dhritic99/vit-base-brain-alzheimer-detection
Browse files- README.md +2 -2
- all_results.json +11 -11
- eval_results.json +6 -6
- train_results.json +6 -6
- trainer_state.json +322 -246
README.md
CHANGED
|
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Accuracy: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.2428
|
| 21 |
+
- Accuracy: 0.9508
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 13.
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 22.
|
| 8 |
-
"total_flos": 4.
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second":
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 13.28125,
|
| 3 |
+
"eval_accuracy": 0.95078125,
|
| 4 |
+
"eval_loss": 0.24280217289924622,
|
| 5 |
+
"eval_runtime": 6.969,
|
| 6 |
+
"eval_samples_per_second": 183.67,
|
| 7 |
+
"eval_steps_per_second": 22.959,
|
| 8 |
+
"total_flos": 4.2156398005714944e+18,
|
| 9 |
+
"train_loss": 0.23371873112285838,
|
| 10 |
+
"train_runtime": 1043.0698,
|
| 11 |
+
"train_samples_per_second": 78.537,
|
| 12 |
+
"train_steps_per_second": 4.909
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 13.
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 22.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 13.28125,
|
| 3 |
+
"eval_accuracy": 0.95078125,
|
| 4 |
+
"eval_loss": 0.24280217289924622,
|
| 5 |
+
"eval_runtime": 6.969,
|
| 6 |
+
"eval_samples_per_second": 183.67,
|
| 7 |
+
"eval_steps_per_second": 22.959
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 13.
|
| 3 |
-
"total_flos": 4.
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 13.28125,
|
| 3 |
+
"total_flos": 4.2156398005714944e+18,
|
| 4 |
+
"train_loss": 0.23371873112285838,
|
| 5 |
+
"train_runtime": 1043.0698,
|
| 6 |
+
"train_samples_per_second": 78.537,
|
| 7 |
+
"train_steps_per_second": 4.909
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,581 +1,657 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "./vit-base-brain-alzheimer-detection/checkpoint-
|
| 4 |
-
"epoch": 13.
|
| 5 |
-
"eval_steps":
|
| 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": 0.1953125,
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 2.0000000000000003e-06,
|
| 15 |
-
"loss":
|
| 16 |
"step": 50
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.390625,
|
| 20 |
-
"grad_norm":
|
| 21 |
"learning_rate": 4.000000000000001e-06,
|
| 22 |
-
"loss":
|
| 23 |
"step": 100
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.5859375,
|
| 27 |
-
"grad_norm":
|
| 28 |
"learning_rate": 6e-06,
|
| 29 |
-
"loss":
|
| 30 |
"step": 150
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.78125,
|
| 34 |
-
"grad_norm":
|
| 35 |
"learning_rate": 8.000000000000001e-06,
|
| 36 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
"step": 200
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.9765625,
|
| 41 |
-
"grad_norm":
|
| 42 |
"learning_rate": 1e-05,
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 250
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 1.171875,
|
| 48 |
-
"grad_norm":
|
| 49 |
"learning_rate": 1.2e-05,
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 300
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 1.3671875,
|
| 55 |
-
"grad_norm":
|
| 56 |
"learning_rate": 1.4e-05,
|
| 57 |
-
"loss": 0.
|
| 58 |
"step": 350
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 1.5625,
|
| 62 |
-
"grad_norm":
|
| 63 |
"learning_rate": 1.6000000000000003e-05,
|
| 64 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 65 |
"step": 400
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 1.7578125,
|
| 69 |
-
"grad_norm":
|
| 70 |
"learning_rate": 1.8e-05,
|
| 71 |
-
"loss": 0.
|
| 72 |
"step": 450
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 1.953125,
|
| 76 |
-
"grad_norm":
|
| 77 |
-
"learning_rate":
|
| 78 |
-
"loss": 0.
|
| 79 |
-
"step": 500
|
| 80 |
-
},
|
| 81 |
-
{
|
| 82 |
-
"epoch": 1.953125,
|
| 83 |
-
"eval_accuracy": 0.8955078125,
|
| 84 |
-
"eval_loss": 0.3758355975151062,
|
| 85 |
-
"eval_runtime": 7.1938,
|
| 86 |
-
"eval_samples_per_second": 142.346,
|
| 87 |
-
"eval_steps_per_second": 17.793,
|
| 88 |
"step": 500
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"epoch": 2.1484375,
|
| 92 |
-
"grad_norm":
|
| 93 |
-
"learning_rate": 1.
|
| 94 |
-
"loss": 0.
|
| 95 |
"step": 550
|
| 96 |
},
|
| 97 |
{
|
| 98 |
"epoch": 2.34375,
|
| 99 |
-
"grad_norm":
|
| 100 |
-
"learning_rate": 1.
|
| 101 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 102 |
"step": 600
|
| 103 |
},
|
| 104 |
{
|
| 105 |
"epoch": 2.5390625,
|
| 106 |
-
"grad_norm":
|
| 107 |
-
"learning_rate": 1.
|
| 108 |
-
"loss": 0.
|
| 109 |
"step": 650
|
| 110 |
},
|
| 111 |
{
|
| 112 |
"epoch": 2.734375,
|
| 113 |
-
"grad_norm": 10.
|
| 114 |
-
"learning_rate": 1.
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 700
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 2.9296875,
|
| 120 |
-
"grad_norm":
|
| 121 |
-
"learning_rate": 1.
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 750
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 3.125,
|
| 127 |
-
"grad_norm":
|
| 128 |
-
"learning_rate": 1.
|
| 129 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 130 |
"step": 800
|
| 131 |
},
|
| 132 |
{
|
| 133 |
"epoch": 3.3203125,
|
| 134 |
-
"grad_norm":
|
| 135 |
-
"learning_rate": 1.
|
| 136 |
-
"loss": 0.
|
| 137 |
"step": 850
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 3.515625,
|
| 141 |
-
"grad_norm":
|
| 142 |
-
"learning_rate": 1.
|
| 143 |
-
"loss": 0.
|
| 144 |
"step": 900
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"epoch": 3.7109375,
|
| 148 |
-
"grad_norm":
|
| 149 |
-
"learning_rate": 1.
|
| 150 |
-
"loss": 0.
|
| 151 |
"step": 950
|
| 152 |
},
|
| 153 |
{
|
| 154 |
"epoch": 3.90625,
|
| 155 |
-
"grad_norm":
|
| 156 |
-
"learning_rate": 1.
|
| 157 |
-
"loss": 0.
|
| 158 |
"step": 1000
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 3.90625,
|
| 162 |
-
"eval_accuracy": 0.
|
| 163 |
-
"eval_loss": 0.
|
| 164 |
-
"eval_runtime": 5.
|
| 165 |
-
"eval_samples_per_second":
|
| 166 |
-
"eval_steps_per_second":
|
| 167 |
"step": 1000
|
| 168 |
},
|
| 169 |
{
|
| 170 |
"epoch": 4.1015625,
|
| 171 |
-
"grad_norm":
|
| 172 |
-
"learning_rate": 1.
|
| 173 |
-
"loss": 0.
|
| 174 |
"step": 1050
|
| 175 |
},
|
| 176 |
{
|
| 177 |
"epoch": 4.296875,
|
| 178 |
-
"grad_norm":
|
| 179 |
-
"learning_rate": 1.
|
| 180 |
-
"loss": 0.
|
| 181 |
"step": 1100
|
| 182 |
},
|
| 183 |
{
|
| 184 |
"epoch": 4.4921875,
|
| 185 |
-
"grad_norm":
|
| 186 |
-
"learning_rate": 1.
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 1150
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 4.6875,
|
| 192 |
-
"grad_norm": 0.
|
| 193 |
-
"learning_rate": 1.
|
| 194 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 195 |
"step": 1200
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 4.8828125,
|
| 199 |
-
"grad_norm":
|
| 200 |
-
"learning_rate": 1.
|
| 201 |
-
"loss": 0.
|
| 202 |
"step": 1250
|
| 203 |
},
|
| 204 |
{
|
| 205 |
"epoch": 5.078125,
|
| 206 |
-
"grad_norm":
|
| 207 |
-
"learning_rate": 1.
|
| 208 |
-
"loss": 0.
|
| 209 |
"step": 1300
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 5.2734375,
|
| 213 |
-
"grad_norm": 0.
|
| 214 |
-
"learning_rate": 1.
|
| 215 |
-
"loss": 0.
|
| 216 |
"step": 1350
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"epoch": 5.46875,
|
| 220 |
-
"grad_norm":
|
| 221 |
-
"learning_rate": 1.
|
| 222 |
-
"loss": 0.
|
| 223 |
"step": 1400
|
| 224 |
},
|
| 225 |
{
|
| 226 |
-
"epoch": 5.
|
| 227 |
-
"
|
| 228 |
-
"
|
| 229 |
-
"
|
| 230 |
-
"
|
|
|
|
|
|
|
| 231 |
},
|
| 232 |
{
|
| 233 |
-
"epoch": 5.
|
| 234 |
-
"grad_norm": 0.
|
| 235 |
-
"learning_rate": 1.
|
| 236 |
-
"loss": 0.
|
| 237 |
-
"step":
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 5.859375,
|
| 241 |
-
"
|
| 242 |
-
"
|
| 243 |
-
"
|
| 244 |
-
"eval_samples_per_second": 170.813,
|
| 245 |
-
"eval_steps_per_second": 21.352,
|
| 246 |
"step": 1500
|
| 247 |
},
|
| 248 |
{
|
| 249 |
"epoch": 6.0546875,
|
| 250 |
-
"grad_norm":
|
| 251 |
-
"learning_rate": 1.
|
| 252 |
-
"loss": 0.
|
| 253 |
"step": 1550
|
| 254 |
},
|
| 255 |
{
|
| 256 |
"epoch": 6.25,
|
| 257 |
-
"grad_norm":
|
| 258 |
-
"learning_rate": 1.
|
| 259 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 260 |
"step": 1600
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 6.4453125,
|
| 264 |
-
"grad_norm": 0.
|
| 265 |
-
"learning_rate": 1.
|
| 266 |
-
"loss": 0.
|
| 267 |
"step": 1650
|
| 268 |
},
|
| 269 |
{
|
| 270 |
"epoch": 6.640625,
|
| 271 |
-
"grad_norm": 0.
|
| 272 |
-
"learning_rate": 1.
|
| 273 |
-
"loss": 0.
|
| 274 |
"step": 1700
|
| 275 |
},
|
| 276 |
{
|
| 277 |
"epoch": 6.8359375,
|
| 278 |
-
"grad_norm":
|
| 279 |
-
"learning_rate": 1.
|
| 280 |
-
"loss": 0.
|
| 281 |
"step": 1750
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 7.03125,
|
| 285 |
-
"grad_norm": 0.
|
| 286 |
-
"learning_rate": 1.
|
| 287 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 288 |
"step": 1800
|
| 289 |
},
|
| 290 |
{
|
| 291 |
"epoch": 7.2265625,
|
| 292 |
-
"grad_norm": 0.
|
| 293 |
-
"learning_rate": 1.
|
| 294 |
-
"loss": 0.
|
| 295 |
"step": 1850
|
| 296 |
},
|
| 297 |
{
|
| 298 |
"epoch": 7.421875,
|
| 299 |
-
"grad_norm": 0.
|
| 300 |
-
"learning_rate": 1.
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 1900
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 7.6171875,
|
| 306 |
-
"grad_norm": 0.
|
| 307 |
-
"learning_rate": 1.
|
| 308 |
-
"loss": 0.
|
| 309 |
"step": 1950
|
| 310 |
},
|
| 311 |
{
|
| 312 |
"epoch": 7.8125,
|
| 313 |
-
"grad_norm": 0.
|
| 314 |
-
"learning_rate": 1.
|
| 315 |
-
"loss": 0.
|
| 316 |
"step": 2000
|
| 317 |
},
|
| 318 |
{
|
| 319 |
"epoch": 7.8125,
|
| 320 |
-
"eval_accuracy": 0.
|
| 321 |
-
"eval_loss": 0.
|
| 322 |
-
"eval_runtime":
|
| 323 |
-
"eval_samples_per_second":
|
| 324 |
-
"eval_steps_per_second":
|
| 325 |
"step": 2000
|
| 326 |
},
|
| 327 |
{
|
| 328 |
"epoch": 8.0078125,
|
| 329 |
-
"grad_norm": 0.
|
| 330 |
-
"learning_rate": 1.
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 2050
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 8.203125,
|
| 336 |
-
"grad_norm": 0.
|
| 337 |
-
"learning_rate": 1.
|
| 338 |
-
"loss": 0.
|
| 339 |
"step": 2100
|
| 340 |
},
|
| 341 |
{
|
| 342 |
"epoch": 8.3984375,
|
| 343 |
-
"grad_norm": 0.
|
| 344 |
"learning_rate": 1.2861471861471862e-05,
|
| 345 |
-
"loss": 0.
|
| 346 |
"step": 2150
|
| 347 |
},
|
| 348 |
{
|
| 349 |
"epoch": 8.59375,
|
| 350 |
-
"grad_norm": 0.
|
| 351 |
"learning_rate": 1.2645021645021646e-05,
|
| 352 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 353 |
"step": 2200
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 8.7890625,
|
| 357 |
-
"grad_norm": 0.
|
| 358 |
"learning_rate": 1.242857142857143e-05,
|
| 359 |
-
"loss": 0.
|
| 360 |
"step": 2250
|
| 361 |
},
|
| 362 |
{
|
| 363 |
"epoch": 8.984375,
|
| 364 |
-
"grad_norm": 0.
|
| 365 |
"learning_rate": 1.2212121212121213e-05,
|
| 366 |
-
"loss": 0.
|
| 367 |
"step": 2300
|
| 368 |
},
|
| 369 |
{
|
| 370 |
"epoch": 9.1796875,
|
| 371 |
-
"grad_norm": 0.
|
| 372 |
"learning_rate": 1.1995670995670996e-05,
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 2350
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 9.375,
|
| 378 |
-
"grad_norm": 0.
|
| 379 |
"learning_rate": 1.1779220779220779e-05,
|
| 380 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 381 |
"step": 2400
|
| 382 |
},
|
| 383 |
{
|
| 384 |
"epoch": 9.5703125,
|
| 385 |
-
"grad_norm": 0.
|
| 386 |
"learning_rate": 1.1562770562770564e-05,
|
| 387 |
-
"loss": 0.
|
| 388 |
"step": 2450
|
| 389 |
},
|
| 390 |
{
|
| 391 |
"epoch": 9.765625,
|
| 392 |
-
"grad_norm": 0.
|
| 393 |
"learning_rate": 1.1346320346320347e-05,
|
| 394 |
-
"loss": 0.
|
| 395 |
-
"step": 2500
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 9.765625,
|
| 399 |
-
"eval_accuracy": 0.9697265625,
|
| 400 |
-
"eval_loss": 0.20150402188301086,
|
| 401 |
-
"eval_runtime": 10.8662,
|
| 402 |
-
"eval_samples_per_second": 94.237,
|
| 403 |
-
"eval_steps_per_second": 11.78,
|
| 404 |
"step": 2500
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 9.9609375,
|
| 408 |
-
"grad_norm": 0.
|
| 409 |
"learning_rate": 1.112987012987013e-05,
|
| 410 |
-
"loss": 0.
|
| 411 |
"step": 2550
|
| 412 |
},
|
| 413 |
{
|
| 414 |
"epoch": 10.15625,
|
| 415 |
-
"grad_norm": 0.
|
| 416 |
"learning_rate": 1.0913419913419913e-05,
|
| 417 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 418 |
"step": 2600
|
| 419 |
},
|
| 420 |
{
|
| 421 |
"epoch": 10.3515625,
|
| 422 |
-
"grad_norm": 0.
|
| 423 |
"learning_rate": 1.0696969696969696e-05,
|
| 424 |
-
"loss": 0.
|
| 425 |
"step": 2650
|
| 426 |
},
|
| 427 |
{
|
| 428 |
"epoch": 10.546875,
|
| 429 |
-
"grad_norm": 0.
|
| 430 |
"learning_rate": 1.0480519480519481e-05,
|
| 431 |
-
"loss": 0.
|
| 432 |
"step": 2700
|
| 433 |
},
|
| 434 |
{
|
| 435 |
"epoch": 10.7421875,
|
| 436 |
-
"grad_norm": 0.
|
| 437 |
"learning_rate": 1.0264069264069264e-05,
|
| 438 |
-
"loss": 0.
|
| 439 |
"step": 2750
|
| 440 |
},
|
| 441 |
{
|
| 442 |
"epoch": 10.9375,
|
| 443 |
-
"grad_norm": 0.
|
| 444 |
"learning_rate": 1.0047619047619048e-05,
|
| 445 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 446 |
"step": 2800
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 11.1328125,
|
| 450 |
-
"grad_norm": 0.
|
| 451 |
"learning_rate": 9.831168831168832e-06,
|
| 452 |
-
"loss": 0.
|
| 453 |
"step": 2850
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 11.328125,
|
| 457 |
-
"grad_norm": 0.
|
| 458 |
"learning_rate": 9.614718614718616e-06,
|
| 459 |
-
"loss": 0.
|
| 460 |
"step": 2900
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 11.5234375,
|
| 464 |
-
"grad_norm": 0.
|
| 465 |
"learning_rate": 9.398268398268399e-06,
|
| 466 |
-
"loss": 0.
|
| 467 |
"step": 2950
|
| 468 |
},
|
| 469 |
{
|
| 470 |
"epoch": 11.71875,
|
| 471 |
-
"grad_norm": 0.
|
| 472 |
"learning_rate": 9.181818181818184e-06,
|
| 473 |
-
"loss": 0.
|
| 474 |
"step": 3000
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 11.71875,
|
| 478 |
-
"eval_accuracy": 0.
|
| 479 |
-
"eval_loss": 0.
|
| 480 |
-
"eval_runtime":
|
| 481 |
-
"eval_samples_per_second":
|
| 482 |
-
"eval_steps_per_second":
|
| 483 |
"step": 3000
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 11.9140625,
|
| 487 |
-
"grad_norm": 0.
|
| 488 |
"learning_rate": 8.965367965367967e-06,
|
| 489 |
-
"loss": 0.
|
| 490 |
"step": 3050
|
| 491 |
},
|
| 492 |
{
|
| 493 |
"epoch": 12.109375,
|
| 494 |
-
"grad_norm": 0.
|
| 495 |
"learning_rate": 8.74891774891775e-06,
|
| 496 |
-
"loss": 0.
|
| 497 |
"step": 3100
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 12.3046875,
|
| 501 |
-
"grad_norm": 0.
|
| 502 |
"learning_rate": 8.532467532467533e-06,
|
| 503 |
-
"loss": 0.
|
| 504 |
"step": 3150
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"epoch": 12.5,
|
| 508 |
-
"grad_norm": 0.
|
| 509 |
"learning_rate": 8.316017316017316e-06,
|
| 510 |
-
"loss": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 511 |
"step": 3200
|
| 512 |
},
|
| 513 |
{
|
| 514 |
"epoch": 12.6953125,
|
| 515 |
-
"grad_norm": 0.
|
| 516 |
"learning_rate": 8.099567099567101e-06,
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 3250
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 12.890625,
|
| 522 |
-
"grad_norm": 0.
|
| 523 |
"learning_rate": 7.883116883116884e-06,
|
| 524 |
-
"loss": 0.
|
| 525 |
"step": 3300
|
| 526 |
},
|
| 527 |
{
|
| 528 |
"epoch": 13.0859375,
|
| 529 |
-
"grad_norm": 0.
|
| 530 |
"learning_rate": 7.666666666666667e-06,
|
| 531 |
-
"loss": 0.
|
| 532 |
"step": 3350
|
| 533 |
},
|
| 534 |
{
|
| 535 |
"epoch": 13.28125,
|
| 536 |
-
"grad_norm": 0.
|
| 537 |
"learning_rate": 7.450216450216451e-06,
|
| 538 |
-
"loss": 0.
|
| 539 |
"step": 3400
|
| 540 |
},
|
| 541 |
{
|
| 542 |
-
"epoch": 13.
|
| 543 |
-
"
|
| 544 |
-
"
|
| 545 |
-
"
|
| 546 |
-
"
|
| 547 |
-
|
| 548 |
-
|
| 549 |
-
"epoch": 13.671875,
|
| 550 |
-
"grad_norm": 0.0021076672710478306,
|
| 551 |
-
"learning_rate": 7.017316017316018e-06,
|
| 552 |
-
"loss": 0.0004,
|
| 553 |
-
"step": 3500
|
| 554 |
-
},
|
| 555 |
-
{
|
| 556 |
-
"epoch": 13.671875,
|
| 557 |
-
"eval_accuracy": 0.96875,
|
| 558 |
-
"eval_loss": 0.21559299528598785,
|
| 559 |
-
"eval_runtime": 6.0759,
|
| 560 |
-
"eval_samples_per_second": 168.535,
|
| 561 |
-
"eval_steps_per_second": 21.067,
|
| 562 |
-
"step": 3500
|
| 563 |
},
|
| 564 |
{
|
| 565 |
-
"epoch": 13.
|
| 566 |
-
"step":
|
| 567 |
-
"total_flos": 4.
|
| 568 |
-
"train_loss": 0.
|
| 569 |
-
"train_runtime":
|
| 570 |
-
"train_samples_per_second":
|
| 571 |
-
"train_steps_per_second":
|
| 572 |
}
|
| 573 |
],
|
| 574 |
"logging_steps": 50,
|
| 575 |
"max_steps": 5120,
|
| 576 |
"num_input_tokens_seen": 0,
|
| 577 |
"num_train_epochs": 20,
|
| 578 |
-
"save_steps":
|
| 579 |
"stateful_callbacks": {
|
| 580 |
"EarlyStoppingCallback": {
|
| 581 |
"args": {
|
|
@@ -597,7 +673,7 @@
|
|
| 597 |
"attributes": {}
|
| 598 |
}
|
| 599 |
},
|
| 600 |
-
"total_flos": 4.
|
| 601 |
"train_batch_size": 16,
|
| 602 |
"trial_name": null,
|
| 603 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.19315651059150696,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-brain-alzheimer-detection/checkpoint-2800",
|
| 4 |
+
"epoch": 13.28125,
|
| 5 |
+
"eval_steps": 200,
|
| 6 |
+
"global_step": 3400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.1953125,
|
| 13 |
+
"grad_norm": 3.182270050048828,
|
| 14 |
"learning_rate": 2.0000000000000003e-06,
|
| 15 |
+
"loss": 1.401,
|
| 16 |
"step": 50
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.390625,
|
| 20 |
+
"grad_norm": 2.6782896518707275,
|
| 21 |
"learning_rate": 4.000000000000001e-06,
|
| 22 |
+
"loss": 1.1871,
|
| 23 |
"step": 100
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.5859375,
|
| 27 |
+
"grad_norm": 2.7876715660095215,
|
| 28 |
"learning_rate": 6e-06,
|
| 29 |
+
"loss": 1.035,
|
| 30 |
"step": 150
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.78125,
|
| 34 |
+
"grad_norm": 3.130065441131592,
|
| 35 |
"learning_rate": 8.000000000000001e-06,
|
| 36 |
+
"loss": 0.9772,
|
| 37 |
+
"step": 200
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.78125,
|
| 41 |
+
"eval_accuracy": 0.580078125,
|
| 42 |
+
"eval_loss": 0.9400424957275391,
|
| 43 |
+
"eval_runtime": 5.935,
|
| 44 |
+
"eval_samples_per_second": 172.535,
|
| 45 |
+
"eval_steps_per_second": 21.567,
|
| 46 |
"step": 200
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.9765625,
|
| 50 |
+
"grad_norm": 2.5494868755340576,
|
| 51 |
"learning_rate": 1e-05,
|
| 52 |
+
"loss": 0.8886,
|
| 53 |
"step": 250
|
| 54 |
},
|
| 55 |
{
|
| 56 |
"epoch": 1.171875,
|
| 57 |
+
"grad_norm": 1.62320876121521,
|
| 58 |
"learning_rate": 1.2e-05,
|
| 59 |
+
"loss": 0.8722,
|
| 60 |
"step": 300
|
| 61 |
},
|
| 62 |
{
|
| 63 |
"epoch": 1.3671875,
|
| 64 |
+
"grad_norm": 4.327115535736084,
|
| 65 |
"learning_rate": 1.4e-05,
|
| 66 |
+
"loss": 0.8154,
|
| 67 |
"step": 350
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 1.5625,
|
| 71 |
+
"grad_norm": 7.814743995666504,
|
| 72 |
"learning_rate": 1.6000000000000003e-05,
|
| 73 |
+
"loss": 0.7451,
|
| 74 |
+
"step": 400
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.5625,
|
| 78 |
+
"eval_accuracy": 0.6552734375,
|
| 79 |
+
"eval_loss": 0.7946789264678955,
|
| 80 |
+
"eval_runtime": 7.068,
|
| 81 |
+
"eval_samples_per_second": 144.879,
|
| 82 |
+
"eval_steps_per_second": 18.11,
|
| 83 |
"step": 400
|
| 84 |
},
|
| 85 |
{
|
| 86 |
"epoch": 1.7578125,
|
| 87 |
+
"grad_norm": 15.593164443969727,
|
| 88 |
"learning_rate": 1.8e-05,
|
| 89 |
+
"loss": 0.7383,
|
| 90 |
"step": 450
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"epoch": 1.953125,
|
| 94 |
+
"grad_norm": 5.583272933959961,
|
| 95 |
+
"learning_rate": 2e-05,
|
| 96 |
+
"loss": 0.733,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
"step": 500
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 2.1484375,
|
| 101 |
+
"grad_norm": 10.433585166931152,
|
| 102 |
+
"learning_rate": 1.9783549783549785e-05,
|
| 103 |
+
"loss": 0.5905,
|
| 104 |
"step": 550
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 2.34375,
|
| 108 |
+
"grad_norm": 4.751012802124023,
|
| 109 |
+
"learning_rate": 1.9567099567099568e-05,
|
| 110 |
+
"loss": 0.5701,
|
| 111 |
+
"step": 600
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 2.34375,
|
| 115 |
+
"eval_accuracy": 0.7236328125,
|
| 116 |
+
"eval_loss": 0.7641921639442444,
|
| 117 |
+
"eval_runtime": 5.5404,
|
| 118 |
+
"eval_samples_per_second": 184.824,
|
| 119 |
+
"eval_steps_per_second": 23.103,
|
| 120 |
"step": 600
|
| 121 |
},
|
| 122 |
{
|
| 123 |
"epoch": 2.5390625,
|
| 124 |
+
"grad_norm": 17.00213050842285,
|
| 125 |
+
"learning_rate": 1.9350649350649354e-05,
|
| 126 |
+
"loss": 0.6235,
|
| 127 |
"step": 650
|
| 128 |
},
|
| 129 |
{
|
| 130 |
"epoch": 2.734375,
|
| 131 |
+
"grad_norm": 10.933674812316895,
|
| 132 |
+
"learning_rate": 1.9134199134199138e-05,
|
| 133 |
+
"loss": 0.5385,
|
| 134 |
"step": 700
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 2.9296875,
|
| 138 |
+
"grad_norm": 9.459794044494629,
|
| 139 |
+
"learning_rate": 1.891774891774892e-05,
|
| 140 |
+
"loss": 0.4557,
|
| 141 |
"step": 750
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 3.125,
|
| 145 |
+
"grad_norm": 2.7128560543060303,
|
| 146 |
+
"learning_rate": 1.8701298701298704e-05,
|
| 147 |
+
"loss": 0.3704,
|
| 148 |
+
"step": 800
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 3.125,
|
| 152 |
+
"eval_accuracy": 0.7744140625,
|
| 153 |
+
"eval_loss": 0.5531814098358154,
|
| 154 |
+
"eval_runtime": 6.1892,
|
| 155 |
+
"eval_samples_per_second": 165.45,
|
| 156 |
+
"eval_steps_per_second": 20.681,
|
| 157 |
"step": 800
|
| 158 |
},
|
| 159 |
{
|
| 160 |
"epoch": 3.3203125,
|
| 161 |
+
"grad_norm": 6.759947776794434,
|
| 162 |
+
"learning_rate": 1.8484848484848487e-05,
|
| 163 |
+
"loss": 0.3876,
|
| 164 |
"step": 850
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 3.515625,
|
| 168 |
+
"grad_norm": 4.957170009613037,
|
| 169 |
+
"learning_rate": 1.826839826839827e-05,
|
| 170 |
+
"loss": 0.3014,
|
| 171 |
"step": 900
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 3.7109375,
|
| 175 |
+
"grad_norm": 4.334908485412598,
|
| 176 |
+
"learning_rate": 1.8051948051948053e-05,
|
| 177 |
+
"loss": 0.2523,
|
| 178 |
"step": 950
|
| 179 |
},
|
| 180 |
{
|
| 181 |
"epoch": 3.90625,
|
| 182 |
+
"grad_norm": 12.843859672546387,
|
| 183 |
+
"learning_rate": 1.7835497835497836e-05,
|
| 184 |
+
"loss": 0.2906,
|
| 185 |
"step": 1000
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 3.90625,
|
| 189 |
+
"eval_accuracy": 0.85546875,
|
| 190 |
+
"eval_loss": 0.4422951340675354,
|
| 191 |
+
"eval_runtime": 5.5363,
|
| 192 |
+
"eval_samples_per_second": 184.962,
|
| 193 |
+
"eval_steps_per_second": 23.12,
|
| 194 |
"step": 1000
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 4.1015625,
|
| 198 |
+
"grad_norm": 3.9840188026428223,
|
| 199 |
+
"learning_rate": 1.761904761904762e-05,
|
| 200 |
+
"loss": 0.2216,
|
| 201 |
"step": 1050
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 4.296875,
|
| 205 |
+
"grad_norm": 11.730119705200195,
|
| 206 |
+
"learning_rate": 1.7402597402597403e-05,
|
| 207 |
+
"loss": 0.1722,
|
| 208 |
"step": 1100
|
| 209 |
},
|
| 210 |
{
|
| 211 |
"epoch": 4.4921875,
|
| 212 |
+
"grad_norm": 2.089110851287842,
|
| 213 |
+
"learning_rate": 1.718614718614719e-05,
|
| 214 |
+
"loss": 0.1427,
|
| 215 |
"step": 1150
|
| 216 |
},
|
| 217 |
{
|
| 218 |
"epoch": 4.6875,
|
| 219 |
+
"grad_norm": 0.5950905084609985,
|
| 220 |
+
"learning_rate": 1.6969696969696972e-05,
|
| 221 |
+
"loss": 0.1636,
|
| 222 |
+
"step": 1200
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.6875,
|
| 226 |
+
"eval_accuracy": 0.900390625,
|
| 227 |
+
"eval_loss": 0.32256966829299927,
|
| 228 |
+
"eval_runtime": 6.2703,
|
| 229 |
+
"eval_samples_per_second": 163.31,
|
| 230 |
+
"eval_steps_per_second": 20.414,
|
| 231 |
"step": 1200
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 4.8828125,
|
| 235 |
+
"grad_norm": 4.693203926086426,
|
| 236 |
+
"learning_rate": 1.6753246753246756e-05,
|
| 237 |
+
"loss": 0.1622,
|
| 238 |
"step": 1250
|
| 239 |
},
|
| 240 |
{
|
| 241 |
"epoch": 5.078125,
|
| 242 |
+
"grad_norm": 0.24195541441440582,
|
| 243 |
+
"learning_rate": 1.653679653679654e-05,
|
| 244 |
+
"loss": 0.099,
|
| 245 |
"step": 1300
|
| 246 |
},
|
| 247 |
{
|
| 248 |
"epoch": 5.2734375,
|
| 249 |
+
"grad_norm": 0.16812123358249664,
|
| 250 |
+
"learning_rate": 1.6320346320346322e-05,
|
| 251 |
+
"loss": 0.0979,
|
| 252 |
"step": 1350
|
| 253 |
},
|
| 254 |
{
|
| 255 |
"epoch": 5.46875,
|
| 256 |
+
"grad_norm": 19.697050094604492,
|
| 257 |
+
"learning_rate": 1.6103896103896105e-05,
|
| 258 |
+
"loss": 0.0837,
|
| 259 |
"step": 1400
|
| 260 |
},
|
| 261 |
{
|
| 262 |
+
"epoch": 5.46875,
|
| 263 |
+
"eval_accuracy": 0.90234375,
|
| 264 |
+
"eval_loss": 0.3483480215072632,
|
| 265 |
+
"eval_runtime": 5.5142,
|
| 266 |
+
"eval_samples_per_second": 185.702,
|
| 267 |
+
"eval_steps_per_second": 23.213,
|
| 268 |
+
"step": 1400
|
| 269 |
},
|
| 270 |
{
|
| 271 |
+
"epoch": 5.6640625,
|
| 272 |
+
"grad_norm": 0.35744136571884155,
|
| 273 |
+
"learning_rate": 1.5887445887445888e-05,
|
| 274 |
+
"loss": 0.118,
|
| 275 |
+
"step": 1450
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 5.859375,
|
| 279 |
+
"grad_norm": 0.1951870173215866,
|
| 280 |
+
"learning_rate": 1.567099567099567e-05,
|
| 281 |
+
"loss": 0.0826,
|
|
|
|
|
|
|
| 282 |
"step": 1500
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 6.0546875,
|
| 286 |
+
"grad_norm": 27.74083137512207,
|
| 287 |
+
"learning_rate": 1.5454545454545454e-05,
|
| 288 |
+
"loss": 0.0697,
|
| 289 |
"step": 1550
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 6.25,
|
| 293 |
+
"grad_norm": 1.9396530389785767,
|
| 294 |
+
"learning_rate": 1.523809523809524e-05,
|
| 295 |
+
"loss": 0.0368,
|
| 296 |
+
"step": 1600
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 6.25,
|
| 300 |
+
"eval_accuracy": 0.939453125,
|
| 301 |
+
"eval_loss": 0.24231916666030884,
|
| 302 |
+
"eval_runtime": 5.9487,
|
| 303 |
+
"eval_samples_per_second": 172.139,
|
| 304 |
+
"eval_steps_per_second": 21.517,
|
| 305 |
"step": 1600
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 6.4453125,
|
| 309 |
+
"grad_norm": 0.1338203400373459,
|
| 310 |
+
"learning_rate": 1.5021645021645024e-05,
|
| 311 |
+
"loss": 0.0482,
|
| 312 |
"step": 1650
|
| 313 |
},
|
| 314 |
{
|
| 315 |
"epoch": 6.640625,
|
| 316 |
+
"grad_norm": 0.16084209084510803,
|
| 317 |
+
"learning_rate": 1.4805194805194807e-05,
|
| 318 |
+
"loss": 0.0382,
|
| 319 |
"step": 1700
|
| 320 |
},
|
| 321 |
{
|
| 322 |
"epoch": 6.8359375,
|
| 323 |
+
"grad_norm": 13.326003074645996,
|
| 324 |
+
"learning_rate": 1.458874458874459e-05,
|
| 325 |
+
"loss": 0.1169,
|
| 326 |
"step": 1750
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 7.03125,
|
| 330 |
+
"grad_norm": 0.1118919774889946,
|
| 331 |
+
"learning_rate": 1.4372294372294374e-05,
|
| 332 |
+
"loss": 0.063,
|
| 333 |
+
"step": 1800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 7.03125,
|
| 337 |
+
"eval_accuracy": 0.927734375,
|
| 338 |
+
"eval_loss": 0.30914944410324097,
|
| 339 |
+
"eval_runtime": 5.5136,
|
| 340 |
+
"eval_samples_per_second": 185.722,
|
| 341 |
+
"eval_steps_per_second": 23.215,
|
| 342 |
"step": 1800
|
| 343 |
},
|
| 344 |
{
|
| 345 |
"epoch": 7.2265625,
|
| 346 |
+
"grad_norm": 0.11637281626462936,
|
| 347 |
+
"learning_rate": 1.4155844155844157e-05,
|
| 348 |
+
"loss": 0.0385,
|
| 349 |
"step": 1850
|
| 350 |
},
|
| 351 |
{
|
| 352 |
"epoch": 7.421875,
|
| 353 |
+
"grad_norm": 0.09042005985975266,
|
| 354 |
+
"learning_rate": 1.3939393939393942e-05,
|
| 355 |
+
"loss": 0.0277,
|
| 356 |
"step": 1900
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 7.6171875,
|
| 360 |
+
"grad_norm": 0.07495836168527603,
|
| 361 |
+
"learning_rate": 1.3722943722943725e-05,
|
| 362 |
+
"loss": 0.0214,
|
| 363 |
"step": 1950
|
| 364 |
},
|
| 365 |
{
|
| 366 |
"epoch": 7.8125,
|
| 367 |
+
"grad_norm": 0.06756331771612167,
|
| 368 |
+
"learning_rate": 1.3506493506493508e-05,
|
| 369 |
+
"loss": 0.047,
|
| 370 |
"step": 2000
|
| 371 |
},
|
| 372 |
{
|
| 373 |
"epoch": 7.8125,
|
| 374 |
+
"eval_accuracy": 0.90234375,
|
| 375 |
+
"eval_loss": 0.3906704783439636,
|
| 376 |
+
"eval_runtime": 6.0843,
|
| 377 |
+
"eval_samples_per_second": 168.303,
|
| 378 |
+
"eval_steps_per_second": 21.038,
|
| 379 |
"step": 2000
|
| 380 |
},
|
| 381 |
{
|
| 382 |
"epoch": 8.0078125,
|
| 383 |
+
"grad_norm": 0.3716655671596527,
|
| 384 |
+
"learning_rate": 1.3290043290043291e-05,
|
| 385 |
+
"loss": 0.0262,
|
| 386 |
"step": 2050
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 8.203125,
|
| 390 |
+
"grad_norm": 0.40872329473495483,
|
| 391 |
+
"learning_rate": 1.3073593073593074e-05,
|
| 392 |
+
"loss": 0.0139,
|
| 393 |
"step": 2100
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 8.3984375,
|
| 397 |
+
"grad_norm": 0.05600840970873833,
|
| 398 |
"learning_rate": 1.2861471861471862e-05,
|
| 399 |
+
"loss": 0.0161,
|
| 400 |
"step": 2150
|
| 401 |
},
|
| 402 |
{
|
| 403 |
"epoch": 8.59375,
|
| 404 |
+
"grad_norm": 0.055402472615242004,
|
| 405 |
"learning_rate": 1.2645021645021646e-05,
|
| 406 |
+
"loss": 0.0127,
|
| 407 |
+
"step": 2200
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 8.59375,
|
| 411 |
+
"eval_accuracy": 0.9560546875,
|
| 412 |
+
"eval_loss": 0.20022055506706238,
|
| 413 |
+
"eval_runtime": 5.6277,
|
| 414 |
+
"eval_samples_per_second": 181.956,
|
| 415 |
+
"eval_steps_per_second": 22.744,
|
| 416 |
"step": 2200
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 8.7890625,
|
| 420 |
+
"grad_norm": 0.054736897349357605,
|
| 421 |
"learning_rate": 1.242857142857143e-05,
|
| 422 |
+
"loss": 0.0135,
|
| 423 |
"step": 2250
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"epoch": 8.984375,
|
| 427 |
+
"grad_norm": 0.04909529909491539,
|
| 428 |
"learning_rate": 1.2212121212121213e-05,
|
| 429 |
+
"loss": 0.0209,
|
| 430 |
"step": 2300
|
| 431 |
},
|
| 432 |
{
|
| 433 |
"epoch": 9.1796875,
|
| 434 |
+
"grad_norm": 0.04906412959098816,
|
| 435 |
"learning_rate": 1.1995670995670996e-05,
|
| 436 |
+
"loss": 0.0107,
|
| 437 |
"step": 2350
|
| 438 |
},
|
| 439 |
{
|
| 440 |
"epoch": 9.375,
|
| 441 |
+
"grad_norm": 0.04470803588628769,
|
| 442 |
"learning_rate": 1.1779220779220779e-05,
|
| 443 |
+
"loss": 0.0102,
|
| 444 |
+
"step": 2400
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 9.375,
|
| 448 |
+
"eval_accuracy": 0.9306640625,
|
| 449 |
+
"eval_loss": 0.30005937814712524,
|
| 450 |
+
"eval_runtime": 6.136,
|
| 451 |
+
"eval_samples_per_second": 166.885,
|
| 452 |
+
"eval_steps_per_second": 20.861,
|
| 453 |
"step": 2400
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 9.5703125,
|
| 457 |
+
"grad_norm": 0.04809673875570297,
|
| 458 |
"learning_rate": 1.1562770562770564e-05,
|
| 459 |
+
"loss": 0.01,
|
| 460 |
"step": 2450
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 9.765625,
|
| 464 |
+
"grad_norm": 0.053658001124858856,
|
| 465 |
"learning_rate": 1.1346320346320347e-05,
|
| 466 |
+
"loss": 0.0096,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 467 |
"step": 2500
|
| 468 |
},
|
| 469 |
{
|
| 470 |
"epoch": 9.9609375,
|
| 471 |
+
"grad_norm": 0.05094292759895325,
|
| 472 |
"learning_rate": 1.112987012987013e-05,
|
| 473 |
+
"loss": 0.0085,
|
| 474 |
"step": 2550
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 10.15625,
|
| 478 |
+
"grad_norm": 0.03785080462694168,
|
| 479 |
"learning_rate": 1.0913419913419913e-05,
|
| 480 |
+
"loss": 0.0086,
|
| 481 |
+
"step": 2600
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 10.15625,
|
| 485 |
+
"eval_accuracy": 0.951171875,
|
| 486 |
+
"eval_loss": 0.19980870187282562,
|
| 487 |
+
"eval_runtime": 5.5065,
|
| 488 |
+
"eval_samples_per_second": 185.964,
|
| 489 |
+
"eval_steps_per_second": 23.245,
|
| 490 |
"step": 2600
|
| 491 |
},
|
| 492 |
{
|
| 493 |
"epoch": 10.3515625,
|
| 494 |
+
"grad_norm": 0.048504624515771866,
|
| 495 |
"learning_rate": 1.0696969696969696e-05,
|
| 496 |
+
"loss": 0.0078,
|
| 497 |
"step": 2650
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 10.546875,
|
| 501 |
+
"grad_norm": 0.040160298347473145,
|
| 502 |
"learning_rate": 1.0480519480519481e-05,
|
| 503 |
+
"loss": 0.0136,
|
| 504 |
"step": 2700
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"epoch": 10.7421875,
|
| 508 |
+
"grad_norm": 0.03454400226473808,
|
| 509 |
"learning_rate": 1.0264069264069264e-05,
|
| 510 |
+
"loss": 0.0077,
|
| 511 |
"step": 2750
|
| 512 |
},
|
| 513 |
{
|
| 514 |
"epoch": 10.9375,
|
| 515 |
+
"grad_norm": 0.03354168310761452,
|
| 516 |
"learning_rate": 1.0047619047619048e-05,
|
| 517 |
+
"loss": 0.0073,
|
| 518 |
+
"step": 2800
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 10.9375,
|
| 522 |
+
"eval_accuracy": 0.958984375,
|
| 523 |
+
"eval_loss": 0.19315651059150696,
|
| 524 |
+
"eval_runtime": 6.0024,
|
| 525 |
+
"eval_samples_per_second": 170.598,
|
| 526 |
+
"eval_steps_per_second": 21.325,
|
| 527 |
"step": 2800
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 11.1328125,
|
| 531 |
+
"grad_norm": 0.033880043774843216,
|
| 532 |
"learning_rate": 9.831168831168832e-06,
|
| 533 |
+
"loss": 0.0071,
|
| 534 |
"step": 2850
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 11.328125,
|
| 538 |
+
"grad_norm": 0.032037124037742615,
|
| 539 |
"learning_rate": 9.614718614718616e-06,
|
| 540 |
+
"loss": 0.0066,
|
| 541 |
"step": 2900
|
| 542 |
},
|
| 543 |
{
|
| 544 |
"epoch": 11.5234375,
|
| 545 |
+
"grad_norm": 0.0329328328371048,
|
| 546 |
"learning_rate": 9.398268398268399e-06,
|
| 547 |
+
"loss": 0.0066,
|
| 548 |
"step": 2950
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 11.71875,
|
| 552 |
+
"grad_norm": 0.02779380977153778,
|
| 553 |
"learning_rate": 9.181818181818184e-06,
|
| 554 |
+
"loss": 0.0064,
|
| 555 |
"step": 3000
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 11.71875,
|
| 559 |
+
"eval_accuracy": 0.9560546875,
|
| 560 |
+
"eval_loss": 0.19875724613666534,
|
| 561 |
+
"eval_runtime": 5.8842,
|
| 562 |
+
"eval_samples_per_second": 174.025,
|
| 563 |
+
"eval_steps_per_second": 21.753,
|
| 564 |
"step": 3000
|
| 565 |
},
|
| 566 |
{
|
| 567 |
"epoch": 11.9140625,
|
| 568 |
+
"grad_norm": 0.028905285522341728,
|
| 569 |
"learning_rate": 8.965367965367967e-06,
|
| 570 |
+
"loss": 0.0063,
|
| 571 |
"step": 3050
|
| 572 |
},
|
| 573 |
{
|
| 574 |
"epoch": 12.109375,
|
| 575 |
+
"grad_norm": 0.027350405231118202,
|
| 576 |
"learning_rate": 8.74891774891775e-06,
|
| 577 |
+
"loss": 0.0058,
|
| 578 |
"step": 3100
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 12.3046875,
|
| 582 |
+
"grad_norm": 0.026442719623446465,
|
| 583 |
"learning_rate": 8.532467532467533e-06,
|
| 584 |
+
"loss": 0.0059,
|
| 585 |
"step": 3150
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 12.5,
|
| 589 |
+
"grad_norm": 0.028089182451367378,
|
| 590 |
"learning_rate": 8.316017316017316e-06,
|
| 591 |
+
"loss": 0.0056,
|
| 592 |
+
"step": 3200
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 12.5,
|
| 596 |
+
"eval_accuracy": 0.9580078125,
|
| 597 |
+
"eval_loss": 0.19929003715515137,
|
| 598 |
+
"eval_runtime": 6.1143,
|
| 599 |
+
"eval_samples_per_second": 167.475,
|
| 600 |
+
"eval_steps_per_second": 20.934,
|
| 601 |
"step": 3200
|
| 602 |
},
|
| 603 |
{
|
| 604 |
"epoch": 12.6953125,
|
| 605 |
+
"grad_norm": 0.02532665804028511,
|
| 606 |
"learning_rate": 8.099567099567101e-06,
|
| 607 |
+
"loss": 0.0053,
|
| 608 |
"step": 3250
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 12.890625,
|
| 612 |
+
"grad_norm": 0.024192778393626213,
|
| 613 |
"learning_rate": 7.883116883116884e-06,
|
| 614 |
+
"loss": 0.0055,
|
| 615 |
"step": 3300
|
| 616 |
},
|
| 617 |
{
|
| 618 |
"epoch": 13.0859375,
|
| 619 |
+
"grad_norm": 0.028951533138751984,
|
| 620 |
"learning_rate": 7.666666666666667e-06,
|
| 621 |
+
"loss": 0.0053,
|
| 622 |
"step": 3350
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 13.28125,
|
| 626 |
+
"grad_norm": 0.024818725883960724,
|
| 627 |
"learning_rate": 7.450216450216451e-06,
|
| 628 |
+
"loss": 0.0049,
|
| 629 |
"step": 3400
|
| 630 |
},
|
| 631 |
{
|
| 632 |
+
"epoch": 13.28125,
|
| 633 |
+
"eval_accuracy": 0.958984375,
|
| 634 |
+
"eval_loss": 0.2046782374382019,
|
| 635 |
+
"eval_runtime": 5.5223,
|
| 636 |
+
"eval_samples_per_second": 185.43,
|
| 637 |
+
"eval_steps_per_second": 23.179,
|
| 638 |
+
"step": 3400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 639 |
},
|
| 640 |
{
|
| 641 |
+
"epoch": 13.28125,
|
| 642 |
+
"step": 3400,
|
| 643 |
+
"total_flos": 4.2156398005714944e+18,
|
| 644 |
+
"train_loss": 0.23371873112285838,
|
| 645 |
+
"train_runtime": 1043.0698,
|
| 646 |
+
"train_samples_per_second": 78.537,
|
| 647 |
+
"train_steps_per_second": 4.909
|
| 648 |
}
|
| 649 |
],
|
| 650 |
"logging_steps": 50,
|
| 651 |
"max_steps": 5120,
|
| 652 |
"num_input_tokens_seen": 0,
|
| 653 |
"num_train_epochs": 20,
|
| 654 |
+
"save_steps": 200,
|
| 655 |
"stateful_callbacks": {
|
| 656 |
"EarlyStoppingCallback": {
|
| 657 |
"args": {
|
|
|
|
| 673 |
"attributes": {}
|
| 674 |
}
|
| 675 |
},
|
| 676 |
+
"total_flos": 4.2156398005714944e+18,
|
| 677 |
"train_batch_size": 16,
|
| 678 |
"trial_name": null,
|
| 679 |
"trial_params": null
|