 |
+
"loss": 0.5445,
|
| 169 |
+
"step": 100
|
| 170 |
},
|
| 171 |
{
|
| 172 |
+
"epoch": 10.974358974358974,
|
| 173 |
+
"eval_accuracy": 0.8529411764705882,
|
| 174 |
+
"eval_loss": 0.5995692610740662,
|
| 175 |
+
"eval_runtime": 31.2634,
|
| 176 |
+
"eval_samples_per_second": 4.35,
|
| 177 |
+
"eval_steps_per_second": 0.16,
|
| 178 |
+
"step": 107
|
| 179 |
},
|
| 180 |
{
|
| 181 |
+
"epoch": 11.282051282051283,
|
| 182 |
+
"grad_norm": 2.380070447921753,
|
| 183 |
+
"learning_rate": 4.197530864197531e-05,
|
| 184 |
+
"loss": 0.4935,
|
| 185 |
+
"step": 110
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 12.0,
|
| 189 |
+
"eval_accuracy": 0.8529411764705882,
|
| 190 |
+
"eval_loss": 0.5513917207717896,
|
| 191 |
+
"eval_runtime": 30.5663,
|
| 192 |
+
"eval_samples_per_second": 4.449,
|
| 193 |
+
"eval_steps_per_second": 0.164,
|
| 194 |
+
"step": 117
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 12.307692307692308,
|
| 198 |
+
"grad_norm": 2.3905367851257324,
|
| 199 |
+
"learning_rate": 4.074074074074074e-05,
|
| 200 |
+
"loss": 0.4131,
|
| 201 |
+
"step": 120
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 12.923076923076923,
|
| 205 |
+
"eval_accuracy": 0.8602941176470589,
|
| 206 |
+
"eval_loss": 0.5028533339500427,
|
| 207 |
+
"eval_runtime": 29.2998,
|
| 208 |
+
"eval_samples_per_second": 4.642,
|
| 209 |
+
"eval_steps_per_second": 0.171,
|
| 210 |
+
"step": 126
|
| 211 |
},
|
| 212 |
{
|
| 213 |
"epoch": 13.333333333333334,
|
| 214 |
+
"grad_norm": 2.1402792930603027,
|
| 215 |
+
"learning_rate": 3.950617283950617e-05,
|
| 216 |
+
"loss": 0.4012,
|
| 217 |
+
"step": 130
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 218 |
},
|
| 219 |
{
|
| 220 |
+
"epoch": 13.948717948717949,
|
| 221 |
+
"eval_accuracy": 0.8382352941176471,
|
| 222 |
+
"eval_loss": 0.5566176176071167,
|
| 223 |
+
"eval_runtime": 31.1116,
|
| 224 |
+
"eval_samples_per_second": 4.371,
|
| 225 |
+
"eval_steps_per_second": 0.161,
|
| 226 |
+
"step": 136
|
| 227 |
},
|
| 228 |
{
|
| 229 |
+
"epoch": 14.35897435897436,
|
| 230 |
+
"grad_norm": 2.5639398097991943,
|
| 231 |
+
"learning_rate": 3.82716049382716e-05,
|
| 232 |
+
"loss": 0.3689,
|
| 233 |
+
"step": 140
|
| 234 |
},
|
| 235 |
{
|
| 236 |
+
"epoch": 14.974358974358974,
|
| 237 |
+
"eval_accuracy": 0.8382352941176471,
|
| 238 |
+
"eval_loss": 0.5533155798912048,
|
| 239 |
+
"eval_runtime": 29.2502,
|
| 240 |
+
"eval_samples_per_second": 4.65,
|
| 241 |
+
"eval_steps_per_second": 0.171,
|
| 242 |
+
"step": 146
|
| 243 |
},
|
| 244 |
{
|
| 245 |
+
"epoch": 15.384615384615385,
|
| 246 |
+
"grad_norm": 3.994143009185791,
|
| 247 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 248 |
+
"loss": 0.3533,
|
| 249 |
+
"step": 150
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 16.0,
|
| 253 |
"eval_accuracy": 0.8970588235294118,
|
| 254 |
+
"eval_loss": 0.423177033662796,
|
| 255 |
+
"eval_runtime": 29.8411,
|
| 256 |
+
"eval_samples_per_second": 4.557,
|
| 257 |
+
"eval_steps_per_second": 0.168,
|
| 258 |
+
"step": 156
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 259 |
},
|
| 260 |
{
|
| 261 |
+
"epoch": 16.41025641025641,
|
| 262 |
+
"grad_norm": 3.873570203781128,
|
| 263 |
+
"learning_rate": 3.580246913580247e-05,
|
| 264 |
+
"loss": 0.2954,
|
| 265 |
+
"step": 160
|
| 266 |
},
|
| 267 |
{
|
| 268 |
+
"epoch": 16.923076923076923,
|
| 269 |
+
"eval_accuracy": 0.8897058823529411,
|
| 270 |
+
"eval_loss": 0.4589211940765381,
|
| 271 |
+
"eval_runtime": 29.2531,
|
| 272 |
+
"eval_samples_per_second": 4.649,
|
| 273 |
+
"eval_steps_per_second": 0.171,
|
| 274 |
+
"step": 165
|
| 275 |
},
|
| 276 |
{
|
| 277 |
+
"epoch": 17.435897435897434,
|
| 278 |
+
"grad_norm": 2.3892805576324463,
|
| 279 |
+
"learning_rate": 3.45679012345679e-05,
|
| 280 |
+
"loss": 0.2907,
|
| 281 |
+
"step": 170
|
|
|
|
|
|
|
| 282 |
},
|
| 283 |
{
|
| 284 |
+
"epoch": 17.94871794871795,
|
| 285 |
+
"eval_accuracy": 0.8970588235294118,
|
| 286 |
+
"eval_loss": 0.4223209321498871,
|
| 287 |
+
"eval_runtime": 28.9917,
|
| 288 |
+
"eval_samples_per_second": 4.691,
|
| 289 |
+
"eval_steps_per_second": 0.172,
|
| 290 |
+
"step": 175
|
| 291 |
},
|
| 292 |
{
|
| 293 |
+
"epoch": 18.46153846153846,
|
| 294 |
+
"grad_norm": 0.7764536142349243,
|
| 295 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 296 |
+
"loss": 0.2804,
|
| 297 |
+
"step": 180
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 18.974358974358974,
|
| 301 |
+
"eval_accuracy": 0.8970588235294118,
|
| 302 |
+
"eval_loss": 0.4056432843208313,
|
| 303 |
+
"eval_runtime": 28.7309,
|
| 304 |
+
"eval_samples_per_second": 4.734,
|
| 305 |
+
"eval_steps_per_second": 0.174,
|
| 306 |
+
"step": 185
|
| 307 |
},
|
| 308 |
{
|
| 309 |
+
"epoch": 19.487179487179485,
|
| 310 |
+
"grad_norm": 1.780919075012207,
|
| 311 |
+
"learning_rate": 3.209876543209876e-05,
|
| 312 |
+
"loss": 0.2469,
|
| 313 |
+
"step": 190
|
| 314 |
},
|
| 315 |
{
|
| 316 |
+
"epoch": 20.0,
|
| 317 |
"eval_accuracy": 0.9117647058823529,
|
| 318 |
+
"eval_loss": 0.390433132648468,
|
| 319 |
+
"eval_runtime": 29.1361,
|
| 320 |
+
"eval_samples_per_second": 4.668,
|
| 321 |
+
"eval_steps_per_second": 0.172,
|
| 322 |
+
"step": 195
|
| 323 |
},
|
| 324 |
{
|
| 325 |
+
"epoch": 20.51282051282051,
|
| 326 |
+
"grad_norm": 3.051804304122925,
|
| 327 |
+
"learning_rate": 3.08641975308642e-05,
|
| 328 |
+
"loss": 0.2643,
|
| 329 |
+
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 330 |
},
|
| 331 |
{
|
| 332 |
+
"epoch": 20.923076923076923,
|
| 333 |
"eval_accuracy": 0.9044117647058824,
|
| 334 |
+
"eval_loss": 0.38655146956443787,
|
| 335 |
+
"eval_runtime": 28.9046,
|
| 336 |
+
"eval_samples_per_second": 4.705,
|
| 337 |
+
"eval_steps_per_second": 0.173,
|
| 338 |
+
"step": 204
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 339 |
},
|
| 340 |
{
|
| 341 |
+
"epoch": 21.53846153846154,
|
| 342 |
+
"grad_norm": 2.3158023357391357,
|
| 343 |
+
"learning_rate": 2.962962962962963e-05,
|
| 344 |
+
"loss": 0.2212,
|
| 345 |
+
"step": 210
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 21.94871794871795,
|
| 349 |
+
"eval_accuracy": 0.875,
|
| 350 |
+
"eval_loss": 0.4173263609409332,
|
| 351 |
+
"eval_runtime": 29.2277,
|
| 352 |
+
"eval_samples_per_second": 4.653,
|
| 353 |
+
"eval_steps_per_second": 0.171,
|
| 354 |
+
"step": 214
|
| 355 |
},
|
| 356 |
{
|
| 357 |
+
"epoch": 22.564102564102566,
|
| 358 |
+
"grad_norm": 3.5362863540649414,
|
| 359 |
+
"learning_rate": 2.839506172839506e-05,
|
| 360 |
+
"loss": 0.2476,
|
| 361 |
+
"step": 220
|
| 362 |
},
|
| 363 |
{
|
| 364 |
+
"epoch": 22.974358974358974,
|
| 365 |
+
"eval_accuracy": 0.8014705882352942,
|
| 366 |
+
"eval_loss": 0.6001182794570923,
|
| 367 |
+
"eval_runtime": 29.5734,
|
| 368 |
+
"eval_samples_per_second": 4.599,
|
| 369 |
+
"eval_steps_per_second": 0.169,
|
| 370 |
+
"step": 224
|
| 371 |
},
|
| 372 |
{
|
| 373 |
+
"epoch": 23.58974358974359,
|
| 374 |
+
"grad_norm": 2.7392866611480713,
|
| 375 |
+
"learning_rate": 2.7160493827160493e-05,
|
| 376 |
+
"loss": 0.2347,
|
| 377 |
+
"step": 230
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 24.0,
|
| 381 |
+
"eval_accuracy": 0.9044117647058824,
|
| 382 |
+
"eval_loss": 0.39004915952682495,
|
| 383 |
+
"eval_runtime": 28.5909,
|
| 384 |
+
"eval_samples_per_second": 4.757,
|
| 385 |
+
"eval_steps_per_second": 0.175,
|
| 386 |
+
"step": 234
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 24.615384615384617,
|
| 390 |
+
"grad_norm": 1.7161147594451904,
|
| 391 |
+
"learning_rate": 2.5925925925925925e-05,
|
| 392 |
+
"loss": 0.207,
|
| 393 |
+
"step": 240
|
| 394 |
},
|
| 395 |
{
|
| 396 |
+
"epoch": 24.923076923076923,
|
| 397 |
+
"eval_accuracy": 0.8897058823529411,
|
| 398 |
+
"eval_loss": 0.40333059430122375,
|
| 399 |
+
"eval_runtime": 29.3709,
|
| 400 |
+
"eval_samples_per_second": 4.63,
|
| 401 |
+
"eval_steps_per_second": 0.17,
|
| 402 |
+
"step": 243
|
| 403 |
},
|
| 404 |
{
|
| 405 |
+
"epoch": 25.641025641025642,
|
| 406 |
+
"grad_norm": 3.595083713531494,
|
| 407 |
+
"learning_rate": 2.4691358024691357e-05,
|
| 408 |
+
"loss": 0.1803,
|
| 409 |
+
"step": 250
|
|
|
|
|
|
|
| 410 |
},
|
| 411 |
{
|
| 412 |
+
"epoch": 25.94871794871795,
|
| 413 |
+
"eval_accuracy": 0.9264705882352942,
|
| 414 |
+
"eval_loss": 0.3509926497936249,
|
| 415 |
+
"eval_runtime": 28.9919,
|
| 416 |
+
"eval_samples_per_second": 4.691,
|
| 417 |
+
"eval_steps_per_second": 0.172,
|
| 418 |
+
"step": 253
|
| 419 |
},
|
| 420 |
{
|
| 421 |
+
"epoch": 26.666666666666668,
|
| 422 |
+
"grad_norm": 3.6539347171783447,
|
| 423 |
+
"learning_rate": 2.345679012345679e-05,
|
| 424 |
+
"loss": 0.1979,
|
| 425 |
+
"step": 260
|
| 426 |
},
|
| 427 |
{
|
| 428 |
+
"epoch": 26.974358974358974,
|
| 429 |
+
"eval_accuracy": 0.9191176470588235,
|
| 430 |
+
"eval_loss": 0.3723020553588867,
|
| 431 |
+
"eval_runtime": 28.5326,
|
| 432 |
+
"eval_samples_per_second": 4.766,
|
| 433 |
+
"eval_steps_per_second": 0.175,
|
| 434 |
+
"step": 263
|
| 435 |
},
|
| 436 |
{
|
| 437 |
+
"epoch": 27.692307692307693,
|
| 438 |
+
"grad_norm": 2.836503505706787,
|
| 439 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 440 |
+
"loss": 0.1821,
|
| 441 |
+
"step": 270
|
| 442 |
},
|
| 443 |
{
|
| 444 |
+
"epoch": 28.0,
|
| 445 |
"eval_accuracy": 0.8823529411764706,
|
| 446 |
+
"eval_loss": 0.4319731295108795,
|
| 447 |
+
"eval_runtime": 29.3401,
|
| 448 |
+
"eval_samples_per_second": 4.635,
|
| 449 |
+
"eval_steps_per_second": 0.17,
|
| 450 |
+
"step": 273
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 451 |
},
|
| 452 |
{
|
| 453 |
+
"epoch": 28.71794871794872,
|
| 454 |
+
"grad_norm": 3.2694666385650635,
|
| 455 |
+
"learning_rate": 2.0987654320987655e-05,
|
| 456 |
+
"loss": 0.1992,
|
| 457 |
+
"step": 280
|
| 458 |
},
|
| 459 |
{
|
| 460 |
+
"epoch": 28.923076923076923,
|
| 461 |
+
"eval_accuracy": 0.9117647058823529,
|
| 462 |
+
"eval_loss": 0.3557020127773285,
|
| 463 |
+
"eval_runtime": 29.5759,
|
| 464 |
+
"eval_samples_per_second": 4.598,
|
| 465 |
+
"eval_steps_per_second": 0.169,
|
| 466 |
+
"step": 282
|
| 467 |
},
|
| 468 |
{
|
| 469 |
+
"epoch": 29.743589743589745,
|
| 470 |
+
"grad_norm": 3.0022389888763428,
|
| 471 |
+
"learning_rate": 1.9753086419753087e-05,
|
| 472 |
+
"loss": 0.2154,
|
| 473 |
+
"step": 290
|
| 474 |
},
|
| 475 |
{
|
| 476 |
+
"epoch": 29.94871794871795,
|
| 477 |
+
"eval_accuracy": 0.9191176470588235,
|
| 478 |
+
"eval_loss": 0.3362026810646057,
|
| 479 |
+
"eval_runtime": 29.8144,
|
| 480 |
+
"eval_samples_per_second": 4.562,
|
| 481 |
+
"eval_steps_per_second": 0.168,
|
| 482 |
+
"step": 292
|
| 483 |
},
|
| 484 |
{
|
| 485 |
+
"epoch": 30.76923076923077,
|
| 486 |
+
"grad_norm": 1.9505140781402588,
|
| 487 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 488 |
+
"loss": 0.1801,
|
| 489 |
+
"step": 300
|
| 490 |
},
|
| 491 |
{
|
| 492 |
+
"epoch": 30.974358974358974,
|
| 493 |
+
"eval_accuracy": 0.875,
|
| 494 |
+
"eval_loss": 0.4357524812221527,
|
| 495 |
+
"eval_runtime": 28.9816,
|
| 496 |
+
"eval_samples_per_second": 4.693,
|
| 497 |
+
"eval_steps_per_second": 0.173,
|
| 498 |
+
"step": 302
|
| 499 |
},
|
| 500 |
{
|
| 501 |
+
"epoch": 31.794871794871796,
|
| 502 |
+
"grad_norm": 2.702116012573242,
|
| 503 |
+
"learning_rate": 1.728395061728395e-05,
|
| 504 |
+
"loss": 0.1794,
|
| 505 |
+
"step": 310
|
| 506 |
},
|
| 507 |
{
|
| 508 |
+
"epoch": 32.0,
|
| 509 |
"eval_accuracy": 0.9191176470588235,
|
| 510 |
+
"eval_loss": 0.3500049114227295,
|
| 511 |
+
"eval_runtime": 29.5602,
|
| 512 |
+
"eval_samples_per_second": 4.601,
|
| 513 |
+
"eval_steps_per_second": 0.169,
|
| 514 |
+
"step": 312
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 515 |
},
|
| 516 |
{
|
| 517 |
+
"epoch": 32.82051282051282,
|
| 518 |
+
"grad_norm": 1.1489578485488892,
|
| 519 |
+
"learning_rate": 1.604938271604938e-05,
|
| 520 |
+
"loss": 0.1566,
|
| 521 |
+
"step": 320
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 522 |
},
|
| 523 |
{
|
| 524 |
+
"epoch": 32.92307692307692,
|
| 525 |
"eval_accuracy": 0.9264705882352942,
|
| 526 |
+
"eval_loss": 0.30455005168914795,
|
| 527 |
+
"eval_runtime": 29.3612,
|
| 528 |
+
"eval_samples_per_second": 4.632,
|
| 529 |
+
"eval_steps_per_second": 0.17,
|
| 530 |
+
"step": 321
|
| 531 |
},
|
| 532 |
{
|
| 533 |
+
"epoch": 33.84615384615385,
|
| 534 |
+
"grad_norm": 1.6699856519699097,
|
| 535 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 536 |
+
"loss": 0.1432,
|
| 537 |
+
"step": 330
|
| 538 |
},
|
| 539 |
{
|
| 540 |
+
"epoch": 33.94871794871795,
|
| 541 |
+
"eval_accuracy": 0.9264705882352942,
|
| 542 |
+
"eval_loss": 0.3239189088344574,
|
| 543 |
+
"eval_runtime": 29.5627,
|
| 544 |
+
"eval_samples_per_second": 4.6,
|
| 545 |
+
"eval_steps_per_second": 0.169,
|
| 546 |
+
"step": 331
|
| 547 |
},
|
| 548 |
{
|
| 549 |
+
"epoch": 34.87179487179487,
|
| 550 |
+
"grad_norm": 1.4475438594818115,
|
| 551 |
+
"learning_rate": 1.3580246913580247e-05,
|
| 552 |
+
"loss": 0.145,
|
| 553 |
+
"step": 340
|
| 554 |
},
|
| 555 |
{
|
| 556 |
+
"epoch": 34.97435897435897,
|
| 557 |
+
"eval_accuracy": 0.9338235294117647,
|
| 558 |
+
"eval_loss": 0.33108076453208923,
|
| 559 |
+
"eval_runtime": 29.2908,
|
| 560 |
+
"eval_samples_per_second": 4.643,
|
| 561 |
+
"eval_steps_per_second": 0.171,
|
| 562 |
+
"step": 341
|
| 563 |
},
|
| 564 |
{
|
| 565 |
+
"epoch": 35.8974358974359,
|
| 566 |
+
"grad_norm": 1.5676722526550293,
|
| 567 |
+
"learning_rate": 1.2345679012345678e-05,
|
| 568 |
+
"loss": 0.1578,
|
| 569 |
+
"step": 350
|
|
|
|
|
|
|
| 570 |
},
|
| 571 |
{
|
| 572 |
+
"epoch": 36.0,
|
| 573 |
+
"eval_accuracy": 0.9338235294117647,
|
| 574 |
+
"eval_loss": 0.30287855863571167,
|
| 575 |
+
"eval_runtime": 29.0519,
|
| 576 |
+
"eval_samples_per_second": 4.681,
|
| 577 |
+
"eval_steps_per_second": 0.172,
|
| 578 |
+
"step": 351
|
| 579 |
},
|
| 580 |
{
|
| 581 |
+
"epoch": 36.92307692307692,
|
| 582 |
+
"grad_norm": 1.0356682538986206,
|
| 583 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 584 |
+
"loss": 0.1511,
|
| 585 |
+
"step": 360
|
| 586 |
},
|
| 587 |
{
|
| 588 |
+
"epoch": 36.92307692307692,
|
| 589 |
+
"eval_accuracy": 0.9338235294117647,
|
| 590 |
+
"eval_loss": 0.30101922154426575,
|
| 591 |
+
"eval_runtime": 31.6784,
|
| 592 |
+
"eval_samples_per_second": 4.293,
|
| 593 |
+
"eval_steps_per_second": 0.158,
|
| 594 |
+
"step": 360
|
| 595 |
},
|
| 596 |
{
|
| 597 |
+
"epoch": 37.94871794871795,
|
| 598 |
+
"grad_norm": 3.916630268096924,
|
| 599 |
+
"learning_rate": 9.876543209876543e-06,
|
| 600 |
+
"loss": 0.139,
|
| 601 |
+
"step": 370
|
| 602 |
},
|
| 603 |
{
|
| 604 |
+
"epoch": 37.94871794871795,
|
| 605 |
"eval_accuracy": 0.9264705882352942,
|
| 606 |
+
"eval_loss": 0.29820650815963745,
|
| 607 |
+
"eval_runtime": 29.7228,
|
| 608 |
+
"eval_samples_per_second": 4.576,
|
| 609 |
+
"eval_steps_per_second": 0.168,
|
| 610 |
+
"step": 370
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 611 |
},
|
| 612 |
{
|
| 613 |
+
"epoch": 38.97435897435897,
|
| 614 |
+
"grad_norm": 0.7166056632995605,
|
| 615 |
+
"learning_rate": 8.641975308641975e-06,
|
| 616 |
+
"loss": 0.1294,
|
| 617 |
+
"step": 380
|
| 618 |
},
|
| 619 |
{
|
| 620 |
+
"epoch": 38.97435897435897,
|
| 621 |
+
"eval_accuracy": 0.9191176470588235,
|
| 622 |
+
"eval_loss": 0.32610374689102173,
|
| 623 |
+
"eval_runtime": 28.752,
|
| 624 |
+
"eval_samples_per_second": 4.73,
|
| 625 |
+
"eval_steps_per_second": 0.174,
|
| 626 |
+
"step": 380
|
| 627 |
},
|
| 628 |
{
|
| 629 |
+
"epoch": 40.0,
|
| 630 |
+
"grad_norm": 2.002183675765991,
|
| 631 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 632 |
+
"loss": 0.1263,
|
| 633 |
+
"step": 390
|
| 634 |
},
|
| 635 |
{
|
| 636 |
+
"epoch": 40.0,
|
| 637 |
+
"eval_accuracy": 0.9338235294117647,
|
| 638 |
+
"eval_loss": 0.2932202219963074,
|
| 639 |
+
"eval_runtime": 29.503,
|
| 640 |
+
"eval_samples_per_second": 4.61,
|
| 641 |
+
"eval_steps_per_second": 0.169,
|
| 642 |
+
"step": 390
|
| 643 |
},
|
| 644 |
{
|
| 645 |
+
"epoch": 40.92307692307692,
|
| 646 |
+
"eval_accuracy": 0.9338235294117647,
|
| 647 |
+
"eval_loss": 0.2943984270095825,
|
| 648 |
+
"eval_runtime": 29.1815,
|
| 649 |
+
"eval_samples_per_second": 4.66,
|
| 650 |
+
"eval_steps_per_second": 0.171,
|
| 651 |
+
"step": 399
|
| 652 |
},
|
| 653 |
{
|
| 654 |
+
"epoch": 41.02564102564103,
|
| 655 |
+
"grad_norm": 0.8789101839065552,
|
| 656 |
+
"learning_rate": 6.172839506172839e-06,
|
| 657 |
+
"loss": 0.1216,
|
| 658 |
+
"step": 400
|
| 659 |
},
|
| 660 |
{
|
| 661 |
+
"epoch": 41.94871794871795,
|
| 662 |
+
"eval_accuracy": 0.9338235294117647,
|
| 663 |
+
"eval_loss": 0.2867204248905182,
|
| 664 |
+
"eval_runtime": 28.6349,
|
| 665 |
+
"eval_samples_per_second": 4.749,
|
| 666 |
+
"eval_steps_per_second": 0.175,
|
| 667 |
+
"step": 409
|
| 668 |
},
|
| 669 |
{
|
| 670 |
+
"epoch": 42.05128205128205,
|
| 671 |
+
"grad_norm": 2.1831233501434326,
|
| 672 |
+
"learning_rate": 4.938271604938272e-06,
|
| 673 |
+
"loss": 0.1199,
|
| 674 |
+
"step": 410
|
| 675 |
},
|
| 676 |
{
|
| 677 |
+
"epoch": 42.97435897435897,
|
| 678 |
+
"eval_accuracy": 0.9338235294117647,
|
| 679 |
+
"eval_loss": 0.28869205713272095,
|
| 680 |
+
"eval_runtime": 28.6692,
|
| 681 |
+
"eval_samples_per_second": 4.744,
|
| 682 |
+
"eval_steps_per_second": 0.174,
|
| 683 |
+
"step": 419
|
| 684 |
},
|
| 685 |
{
|
| 686 |
+
"epoch": 43.07692307692308,
|
| 687 |
+
"grad_norm": 3.3434555530548096,
|
| 688 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 689 |
+
"loss": 0.128,
|
| 690 |
+
"step": 420
|
| 691 |
},
|
| 692 |
{
|
| 693 |
+
"epoch": 44.0,
|
| 694 |
+
"eval_accuracy": 0.9338235294117647,
|
| 695 |
+
"eval_loss": 0.28250837326049805,
|
| 696 |
+
"eval_runtime": 28.5349,
|
| 697 |
+
"eval_samples_per_second": 4.766,
|
| 698 |
+
"eval_steps_per_second": 0.175,
|
| 699 |
+
"step": 429
|
| 700 |
},
|
| 701 |
{
|
| 702 |
+
"epoch": 44.1025641025641,
|
| 703 |
+
"grad_norm": 1.314479112625122,
|
| 704 |
+
"learning_rate": 2.469135802469136e-06,
|
| 705 |
+
"loss": 0.1115,
|
| 706 |
+
"step": 430
|
| 707 |
},
|
| 708 |
{
|
| 709 |
+
"epoch": 44.92307692307692,
|
| 710 |
+
"eval_accuracy": 0.9338235294117647,
|
| 711 |
+
"eval_loss": 0.2879542410373688,
|
| 712 |
+
"eval_runtime": 28.8609,
|
| 713 |
+
"eval_samples_per_second": 4.712,
|
| 714 |
+
"eval_steps_per_second": 0.173,
|
| 715 |
+
"step": 438
|
| 716 |
},
|
| 717 |
{
|
| 718 |
+
"epoch": 45.12820512820513,
|
| 719 |
+
"grad_norm": 1.4549723863601685,
|
| 720 |
+
"learning_rate": 1.234567901234568e-06,
|
| 721 |
+
"loss": 0.1179,
|
| 722 |
+
"step": 440
|
| 723 |
},
|
| 724 |
{
|
| 725 |
+
"epoch": 45.94871794871795,
|
| 726 |
+
"eval_accuracy": 0.9338235294117647,
|
| 727 |
+
"eval_loss": 0.2870730757713318,
|
| 728 |
+
"eval_runtime": 29.6828,
|
| 729 |
+
"eval_samples_per_second": 4.582,
|
| 730 |
+
"eval_steps_per_second": 0.168,
|
| 731 |
+
"step": 448
|
| 732 |
},
|
| 733 |
{
|
| 734 |
+
"epoch": 46.15384615384615,
|
| 735 |
+
"grad_norm": 1.243077039718628,
|
| 736 |
+
"learning_rate": 0.0,
|
| 737 |
+
"loss": 0.12,
|
| 738 |
+
"step": 450
|
| 739 |
},
|
| 740 |
{
|
| 741 |
+
"epoch": 46.15384615384615,
|
| 742 |
+
"eval_accuracy": 0.9338235294117647,
|
| 743 |
+
"eval_loss": 0.28695058822631836,
|
| 744 |
+
"eval_runtime": 31.6423,
|
| 745 |
+
"eval_samples_per_second": 4.298,
|
| 746 |
+
"eval_steps_per_second": 0.158,
|
| 747 |
+
"step": 450
|
| 748 |
},
|
| 749 |
{
|
| 750 |
+
"epoch": 46.15384615384615,
|
| 751 |
+
"step": 450,
|
| 752 |
+
"total_flos": 4.3781443993328026e+18,
|
| 753 |
+
"train_loss": 0.41361336602105037,
|
| 754 |
+
"train_runtime": 14457.8356,
|
| 755 |
+
"train_samples_per_second": 4.233,
|
| 756 |
+
"train_steps_per_second": 0.031
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 757 |
}
|
| 758 |
],
|
| 759 |
"logging_steps": 10,
|
| 760 |
+
"max_steps": 450,
|
| 761 |
"num_input_tokens_seen": 0,
|
| 762 |
+
"num_train_epochs": 50,
|
| 763 |
"save_steps": 500,
|
| 764 |
"stateful_callbacks": {
|
| 765 |
"TrainerControl": {
|
|
|
|
| 773 |
"attributes": {}
|
| 774 |
}
|
| 775 |
},
|
| 776 |
+
"total_flos": 4.3781443993328026e+18,
|
| 777 |
+
"train_batch_size": 32,
|
| 778 |
"trial_name": null,
|
| 779 |
"trial_params": null
|
| 780 |
}
|