End of training
Browse files- all_results.json +9 -9
- eval_results.json +5 -5
- pytorch_model.bin +1 -1
- runs/May23_23-28-14_gbnwp-pod015-2.ipu.graphcore.ai/events.out.tfevents.1653345392.gbnwp-pod015-2.ipu.graphcore.ai +2 -2
- runs/May23_23-28-14_gbnwp-pod015-2.ipu.graphcore.ai/events.out.tfevents.1653346022.gbnwp-pod015-2.ipu.graphcore.ai +3 -0
- train_results.json +4 -4
- trainer_state.json +198 -198
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 5.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second":
|
| 11 |
-
"train_steps_per_second":
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9821639150943396,
|
| 4 |
+
"eval_loss": 0.0848388671875,
|
| 5 |
+
"eval_runtime": 229.7479,
|
| 6 |
+
"eval_samples_per_second": 29.589,
|
| 7 |
+
"eval_steps_per_second": 0.235,
|
| 8 |
+
"train_loss": 0.32014525516290115,
|
| 9 |
+
"train_runtime": 392.8902,
|
| 10 |
+
"train_samples_per_second": 650.233,
|
| 11 |
+
"train_steps_per_second": 5.078
|
| 12 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 5.0,
|
| 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": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9821639150943396,
|
| 4 |
+
"eval_loss": 0.0848388671875,
|
| 5 |
+
"eval_runtime": 229.7479,
|
| 6 |
+
"eval_samples_per_second": 29.589,
|
| 7 |
+
"eval_steps_per_second": 0.235
|
| 8 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 189210261
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2821bc9de1117450dbce5c0cf271cbd824e1fe6097d3bb2161b0f4189ec04635
|
| 3 |
size 189210261
|
runs/May23_23-28-14_gbnwp-pod015-2.ipu.graphcore.ai/events.out.tfevents.1653345392.gbnwp-pod015-2.ipu.graphcore.ai
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b6df7c6b9168aa97507327127d587a5420b73d64ab5e6485d4638a498216df1
|
| 3 |
+
size 35906
|
runs/May23_23-28-14_gbnwp-pod015-2.ipu.graphcore.ai/events.out.tfevents.1653346022.gbnwp-pod015-2.ipu.graphcore.ai
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ffcaf02470b31d560fa1dfff56228916ee00c9f71e92db41d5e6279039b982e
|
| 3 |
+
size 40
|
train_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 5.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples_per_second":
|
| 6 |
-
"train_steps_per_second":
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.32014525516290115,
|
| 4 |
+
"train_runtime": 392.8902,
|
| 5 |
+
"train_samples_per_second": 650.233,
|
| 6 |
+
"train_steps_per_second": 5.078
|
| 7 |
}
|
trainer_state.json
CHANGED
|
@@ -10,55 +10,55 @@
|
|
| 10 |
{
|
| 11 |
"epoch": 0.03,
|
| 12 |
"learning_rate": 1.5e-06,
|
| 13 |
-
"loss": 2.
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.05,
|
| 18 |
"learning_rate": 3e-06,
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.08,
|
| 24 |
"learning_rate": 4.5e-06,
|
| 25 |
-
"loss": 2.
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.1,
|
| 30 |
"learning_rate": 6e-06,
|
| 31 |
-
"loss": 2.
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.13,
|
| 36 |
"learning_rate": 7.5e-06,
|
| 37 |
-
"loss": 2.
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.15,
|
| 42 |
"learning_rate": 9e-06,
|
| 43 |
-
"loss": 1.
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.18,
|
| 48 |
"learning_rate": 1.05e-05,
|
| 49 |
-
"loss": 1.
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.2,
|
| 54 |
"learning_rate": 1.2e-05,
|
| 55 |
-
"loss": 1.
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.23,
|
| 60 |
"learning_rate": 1.3500000000000001e-05,
|
| 61 |
-
"loss": 1.
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
|
@@ -70,19 +70,19 @@
|
|
| 70 |
{
|
| 71 |
"epoch": 0.28,
|
| 72 |
"learning_rate": 1.65e-05,
|
| 73 |
-
"loss": 1.
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.3,
|
| 78 |
"learning_rate": 1.8e-05,
|
| 79 |
-
"loss": 1.
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.33,
|
| 84 |
"learning_rate": 1.95e-05,
|
| 85 |
-
"loss": 1.
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
|
@@ -94,709 +94,709 @@
|
|
| 94 |
{
|
| 95 |
"epoch": 0.38,
|
| 96 |
"learning_rate": 2.25e-05,
|
| 97 |
-
"loss": 1.
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.4,
|
| 102 |
"learning_rate": 2.4e-05,
|
| 103 |
-
"loss": 1.
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.43,
|
| 108 |
"learning_rate": 2.55e-05,
|
| 109 |
-
"loss": 1.
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.45,
|
| 114 |
"learning_rate": 2.7000000000000002e-05,
|
| 115 |
-
"loss": 1.
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.48,
|
| 120 |
"learning_rate": 2.8499999999999998e-05,
|
| 121 |
-
"loss": 1.
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.5,
|
| 126 |
"learning_rate": 3e-05,
|
| 127 |
-
"loss": 1.
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.53,
|
| 132 |
"learning_rate": 2.9832869080779945e-05,
|
| 133 |
-
"loss": 1.
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.55,
|
| 138 |
"learning_rate": 2.9665738161559886e-05,
|
| 139 |
-
"loss": 1.
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.58,
|
| 144 |
"learning_rate": 2.9498607242339834e-05,
|
| 145 |
-
"loss": 1.
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.6,
|
| 150 |
"learning_rate": 2.933147632311978e-05,
|
| 151 |
-
"loss": 1.
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.63,
|
| 156 |
"learning_rate": 2.916434540389972e-05,
|
| 157 |
-
"loss": 1.
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.65,
|
| 162 |
"learning_rate": 2.8997214484679665e-05,
|
| 163 |
-
"loss": 1.
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.68,
|
| 168 |
"learning_rate": 2.8830083565459613e-05,
|
| 169 |
-
"loss": 0.
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.7,
|
| 174 |
"learning_rate": 2.8662952646239554e-05,
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.73,
|
| 180 |
"learning_rate": 2.84958217270195e-05,
|
| 181 |
-
"loss": 0.
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.75,
|
| 186 |
"learning_rate": 2.8328690807799443e-05,
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.78,
|
| 192 |
"learning_rate": 2.8161559888579388e-05,
|
| 193 |
-
"loss": 0.
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.8,
|
| 198 |
"learning_rate": 2.7994428969359332e-05,
|
| 199 |
-
"loss": 0.
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.83,
|
| 204 |
"learning_rate": 2.7827298050139277e-05,
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.85,
|
| 210 |
"learning_rate": 2.7660167130919218e-05,
|
| 211 |
-
"loss": 0.
|
| 212 |
"step": 340
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.88,
|
| 216 |
"learning_rate": 2.7493036211699166e-05,
|
| 217 |
-
"loss": 0.
|
| 218 |
"step": 350
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.9,
|
| 222 |
"learning_rate": 2.732590529247911e-05,
|
| 223 |
-
"loss": 0.
|
| 224 |
"step": 360
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.93,
|
| 228 |
"learning_rate": 2.7158774373259055e-05,
|
| 229 |
-
"loss": 0.
|
| 230 |
"step": 370
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.95,
|
| 234 |
"learning_rate": 2.6991643454038996e-05,
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 380
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.98,
|
| 240 |
"learning_rate": 2.6824512534818944e-05,
|
| 241 |
-
"loss": 0.
|
| 242 |
"step": 390
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 1.0,
|
| 246 |
"learning_rate": 2.665738161559889e-05,
|
| 247 |
-
"loss": 0.
|
| 248 |
"step": 400
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 1.03,
|
| 252 |
"learning_rate": 2.649025069637883e-05,
|
| 253 |
-
"loss": 0.
|
| 254 |
"step": 410
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 1.05,
|
| 258 |
"learning_rate": 2.6323119777158774e-05,
|
| 259 |
-
"loss": 0.
|
| 260 |
"step": 420
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 1.08,
|
| 264 |
"learning_rate": 2.6155988857938722e-05,
|
| 265 |
-
"loss": 0.
|
| 266 |
"step": 430
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 1.1,
|
| 270 |
"learning_rate": 2.5988857938718663e-05,
|
| 271 |
-
"loss": 0.
|
| 272 |
"step": 440
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 1.13,
|
| 276 |
"learning_rate": 2.5821727019498608e-05,
|
| 277 |
-
"loss": 0.
|
| 278 |
"step": 450
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 1.15,
|
| 282 |
"learning_rate": 2.5654596100278553e-05,
|
| 283 |
-
"loss": 0.
|
| 284 |
"step": 460
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 1.18,
|
| 288 |
"learning_rate": 2.5487465181058497e-05,
|
| 289 |
-
"loss": 0.
|
| 290 |
"step": 470
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 1.2,
|
| 294 |
"learning_rate": 2.532033426183844e-05,
|
| 295 |
-
"loss": 0.
|
| 296 |
"step": 480
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 1.23,
|
| 300 |
"learning_rate": 2.5153203342618386e-05,
|
| 301 |
-
"loss": 0.
|
| 302 |
"step": 490
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 1.25,
|
| 306 |
"learning_rate": 2.4986072423398327e-05,
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 1.28,
|
| 312 |
"learning_rate": 2.4818941504178275e-05,
|
| 313 |
-
"loss": 0.
|
| 314 |
"step": 510
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 1.3,
|
| 318 |
"learning_rate": 2.465181058495822e-05,
|
| 319 |
-
"loss": 0.
|
| 320 |
"step": 520
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 1.33,
|
| 324 |
"learning_rate": 2.448467966573816e-05,
|
| 325 |
-
"loss": 0.
|
| 326 |
"step": 530
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 1.35,
|
| 330 |
"learning_rate": 2.4317548746518106e-05,
|
| 331 |
-
"loss": 0.
|
| 332 |
"step": 540
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 1.38,
|
| 336 |
"learning_rate": 2.415041782729805e-05,
|
| 337 |
-
"loss": 0.
|
| 338 |
"step": 550
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 1.4,
|
| 342 |
"learning_rate": 2.3983286908077995e-05,
|
| 343 |
-
"loss": 0.
|
| 344 |
"step": 560
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 1.43,
|
| 348 |
"learning_rate": 2.381615598885794e-05,
|
| 349 |
-
"loss": 0.
|
| 350 |
"step": 570
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 1.45,
|
| 354 |
"learning_rate": 2.3649025069637884e-05,
|
| 355 |
-
"loss": 0.
|
| 356 |
"step": 580
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.48,
|
| 360 |
"learning_rate": 2.3481894150417825e-05,
|
| 361 |
-
"loss": 0.
|
| 362 |
"step": 590
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.5,
|
| 366 |
"learning_rate": 2.3314763231197773e-05,
|
| 367 |
-
"loss": 0.
|
| 368 |
"step": 600
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 1.53,
|
| 372 |
"learning_rate": 2.3147632311977718e-05,
|
| 373 |
-
"loss": 0.
|
| 374 |
"step": 610
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 1.55,
|
| 378 |
"learning_rate": 2.298050139275766e-05,
|
| 379 |
-
"loss": 0.
|
| 380 |
"step": 620
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 1.58,
|
| 384 |
"learning_rate": 2.2813370473537603e-05,
|
| 385 |
-
"loss": 0.
|
| 386 |
"step": 630
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 1.6,
|
| 390 |
"learning_rate": 2.264623955431755e-05,
|
| 391 |
-
"loss": 0.
|
| 392 |
"step": 640
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 1.63,
|
| 396 |
"learning_rate": 2.2479108635097492e-05,
|
| 397 |
-
"loss": 0.
|
| 398 |
"step": 650
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 1.65,
|
| 402 |
"learning_rate": 2.2311977715877437e-05,
|
| 403 |
-
"loss": 0.
|
| 404 |
"step": 660
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 1.68,
|
| 408 |
"learning_rate": 2.214484679665738e-05,
|
| 409 |
-
"loss": 0.
|
| 410 |
"step": 670
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.7,
|
| 414 |
"learning_rate": 2.1977715877437326e-05,
|
| 415 |
-
"loss": 0.
|
| 416 |
"step": 680
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 1.73,
|
| 420 |
"learning_rate": 2.181058495821727e-05,
|
| 421 |
-
"loss": 0.
|
| 422 |
"step": 690
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 1.75,
|
| 426 |
"learning_rate": 2.1643454038997215e-05,
|
| 427 |
-
"loss": 0.
|
| 428 |
"step": 700
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.78,
|
| 432 |
"learning_rate": 2.147632311977716e-05,
|
| 433 |
-
"loss": 0.
|
| 434 |
"step": 710
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 1.8,
|
| 438 |
"learning_rate": 2.1309192200557104e-05,
|
| 439 |
-
"loss": 0.
|
| 440 |
"step": 720
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 1.83,
|
| 444 |
"learning_rate": 2.114206128133705e-05,
|
| 445 |
-
"loss": 0.
|
| 446 |
"step": 730
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.85,
|
| 450 |
"learning_rate": 2.0974930362116993e-05,
|
| 451 |
-
"loss": 0.
|
| 452 |
"step": 740
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 1.88,
|
| 456 |
"learning_rate": 2.0807799442896935e-05,
|
| 457 |
-
"loss": 0.
|
| 458 |
"step": 750
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 1.9,
|
| 462 |
"learning_rate": 2.0640668523676883e-05,
|
| 463 |
-
"loss": 0.
|
| 464 |
"step": 760
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.93,
|
| 468 |
"learning_rate": 2.0473537604456827e-05,
|
| 469 |
-
"loss": 0.
|
| 470 |
"step": 770
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 1.95,
|
| 474 |
"learning_rate": 2.0306406685236768e-05,
|
| 475 |
-
"loss": 0.
|
| 476 |
"step": 780
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 1.98,
|
| 480 |
"learning_rate": 2.0139275766016713e-05,
|
| 481 |
-
"loss": 0.
|
| 482 |
"step": 790
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 2.01,
|
| 486 |
"learning_rate": 1.997214484679666e-05,
|
| 487 |
-
"loss": 0.
|
| 488 |
"step": 800
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 2.03,
|
| 492 |
"learning_rate": 1.9805013927576602e-05,
|
| 493 |
-
"loss": 0.
|
| 494 |
"step": 810
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 2.06,
|
| 498 |
"learning_rate": 1.9637883008356546e-05,
|
| 499 |
-
"loss": 0.
|
| 500 |
"step": 820
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 2.08,
|
| 504 |
"learning_rate": 1.947075208913649e-05,
|
| 505 |
-
"loss": 0.
|
| 506 |
"step": 830
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 2.11,
|
| 510 |
"learning_rate": 1.9303621169916436e-05,
|
| 511 |
-
"loss": 0.
|
| 512 |
"step": 840
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 2.13,
|
| 516 |
"learning_rate": 1.913649025069638e-05,
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 850
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 2.16,
|
| 522 |
"learning_rate": 1.8969359331476325e-05,
|
| 523 |
-
"loss": 0.
|
| 524 |
"step": 860
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 2.18,
|
| 528 |
"learning_rate": 1.8802228412256266e-05,
|
| 529 |
-
"loss": 0.
|
| 530 |
"step": 870
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 2.21,
|
| 534 |
"learning_rate": 1.863509749303621e-05,
|
| 535 |
-
"loss": 0.
|
| 536 |
"step": 880
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 2.23,
|
| 540 |
"learning_rate": 1.846796657381616e-05,
|
| 541 |
-
"loss": 0.
|
| 542 |
"step": 890
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 2.26,
|
| 546 |
"learning_rate": 1.83008356545961e-05,
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 900
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 2.28,
|
| 552 |
"learning_rate": 1.8133704735376044e-05,
|
| 553 |
-
"loss": 0.
|
| 554 |
"step": 910
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 2.31,
|
| 558 |
"learning_rate": 1.796657381615599e-05,
|
| 559 |
-
"loss": 0.
|
| 560 |
"step": 920
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 2.33,
|
| 564 |
"learning_rate": 1.7799442896935933e-05,
|
| 565 |
-
"loss": 0.
|
| 566 |
"step": 930
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 2.36,
|
| 570 |
"learning_rate": 1.7632311977715878e-05,
|
| 571 |
-
"loss": 0.
|
| 572 |
"step": 940
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 2.38,
|
| 576 |
"learning_rate": 1.7465181058495822e-05,
|
| 577 |
-
"loss": 0.
|
| 578 |
"step": 950
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 2.41,
|
| 582 |
"learning_rate": 1.7298050139275764e-05,
|
| 583 |
-
"loss": 0.
|
| 584 |
"step": 960
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 2.43,
|
| 588 |
"learning_rate": 1.713091922005571e-05,
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 970
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 2.46,
|
| 594 |
"learning_rate": 1.6963788300835656e-05,
|
| 595 |
-
"loss": 0.
|
| 596 |
"step": 980
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 2.48,
|
| 600 |
"learning_rate": 1.6796657381615597e-05,
|
| 601 |
-
"loss": 0.
|
| 602 |
"step": 990
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 2.51,
|
| 606 |
"learning_rate": 1.6629526462395542e-05,
|
| 607 |
-
"loss": 0.
|
| 608 |
"step": 1000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 2.53,
|
| 612 |
"learning_rate": 1.646239554317549e-05,
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 1010
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 2.56,
|
| 618 |
"learning_rate": 1.6295264623955434e-05,
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 1020
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 2.58,
|
| 624 |
"learning_rate": 1.6128133704735375e-05,
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 1030
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 2.61,
|
| 630 |
"learning_rate": 1.596100278551532e-05,
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 1040
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 2.63,
|
| 636 |
"learning_rate": 1.5793871866295268e-05,
|
| 637 |
-
"loss": 0.
|
| 638 |
"step": 1050
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 2.66,
|
| 642 |
"learning_rate": 1.562674094707521e-05,
|
| 643 |
-
"loss": 0.
|
| 644 |
"step": 1060
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 2.68,
|
| 648 |
"learning_rate": 1.5459610027855154e-05,
|
| 649 |
-
"loss": 0.
|
| 650 |
"step": 1070
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 2.71,
|
| 654 |
"learning_rate": 1.5292479108635098e-05,
|
| 655 |
-
"loss": 0.
|
| 656 |
"step": 1080
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 2.73,
|
| 660 |
"learning_rate": 1.5125348189415043e-05,
|
| 661 |
-
"loss": 0.
|
| 662 |
"step": 1090
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 2.76,
|
| 666 |
"learning_rate": 1.4958217270194987e-05,
|
| 667 |
-
"loss": 0.
|
| 668 |
"step": 1100
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 2.78,
|
| 672 |
"learning_rate": 1.479108635097493e-05,
|
| 673 |
-
"loss": 0.
|
| 674 |
"step": 1110
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 2.81,
|
| 678 |
"learning_rate": 1.4623955431754876e-05,
|
| 679 |
-
"loss": 0.
|
| 680 |
"step": 1120
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 2.83,
|
| 684 |
"learning_rate": 1.445682451253482e-05,
|
| 685 |
-
"loss": 0.
|
| 686 |
"step": 1130
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 2.86,
|
| 690 |
"learning_rate": 1.4289693593314764e-05,
|
| 691 |
-
"loss": 0.
|
| 692 |
"step": 1140
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 2.88,
|
| 696 |
"learning_rate": 1.4122562674094708e-05,
|
| 697 |
-
"loss": 0.
|
| 698 |
"step": 1150
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 2.91,
|
| 702 |
"learning_rate": 1.3955431754874653e-05,
|
| 703 |
-
"loss": 0.
|
| 704 |
"step": 1160
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 2.93,
|
| 708 |
"learning_rate": 1.3788300835654596e-05,
|
| 709 |
-
"loss": 0.
|
| 710 |
"step": 1170
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 2.96,
|
| 714 |
"learning_rate": 1.362116991643454e-05,
|
| 715 |
-
"loss": 0.
|
| 716 |
"step": 1180
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 2.98,
|
| 720 |
"learning_rate": 1.3454038997214485e-05,
|
| 721 |
-
"loss": 0.
|
| 722 |
"step": 1190
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 3.01,
|
| 726 |
"learning_rate": 1.3286908077994428e-05,
|
| 727 |
-
"loss": 0.
|
| 728 |
"step": 1200
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 3.03,
|
| 732 |
"learning_rate": 1.3119777158774374e-05,
|
| 733 |
-
"loss": 0.
|
| 734 |
"step": 1210
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 3.06,
|
| 738 |
"learning_rate": 1.2952646239554317e-05,
|
| 739 |
-
"loss": 0.
|
| 740 |
"step": 1220
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 3.08,
|
| 744 |
"learning_rate": 1.2785515320334262e-05,
|
| 745 |
-
"loss": 0.
|
| 746 |
"step": 1230
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 3.11,
|
| 750 |
"learning_rate": 1.2618384401114206e-05,
|
| 751 |
-
"loss": 0.
|
| 752 |
"step": 1240
|
| 753 |
},
|
| 754 |
{
|
| 755 |
"epoch": 3.13,
|
| 756 |
"learning_rate": 1.245125348189415e-05,
|
| 757 |
-
"loss": 0.
|
| 758 |
"step": 1250
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 3.16,
|
| 762 |
"learning_rate": 1.2284122562674095e-05,
|
| 763 |
-
"loss": 0.
|
| 764 |
"step": 1260
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 3.18,
|
| 768 |
"learning_rate": 1.211699164345404e-05,
|
| 769 |
-
"loss": 0.
|
| 770 |
"step": 1270
|
| 771 |
},
|
| 772 |
{
|
| 773 |
"epoch": 3.21,
|
| 774 |
"learning_rate": 1.1949860724233983e-05,
|
| 775 |
-
"loss": 0.
|
| 776 |
"step": 1280
|
| 777 |
},
|
| 778 |
{
|
| 779 |
"epoch": 3.23,
|
| 780 |
"learning_rate": 1.1782729805013929e-05,
|
| 781 |
-
"loss": 0.
|
| 782 |
"step": 1290
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 3.26,
|
| 786 |
"learning_rate": 1.1615598885793872e-05,
|
| 787 |
-
"loss": 0.
|
| 788 |
"step": 1300
|
| 789 |
},
|
| 790 |
{
|
| 791 |
"epoch": 3.28,
|
| 792 |
"learning_rate": 1.1448467966573816e-05,
|
| 793 |
-
"loss": 0.
|
| 794 |
"step": 1310
|
| 795 |
},
|
| 796 |
{
|
| 797 |
"epoch": 3.31,
|
| 798 |
"learning_rate": 1.1281337047353761e-05,
|
| 799 |
-
"loss": 0.
|
| 800 |
"step": 1320
|
| 801 |
},
|
| 802 |
{
|
|
@@ -808,367 +808,367 @@
|
|
| 808 |
{
|
| 809 |
"epoch": 3.36,
|
| 810 |
"learning_rate": 1.0947075208913648e-05,
|
| 811 |
-
"loss": 0.
|
| 812 |
"step": 1340
|
| 813 |
},
|
| 814 |
{
|
| 815 |
"epoch": 3.38,
|
| 816 |
"learning_rate": 1.0779944289693595e-05,
|
| 817 |
-
"loss": 0.
|
| 818 |
"step": 1350
|
| 819 |
},
|
| 820 |
{
|
| 821 |
"epoch": 3.41,
|
| 822 |
"learning_rate": 1.0612813370473537e-05,
|
| 823 |
-
"loss": 0.
|
| 824 |
"step": 1360
|
| 825 |
},
|
| 826 |
{
|
| 827 |
"epoch": 3.43,
|
| 828 |
"learning_rate": 1.0445682451253482e-05,
|
| 829 |
-
"loss": 0.
|
| 830 |
"step": 1370
|
| 831 |
},
|
| 832 |
{
|
| 833 |
"epoch": 3.46,
|
| 834 |
"learning_rate": 1.0278551532033427e-05,
|
| 835 |
-
"loss": 0.
|
| 836 |
"step": 1380
|
| 837 |
},
|
| 838 |
{
|
| 839 |
"epoch": 3.48,
|
| 840 |
"learning_rate": 1.0111420612813371e-05,
|
| 841 |
-
"loss": 0.
|
| 842 |
"step": 1390
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 3.51,
|
| 846 |
"learning_rate": 9.944289693593314e-06,
|
| 847 |
-
"loss": 0.
|
| 848 |
"step": 1400
|
| 849 |
},
|
| 850 |
{
|
| 851 |
"epoch": 3.53,
|
| 852 |
"learning_rate": 9.77715877437326e-06,
|
| 853 |
-
"loss": 0.
|
| 854 |
"step": 1410
|
| 855 |
},
|
| 856 |
{
|
| 857 |
"epoch": 3.56,
|
| 858 |
"learning_rate": 9.610027855153203e-06,
|
| 859 |
-
"loss": 0.
|
| 860 |
"step": 1420
|
| 861 |
},
|
| 862 |
{
|
| 863 |
"epoch": 3.58,
|
| 864 |
"learning_rate": 9.44289693593315e-06,
|
| 865 |
-
"loss": 0.
|
| 866 |
"step": 1430
|
| 867 |
},
|
| 868 |
{
|
| 869 |
"epoch": 3.61,
|
| 870 |
"learning_rate": 9.275766016713092e-06,
|
| 871 |
-
"loss": 0.
|
| 872 |
"step": 1440
|
| 873 |
},
|
| 874 |
{
|
| 875 |
"epoch": 3.63,
|
| 876 |
"learning_rate": 9.108635097493037e-06,
|
| 877 |
-
"loss": 0.
|
| 878 |
"step": 1450
|
| 879 |
},
|
| 880 |
{
|
| 881 |
"epoch": 3.66,
|
| 882 |
"learning_rate": 8.941504178272981e-06,
|
| 883 |
-
"loss": 0.
|
| 884 |
"step": 1460
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 3.68,
|
| 888 |
"learning_rate": 8.774373259052926e-06,
|
| 889 |
-
"loss": 0.
|
| 890 |
"step": 1470
|
| 891 |
},
|
| 892 |
{
|
| 893 |
"epoch": 3.71,
|
| 894 |
"learning_rate": 8.607242339832869e-06,
|
| 895 |
-
"loss": 0.
|
| 896 |
"step": 1480
|
| 897 |
},
|
| 898 |
{
|
| 899 |
"epoch": 3.73,
|
| 900 |
"learning_rate": 8.440111420612815e-06,
|
| 901 |
-
"loss": 0.
|
| 902 |
"step": 1490
|
| 903 |
},
|
| 904 |
{
|
| 905 |
"epoch": 3.76,
|
| 906 |
"learning_rate": 8.272980501392758e-06,
|
| 907 |
-
"loss": 0.
|
| 908 |
"step": 1500
|
| 909 |
},
|
| 910 |
{
|
| 911 |
"epoch": 3.78,
|
| 912 |
"learning_rate": 8.1058495821727e-06,
|
| 913 |
-
"loss": 0.
|
| 914 |
"step": 1510
|
| 915 |
},
|
| 916 |
{
|
| 917 |
"epoch": 3.81,
|
| 918 |
"learning_rate": 7.938718662952647e-06,
|
| 919 |
-
"loss": 0.
|
| 920 |
"step": 1520
|
| 921 |
},
|
| 922 |
{
|
| 923 |
"epoch": 3.83,
|
| 924 |
"learning_rate": 7.77158774373259e-06,
|
| 925 |
-
"loss": 0.
|
| 926 |
"step": 1530
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 3.86,
|
| 930 |
"learning_rate": 7.604456824512535e-06,
|
| 931 |
-
"loss": 0.
|
| 932 |
"step": 1540
|
| 933 |
},
|
| 934 |
{
|
| 935 |
"epoch": 3.88,
|
| 936 |
"learning_rate": 7.43732590529248e-06,
|
| 937 |
-
"loss": 0.
|
| 938 |
"step": 1550
|
| 939 |
},
|
| 940 |
{
|
| 941 |
"epoch": 3.91,
|
| 942 |
"learning_rate": 7.2701949860724235e-06,
|
| 943 |
-
"loss": 0.
|
| 944 |
"step": 1560
|
| 945 |
},
|
| 946 |
{
|
| 947 |
"epoch": 3.93,
|
| 948 |
"learning_rate": 7.103064066852368e-06,
|
| 949 |
-
"loss": 0.
|
| 950 |
"step": 1570
|
| 951 |
},
|
| 952 |
{
|
| 953 |
"epoch": 3.96,
|
| 954 |
"learning_rate": 6.935933147632313e-06,
|
| 955 |
-
"loss": 0.
|
| 956 |
"step": 1580
|
| 957 |
},
|
| 958 |
{
|
| 959 |
"epoch": 3.98,
|
| 960 |
"learning_rate": 6.768802228412256e-06,
|
| 961 |
-
"loss": 0.
|
| 962 |
"step": 1590
|
| 963 |
},
|
| 964 |
{
|
| 965 |
"epoch": 4.01,
|
| 966 |
"learning_rate": 6.601671309192201e-06,
|
| 967 |
-
"loss": 0.
|
| 968 |
"step": 1600
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 4.04,
|
| 972 |
"learning_rate": 6.4345403899721455e-06,
|
| 973 |
-
"loss": 0.
|
| 974 |
"step": 1610
|
| 975 |
},
|
| 976 |
{
|
| 977 |
"epoch": 4.06,
|
| 978 |
"learning_rate": 6.267409470752089e-06,
|
| 979 |
-
"loss": 0.
|
| 980 |
"step": 1620
|
| 981 |
},
|
| 982 |
{
|
| 983 |
"epoch": 4.09,
|
| 984 |
"learning_rate": 6.100278551532034e-06,
|
| 985 |
-
"loss": 0.
|
| 986 |
"step": 1630
|
| 987 |
},
|
| 988 |
{
|
| 989 |
"epoch": 4.11,
|
| 990 |
"learning_rate": 5.933147632311978e-06,
|
| 991 |
-
"loss": 0.
|
| 992 |
"step": 1640
|
| 993 |
},
|
| 994 |
{
|
| 995 |
"epoch": 4.14,
|
| 996 |
"learning_rate": 5.766016713091923e-06,
|
| 997 |
-
"loss": 0.
|
| 998 |
"step": 1650
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
"epoch": 4.16,
|
| 1002 |
"learning_rate": 5.598885793871867e-06,
|
| 1003 |
-
"loss": 0.
|
| 1004 |
"step": 1660
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
"epoch": 4.19,
|
| 1008 |
"learning_rate": 5.43175487465181e-06,
|
| 1009 |
-
"loss": 0.
|
| 1010 |
"step": 1670
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 4.21,
|
| 1014 |
"learning_rate": 5.264623955431755e-06,
|
| 1015 |
-
"loss": 0.
|
| 1016 |
"step": 1680
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
"epoch": 4.24,
|
| 1020 |
"learning_rate": 5.0974930362116986e-06,
|
| 1021 |
-
"loss": 0.
|
| 1022 |
"step": 1690
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
"epoch": 4.26,
|
| 1026 |
"learning_rate": 4.930362116991643e-06,
|
| 1027 |
-
"loss": 0.
|
| 1028 |
"step": 1700
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 4.29,
|
| 1032 |
"learning_rate": 4.763231197771588e-06,
|
| 1033 |
-
"loss": 0.
|
| 1034 |
"step": 1710
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
"epoch": 4.31,
|
| 1038 |
"learning_rate": 4.596100278551532e-06,
|
| 1039 |
-
"loss": 0.
|
| 1040 |
"step": 1720
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
"epoch": 4.34,
|
| 1044 |
"learning_rate": 4.428969359331476e-06,
|
| 1045 |
-
"loss": 0.
|
| 1046 |
"step": 1730
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 4.36,
|
| 1050 |
"learning_rate": 4.2618384401114205e-06,
|
| 1051 |
-
"loss": 0.
|
| 1052 |
"step": 1740
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 4.39,
|
| 1056 |
"learning_rate": 4.094707520891365e-06,
|
| 1057 |
-
"loss": 0.
|
| 1058 |
"step": 1750
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
"epoch": 4.41,
|
| 1062 |
"learning_rate": 3.927576601671309e-06,
|
| 1063 |
-
"loss": 0.
|
| 1064 |
"step": 1760
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 4.44,
|
| 1068 |
"learning_rate": 3.7604456824512533e-06,
|
| 1069 |
-
"loss": 0.
|
| 1070 |
"step": 1770
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
"epoch": 4.46,
|
| 1074 |
"learning_rate": 3.593314763231198e-06,
|
| 1075 |
-
"loss": 0.
|
| 1076 |
"step": 1780
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
"epoch": 4.49,
|
| 1080 |
"learning_rate": 3.426183844011142e-06,
|
| 1081 |
-
"loss": 0.
|
| 1082 |
"step": 1790
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 4.51,
|
| 1086 |
"learning_rate": 3.259052924791086e-06,
|
| 1087 |
-
"loss": 0.
|
| 1088 |
"step": 1800
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
"epoch": 4.54,
|
| 1092 |
"learning_rate": 3.0919220055710307e-06,
|
| 1093 |
-
"loss": 0.
|
| 1094 |
"step": 1810
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 4.56,
|
| 1098 |
"learning_rate": 2.924791086350975e-06,
|
| 1099 |
-
"loss": 0.
|
| 1100 |
"step": 1820
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 4.59,
|
| 1104 |
"learning_rate": 2.7576601671309194e-06,
|
| 1105 |
-
"loss": 0.
|
| 1106 |
"step": 1830
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
"epoch": 4.61,
|
| 1110 |
"learning_rate": 2.5905292479108636e-06,
|
| 1111 |
-
"loss": 0.
|
| 1112 |
"step": 1840
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
"epoch": 4.64,
|
| 1116 |
"learning_rate": 2.4233983286908077e-06,
|
| 1117 |
-
"loss": 0.
|
| 1118 |
"step": 1850
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
"epoch": 4.66,
|
| 1122 |
"learning_rate": 2.2562674094707523e-06,
|
| 1123 |
-
"loss": 0.
|
| 1124 |
"step": 1860
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
"epoch": 4.69,
|
| 1128 |
"learning_rate": 2.0891364902506964e-06,
|
| 1129 |
-
"loss": 0.
|
| 1130 |
"step": 1870
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
"epoch": 4.71,
|
| 1134 |
"learning_rate": 1.922005571030641e-06,
|
| 1135 |
-
"loss": 0.
|
| 1136 |
"step": 1880
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 4.74,
|
| 1140 |
"learning_rate": 1.7548746518105849e-06,
|
| 1141 |
-
"loss": 0.
|
| 1142 |
"step": 1890
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
"epoch": 4.76,
|
| 1146 |
"learning_rate": 1.5877437325905292e-06,
|
| 1147 |
-
"loss": 0.
|
| 1148 |
"step": 1900
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
"epoch": 4.79,
|
| 1152 |
"learning_rate": 1.4206128133704736e-06,
|
| 1153 |
-
"loss": 0.
|
| 1154 |
"step": 1910
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
"epoch": 4.81,
|
| 1158 |
"learning_rate": 1.253481894150418e-06,
|
| 1159 |
-
"loss": 0.
|
| 1160 |
"step": 1920
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 4.84,
|
| 1164 |
"learning_rate": 1.0863509749303623e-06,
|
| 1165 |
-
"loss": 0.
|
| 1166 |
"step": 1930
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
"epoch": 4.86,
|
| 1170 |
"learning_rate": 9.192200557103064e-07,
|
| 1171 |
-
"loss": 0.
|
| 1172 |
"step": 1940
|
| 1173 |
},
|
| 1174 |
{
|
|
@@ -1186,29 +1186,29 @@
|
|
| 1186 |
{
|
| 1187 |
"epoch": 4.94,
|
| 1188 |
"learning_rate": 4.178272980501393e-07,
|
| 1189 |
-
"loss": 0.
|
| 1190 |
"step": 1970
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
"epoch": 4.96,
|
| 1194 |
"learning_rate": 2.506963788300836e-07,
|
| 1195 |
-
"loss": 0.
|
| 1196 |
"step": 1980
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
"epoch": 4.99,
|
| 1200 |
"learning_rate": 8.356545961002785e-08,
|
| 1201 |
-
"loss": 0.
|
| 1202 |
"step": 1990
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
"epoch": 5.0,
|
| 1206 |
"step": 1995,
|
| 1207 |
"total_flos": 1.4837528359600128e+20,
|
| 1208 |
-
"train_loss": 0.
|
| 1209 |
-
"train_runtime":
|
| 1210 |
-
"train_samples_per_second":
|
| 1211 |
-
"train_steps_per_second":
|
| 1212 |
}
|
| 1213 |
],
|
| 1214 |
"max_steps": 1995,
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 0.03,
|
| 12 |
"learning_rate": 1.5e-06,
|
| 13 |
+
"loss": 2.4689,
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 0.05,
|
| 18 |
"learning_rate": 3e-06,
|
| 19 |
+
"loss": 2.4537,
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
"epoch": 0.08,
|
| 24 |
"learning_rate": 4.5e-06,
|
| 25 |
+
"loss": 2.3789,
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 0.1,
|
| 30 |
"learning_rate": 6e-06,
|
| 31 |
+
"loss": 2.2205,
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
"epoch": 0.13,
|
| 36 |
"learning_rate": 7.5e-06,
|
| 37 |
+
"loss": 2.0825,
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
"epoch": 0.15,
|
| 42 |
"learning_rate": 9e-06,
|
| 43 |
+
"loss": 1.9133,
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.18,
|
| 48 |
"learning_rate": 1.05e-05,
|
| 49 |
+
"loss": 1.8169,
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
"epoch": 0.2,
|
| 54 |
"learning_rate": 1.2e-05,
|
| 55 |
+
"loss": 1.8079,
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
"epoch": 0.23,
|
| 60 |
"learning_rate": 1.3500000000000001e-05,
|
| 61 |
+
"loss": 1.5825,
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
|
|
|
| 70 |
{
|
| 71 |
"epoch": 0.28,
|
| 72 |
"learning_rate": 1.65e-05,
|
| 73 |
+
"loss": 1.5597,
|
| 74 |
"step": 110
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 0.3,
|
| 78 |
"learning_rate": 1.8e-05,
|
| 79 |
+
"loss": 1.6068,
|
| 80 |
"step": 120
|
| 81 |
},
|
| 82 |
{
|
| 83 |
"epoch": 0.33,
|
| 84 |
"learning_rate": 1.95e-05,
|
| 85 |
+
"loss": 1.3703,
|
| 86 |
"step": 130
|
| 87 |
},
|
| 88 |
{
|
|
|
|
| 94 |
{
|
| 95 |
"epoch": 0.38,
|
| 96 |
"learning_rate": 2.25e-05,
|
| 97 |
+
"loss": 1.4707,
|
| 98 |
"step": 150
|
| 99 |
},
|
| 100 |
{
|
| 101 |
"epoch": 0.4,
|
| 102 |
"learning_rate": 2.4e-05,
|
| 103 |
+
"loss": 1.5798,
|
| 104 |
"step": 160
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 0.43,
|
| 108 |
"learning_rate": 2.55e-05,
|
| 109 |
+
"loss": 1.4417,
|
| 110 |
"step": 170
|
| 111 |
},
|
| 112 |
{
|
| 113 |
"epoch": 0.45,
|
| 114 |
"learning_rate": 2.7000000000000002e-05,
|
| 115 |
+
"loss": 1.3562,
|
| 116 |
"step": 180
|
| 117 |
},
|
| 118 |
{
|
| 119 |
"epoch": 0.48,
|
| 120 |
"learning_rate": 2.8499999999999998e-05,
|
| 121 |
+
"loss": 1.4646,
|
| 122 |
"step": 190
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 0.5,
|
| 126 |
"learning_rate": 3e-05,
|
| 127 |
+
"loss": 1.1936,
|
| 128 |
"step": 200
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.53,
|
| 132 |
"learning_rate": 2.9832869080779945e-05,
|
| 133 |
+
"loss": 1.3362,
|
| 134 |
"step": 210
|
| 135 |
},
|
| 136 |
{
|
| 137 |
"epoch": 0.55,
|
| 138 |
"learning_rate": 2.9665738161559886e-05,
|
| 139 |
+
"loss": 1.1757,
|
| 140 |
"step": 220
|
| 141 |
},
|
| 142 |
{
|
| 143 |
"epoch": 0.58,
|
| 144 |
"learning_rate": 2.9498607242339834e-05,
|
| 145 |
+
"loss": 1.1686,
|
| 146 |
"step": 230
|
| 147 |
},
|
| 148 |
{
|
| 149 |
"epoch": 0.6,
|
| 150 |
"learning_rate": 2.933147632311978e-05,
|
| 151 |
+
"loss": 1.147,
|
| 152 |
"step": 240
|
| 153 |
},
|
| 154 |
{
|
| 155 |
"epoch": 0.63,
|
| 156 |
"learning_rate": 2.916434540389972e-05,
|
| 157 |
+
"loss": 1.0784,
|
| 158 |
"step": 250
|
| 159 |
},
|
| 160 |
{
|
| 161 |
"epoch": 0.65,
|
| 162 |
"learning_rate": 2.8997214484679665e-05,
|
| 163 |
+
"loss": 1.0562,
|
| 164 |
"step": 260
|
| 165 |
},
|
| 166 |
{
|
| 167 |
"epoch": 0.68,
|
| 168 |
"learning_rate": 2.8830083565459613e-05,
|
| 169 |
+
"loss": 0.9743,
|
| 170 |
"step": 270
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.7,
|
| 174 |
"learning_rate": 2.8662952646239554e-05,
|
| 175 |
+
"loss": 0.7418,
|
| 176 |
"step": 280
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 0.73,
|
| 180 |
"learning_rate": 2.84958217270195e-05,
|
| 181 |
+
"loss": 0.814,
|
| 182 |
"step": 290
|
| 183 |
},
|
| 184 |
{
|
| 185 |
"epoch": 0.75,
|
| 186 |
"learning_rate": 2.8328690807799443e-05,
|
| 187 |
+
"loss": 0.9595,
|
| 188 |
"step": 300
|
| 189 |
},
|
| 190 |
{
|
| 191 |
"epoch": 0.78,
|
| 192 |
"learning_rate": 2.8161559888579388e-05,
|
| 193 |
+
"loss": 0.7455,
|
| 194 |
"step": 310
|
| 195 |
},
|
| 196 |
{
|
| 197 |
"epoch": 0.8,
|
| 198 |
"learning_rate": 2.7994428969359332e-05,
|
| 199 |
+
"loss": 0.616,
|
| 200 |
"step": 320
|
| 201 |
},
|
| 202 |
{
|
| 203 |
"epoch": 0.83,
|
| 204 |
"learning_rate": 2.7827298050139277e-05,
|
| 205 |
+
"loss": 0.6448,
|
| 206 |
"step": 330
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 0.85,
|
| 210 |
"learning_rate": 2.7660167130919218e-05,
|
| 211 |
+
"loss": 0.5373,
|
| 212 |
"step": 340
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.88,
|
| 216 |
"learning_rate": 2.7493036211699166e-05,
|
| 217 |
+
"loss": 0.5143,
|
| 218 |
"step": 350
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 0.9,
|
| 222 |
"learning_rate": 2.732590529247911e-05,
|
| 223 |
+
"loss": 0.5485,
|
| 224 |
"step": 360
|
| 225 |
},
|
| 226 |
{
|
| 227 |
"epoch": 0.93,
|
| 228 |
"learning_rate": 2.7158774373259055e-05,
|
| 229 |
+
"loss": 0.3994,
|
| 230 |
"step": 370
|
| 231 |
},
|
| 232 |
{
|
| 233 |
"epoch": 0.95,
|
| 234 |
"learning_rate": 2.6991643454038996e-05,
|
| 235 |
+
"loss": 0.4221,
|
| 236 |
"step": 380
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 0.98,
|
| 240 |
"learning_rate": 2.6824512534818944e-05,
|
| 241 |
+
"loss": 0.3865,
|
| 242 |
"step": 390
|
| 243 |
},
|
| 244 |
{
|
| 245 |
"epoch": 1.0,
|
| 246 |
"learning_rate": 2.665738161559889e-05,
|
| 247 |
+
"loss": 0.3756,
|
| 248 |
"step": 400
|
| 249 |
},
|
| 250 |
{
|
| 251 |
"epoch": 1.03,
|
| 252 |
"learning_rate": 2.649025069637883e-05,
|
| 253 |
+
"loss": 0.2686,
|
| 254 |
"step": 410
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 1.05,
|
| 258 |
"learning_rate": 2.6323119777158774e-05,
|
| 259 |
+
"loss": 0.2645,
|
| 260 |
"step": 420
|
| 261 |
},
|
| 262 |
{
|
| 263 |
"epoch": 1.08,
|
| 264 |
"learning_rate": 2.6155988857938722e-05,
|
| 265 |
+
"loss": 0.2747,
|
| 266 |
"step": 430
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 1.1,
|
| 270 |
"learning_rate": 2.5988857938718663e-05,
|
| 271 |
+
"loss": 0.3327,
|
| 272 |
"step": 440
|
| 273 |
},
|
| 274 |
{
|
| 275 |
"epoch": 1.13,
|
| 276 |
"learning_rate": 2.5821727019498608e-05,
|
| 277 |
+
"loss": 0.2937,
|
| 278 |
"step": 450
|
| 279 |
},
|
| 280 |
{
|
| 281 |
"epoch": 1.15,
|
| 282 |
"learning_rate": 2.5654596100278553e-05,
|
| 283 |
+
"loss": 0.2593,
|
| 284 |
"step": 460
|
| 285 |
},
|
| 286 |
{
|
| 287 |
"epoch": 1.18,
|
| 288 |
"learning_rate": 2.5487465181058497e-05,
|
| 289 |
+
"loss": 0.1705,
|
| 290 |
"step": 470
|
| 291 |
},
|
| 292 |
{
|
| 293 |
"epoch": 1.2,
|
| 294 |
"learning_rate": 2.532033426183844e-05,
|
| 295 |
+
"loss": 0.1641,
|
| 296 |
"step": 480
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 1.23,
|
| 300 |
"learning_rate": 2.5153203342618386e-05,
|
| 301 |
+
"loss": 0.3088,
|
| 302 |
"step": 490
|
| 303 |
},
|
| 304 |
{
|
| 305 |
"epoch": 1.25,
|
| 306 |
"learning_rate": 2.4986072423398327e-05,
|
| 307 |
+
"loss": 0.2023,
|
| 308 |
"step": 500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 1.28,
|
| 312 |
"learning_rate": 2.4818941504178275e-05,
|
| 313 |
+
"loss": 0.1896,
|
| 314 |
"step": 510
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 1.3,
|
| 318 |
"learning_rate": 2.465181058495822e-05,
|
| 319 |
+
"loss": 0.1476,
|
| 320 |
"step": 520
|
| 321 |
},
|
| 322 |
{
|
| 323 |
"epoch": 1.33,
|
| 324 |
"learning_rate": 2.448467966573816e-05,
|
| 325 |
+
"loss": 0.2023,
|
| 326 |
"step": 530
|
| 327 |
},
|
| 328 |
{
|
| 329 |
"epoch": 1.35,
|
| 330 |
"learning_rate": 2.4317548746518106e-05,
|
| 331 |
+
"loss": 0.1005,
|
| 332 |
"step": 540
|
| 333 |
},
|
| 334 |
{
|
| 335 |
"epoch": 1.38,
|
| 336 |
"learning_rate": 2.415041782729805e-05,
|
| 337 |
+
"loss": 0.2535,
|
| 338 |
"step": 550
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 1.4,
|
| 342 |
"learning_rate": 2.3983286908077995e-05,
|
| 343 |
+
"loss": 0.1387,
|
| 344 |
"step": 560
|
| 345 |
},
|
| 346 |
{
|
| 347 |
"epoch": 1.43,
|
| 348 |
"learning_rate": 2.381615598885794e-05,
|
| 349 |
+
"loss": 0.2567,
|
| 350 |
"step": 570
|
| 351 |
},
|
| 352 |
{
|
| 353 |
"epoch": 1.45,
|
| 354 |
"learning_rate": 2.3649025069637884e-05,
|
| 355 |
+
"loss": 0.0976,
|
| 356 |
"step": 580
|
| 357 |
},
|
| 358 |
{
|
| 359 |
"epoch": 1.48,
|
| 360 |
"learning_rate": 2.3481894150417825e-05,
|
| 361 |
+
"loss": 0.1436,
|
| 362 |
"step": 590
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 1.5,
|
| 366 |
"learning_rate": 2.3314763231197773e-05,
|
| 367 |
+
"loss": 0.0673,
|
| 368 |
"step": 600
|
| 369 |
},
|
| 370 |
{
|
| 371 |
"epoch": 1.53,
|
| 372 |
"learning_rate": 2.3147632311977718e-05,
|
| 373 |
+
"loss": 0.1552,
|
| 374 |
"step": 610
|
| 375 |
},
|
| 376 |
{
|
| 377 |
"epoch": 1.55,
|
| 378 |
"learning_rate": 2.298050139275766e-05,
|
| 379 |
+
"loss": 0.1199,
|
| 380 |
"step": 620
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 1.58,
|
| 384 |
"learning_rate": 2.2813370473537603e-05,
|
| 385 |
+
"loss": 0.1526,
|
| 386 |
"step": 630
|
| 387 |
},
|
| 388 |
{
|
| 389 |
"epoch": 1.6,
|
| 390 |
"learning_rate": 2.264623955431755e-05,
|
| 391 |
+
"loss": 0.065,
|
| 392 |
"step": 640
|
| 393 |
},
|
| 394 |
{
|
| 395 |
"epoch": 1.63,
|
| 396 |
"learning_rate": 2.2479108635097492e-05,
|
| 397 |
+
"loss": 0.1841,
|
| 398 |
"step": 650
|
| 399 |
},
|
| 400 |
{
|
| 401 |
"epoch": 1.65,
|
| 402 |
"learning_rate": 2.2311977715877437e-05,
|
| 403 |
+
"loss": 0.3076,
|
| 404 |
"step": 660
|
| 405 |
},
|
| 406 |
{
|
| 407 |
"epoch": 1.68,
|
| 408 |
"learning_rate": 2.214484679665738e-05,
|
| 409 |
+
"loss": 0.0943,
|
| 410 |
"step": 670
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 1.7,
|
| 414 |
"learning_rate": 2.1977715877437326e-05,
|
| 415 |
+
"loss": 0.0507,
|
| 416 |
"step": 680
|
| 417 |
},
|
| 418 |
{
|
| 419 |
"epoch": 1.73,
|
| 420 |
"learning_rate": 2.181058495821727e-05,
|
| 421 |
+
"loss": 0.0748,
|
| 422 |
"step": 690
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 1.75,
|
| 426 |
"learning_rate": 2.1643454038997215e-05,
|
| 427 |
+
"loss": 0.0612,
|
| 428 |
"step": 700
|
| 429 |
},
|
| 430 |
{
|
| 431 |
"epoch": 1.78,
|
| 432 |
"learning_rate": 2.147632311977716e-05,
|
| 433 |
+
"loss": 0.1332,
|
| 434 |
"step": 710
|
| 435 |
},
|
| 436 |
{
|
| 437 |
"epoch": 1.8,
|
| 438 |
"learning_rate": 2.1309192200557104e-05,
|
| 439 |
+
"loss": 0.1505,
|
| 440 |
"step": 720
|
| 441 |
},
|
| 442 |
{
|
| 443 |
"epoch": 1.83,
|
| 444 |
"learning_rate": 2.114206128133705e-05,
|
| 445 |
+
"loss": 0.0881,
|
| 446 |
"step": 730
|
| 447 |
},
|
| 448 |
{
|
| 449 |
"epoch": 1.85,
|
| 450 |
"learning_rate": 2.0974930362116993e-05,
|
| 451 |
+
"loss": 0.1235,
|
| 452 |
"step": 740
|
| 453 |
},
|
| 454 |
{
|
| 455 |
"epoch": 1.88,
|
| 456 |
"learning_rate": 2.0807799442896935e-05,
|
| 457 |
+
"loss": 0.0482,
|
| 458 |
"step": 750
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 1.9,
|
| 462 |
"learning_rate": 2.0640668523676883e-05,
|
| 463 |
+
"loss": 0.1344,
|
| 464 |
"step": 760
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 1.93,
|
| 468 |
"learning_rate": 2.0473537604456827e-05,
|
| 469 |
+
"loss": 0.05,
|
| 470 |
"step": 770
|
| 471 |
},
|
| 472 |
{
|
| 473 |
"epoch": 1.95,
|
| 474 |
"learning_rate": 2.0306406685236768e-05,
|
| 475 |
+
"loss": 0.0645,
|
| 476 |
"step": 780
|
| 477 |
},
|
| 478 |
{
|
| 479 |
"epoch": 1.98,
|
| 480 |
"learning_rate": 2.0139275766016713e-05,
|
| 481 |
+
"loss": 0.0952,
|
| 482 |
"step": 790
|
| 483 |
},
|
| 484 |
{
|
| 485 |
"epoch": 2.01,
|
| 486 |
"learning_rate": 1.997214484679666e-05,
|
| 487 |
+
"loss": 0.0444,
|
| 488 |
"step": 800
|
| 489 |
},
|
| 490 |
{
|
| 491 |
"epoch": 2.03,
|
| 492 |
"learning_rate": 1.9805013927576602e-05,
|
| 493 |
+
"loss": 0.0892,
|
| 494 |
"step": 810
|
| 495 |
},
|
| 496 |
{
|
| 497 |
"epoch": 2.06,
|
| 498 |
"learning_rate": 1.9637883008356546e-05,
|
| 499 |
+
"loss": 0.0361,
|
| 500 |
"step": 820
|
| 501 |
},
|
| 502 |
{
|
| 503 |
"epoch": 2.08,
|
| 504 |
"learning_rate": 1.947075208913649e-05,
|
| 505 |
+
"loss": 0.1341,
|
| 506 |
"step": 830
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 2.11,
|
| 510 |
"learning_rate": 1.9303621169916436e-05,
|
| 511 |
+
"loss": 0.101,
|
| 512 |
"step": 840
|
| 513 |
},
|
| 514 |
{
|
| 515 |
"epoch": 2.13,
|
| 516 |
"learning_rate": 1.913649025069638e-05,
|
| 517 |
+
"loss": 0.0945,
|
| 518 |
"step": 850
|
| 519 |
},
|
| 520 |
{
|
| 521 |
"epoch": 2.16,
|
| 522 |
"learning_rate": 1.8969359331476325e-05,
|
| 523 |
+
"loss": 0.0405,
|
| 524 |
"step": 860
|
| 525 |
},
|
| 526 |
{
|
| 527 |
"epoch": 2.18,
|
| 528 |
"learning_rate": 1.8802228412256266e-05,
|
| 529 |
+
"loss": 0.1336,
|
| 530 |
"step": 870
|
| 531 |
},
|
| 532 |
{
|
| 533 |
"epoch": 2.21,
|
| 534 |
"learning_rate": 1.863509749303621e-05,
|
| 535 |
+
"loss": 0.0359,
|
| 536 |
"step": 880
|
| 537 |
},
|
| 538 |
{
|
| 539 |
"epoch": 2.23,
|
| 540 |
"learning_rate": 1.846796657381616e-05,
|
| 541 |
+
"loss": 0.0371,
|
| 542 |
"step": 890
|
| 543 |
},
|
| 544 |
{
|
| 545 |
"epoch": 2.26,
|
| 546 |
"learning_rate": 1.83008356545961e-05,
|
| 547 |
+
"loss": 0.09,
|
| 548 |
"step": 900
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 2.28,
|
| 552 |
"learning_rate": 1.8133704735376044e-05,
|
| 553 |
+
"loss": 0.079,
|
| 554 |
"step": 910
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 2.31,
|
| 558 |
"learning_rate": 1.796657381615599e-05,
|
| 559 |
+
"loss": 0.04,
|
| 560 |
"step": 920
|
| 561 |
},
|
| 562 |
{
|
| 563 |
"epoch": 2.33,
|
| 564 |
"learning_rate": 1.7799442896935933e-05,
|
| 565 |
+
"loss": 0.0835,
|
| 566 |
"step": 930
|
| 567 |
},
|
| 568 |
{
|
| 569 |
"epoch": 2.36,
|
| 570 |
"learning_rate": 1.7632311977715878e-05,
|
| 571 |
+
"loss": 0.0233,
|
| 572 |
"step": 940
|
| 573 |
},
|
| 574 |
{
|
| 575 |
"epoch": 2.38,
|
| 576 |
"learning_rate": 1.7465181058495822e-05,
|
| 577 |
+
"loss": 0.0377,
|
| 578 |
"step": 950
|
| 579 |
},
|
| 580 |
{
|
| 581 |
"epoch": 2.41,
|
| 582 |
"learning_rate": 1.7298050139275764e-05,
|
| 583 |
+
"loss": 0.0422,
|
| 584 |
"step": 960
|
| 585 |
},
|
| 586 |
{
|
| 587 |
"epoch": 2.43,
|
| 588 |
"learning_rate": 1.713091922005571e-05,
|
| 589 |
+
"loss": 0.052,
|
| 590 |
"step": 970
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 2.46,
|
| 594 |
"learning_rate": 1.6963788300835656e-05,
|
| 595 |
+
"loss": 0.0272,
|
| 596 |
"step": 980
|
| 597 |
},
|
| 598 |
{
|
| 599 |
"epoch": 2.48,
|
| 600 |
"learning_rate": 1.6796657381615597e-05,
|
| 601 |
+
"loss": 0.0206,
|
| 602 |
"step": 990
|
| 603 |
},
|
| 604 |
{
|
| 605 |
"epoch": 2.51,
|
| 606 |
"learning_rate": 1.6629526462395542e-05,
|
| 607 |
+
"loss": 0.0296,
|
| 608 |
"step": 1000
|
| 609 |
},
|
| 610 |
{
|
| 611 |
"epoch": 2.53,
|
| 612 |
"learning_rate": 1.646239554317549e-05,
|
| 613 |
+
"loss": 0.0632,
|
| 614 |
"step": 1010
|
| 615 |
},
|
| 616 |
{
|
| 617 |
"epoch": 2.56,
|
| 618 |
"learning_rate": 1.6295264623955434e-05,
|
| 619 |
+
"loss": 0.0719,
|
| 620 |
"step": 1020
|
| 621 |
},
|
| 622 |
{
|
| 623 |
"epoch": 2.58,
|
| 624 |
"learning_rate": 1.6128133704735375e-05,
|
| 625 |
+
"loss": 0.0835,
|
| 626 |
"step": 1030
|
| 627 |
},
|
| 628 |
{
|
| 629 |
"epoch": 2.61,
|
| 630 |
"learning_rate": 1.596100278551532e-05,
|
| 631 |
+
"loss": 0.0492,
|
| 632 |
"step": 1040
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 2.63,
|
| 636 |
"learning_rate": 1.5793871866295268e-05,
|
| 637 |
+
"loss": 0.0275,
|
| 638 |
"step": 1050
|
| 639 |
},
|
| 640 |
{
|
| 641 |
"epoch": 2.66,
|
| 642 |
"learning_rate": 1.562674094707521e-05,
|
| 643 |
+
"loss": 0.0242,
|
| 644 |
"step": 1060
|
| 645 |
},
|
| 646 |
{
|
| 647 |
"epoch": 2.68,
|
| 648 |
"learning_rate": 1.5459610027855154e-05,
|
| 649 |
+
"loss": 0.0278,
|
| 650 |
"step": 1070
|
| 651 |
},
|
| 652 |
{
|
| 653 |
"epoch": 2.71,
|
| 654 |
"learning_rate": 1.5292479108635098e-05,
|
| 655 |
+
"loss": 0.0592,
|
| 656 |
"step": 1080
|
| 657 |
},
|
| 658 |
{
|
| 659 |
"epoch": 2.73,
|
| 660 |
"learning_rate": 1.5125348189415043e-05,
|
| 661 |
+
"loss": 0.0734,
|
| 662 |
"step": 1090
|
| 663 |
},
|
| 664 |
{
|
| 665 |
"epoch": 2.76,
|
| 666 |
"learning_rate": 1.4958217270194987e-05,
|
| 667 |
+
"loss": 0.2112,
|
| 668 |
"step": 1100
|
| 669 |
},
|
| 670 |
{
|
| 671 |
"epoch": 2.78,
|
| 672 |
"learning_rate": 1.479108635097493e-05,
|
| 673 |
+
"loss": 0.0753,
|
| 674 |
"step": 1110
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 2.81,
|
| 678 |
"learning_rate": 1.4623955431754876e-05,
|
| 679 |
+
"loss": 0.072,
|
| 680 |
"step": 1120
|
| 681 |
},
|
| 682 |
{
|
| 683 |
"epoch": 2.83,
|
| 684 |
"learning_rate": 1.445682451253482e-05,
|
| 685 |
+
"loss": 0.0146,
|
| 686 |
"step": 1130
|
| 687 |
},
|
| 688 |
{
|
| 689 |
"epoch": 2.86,
|
| 690 |
"learning_rate": 1.4289693593314764e-05,
|
| 691 |
+
"loss": 0.1671,
|
| 692 |
"step": 1140
|
| 693 |
},
|
| 694 |
{
|
| 695 |
"epoch": 2.88,
|
| 696 |
"learning_rate": 1.4122562674094708e-05,
|
| 697 |
+
"loss": 0.0701,
|
| 698 |
"step": 1150
|
| 699 |
},
|
| 700 |
{
|
| 701 |
"epoch": 2.91,
|
| 702 |
"learning_rate": 1.3955431754874653e-05,
|
| 703 |
+
"loss": 0.1404,
|
| 704 |
"step": 1160
|
| 705 |
},
|
| 706 |
{
|
| 707 |
"epoch": 2.93,
|
| 708 |
"learning_rate": 1.3788300835654596e-05,
|
| 709 |
+
"loss": 0.08,
|
| 710 |
"step": 1170
|
| 711 |
},
|
| 712 |
{
|
| 713 |
"epoch": 2.96,
|
| 714 |
"learning_rate": 1.362116991643454e-05,
|
| 715 |
+
"loss": 0.083,
|
| 716 |
"step": 1180
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 2.98,
|
| 720 |
"learning_rate": 1.3454038997214485e-05,
|
| 721 |
+
"loss": 0.0619,
|
| 722 |
"step": 1190
|
| 723 |
},
|
| 724 |
{
|
| 725 |
"epoch": 3.01,
|
| 726 |
"learning_rate": 1.3286908077994428e-05,
|
| 727 |
+
"loss": 0.0569,
|
| 728 |
"step": 1200
|
| 729 |
},
|
| 730 |
{
|
| 731 |
"epoch": 3.03,
|
| 732 |
"learning_rate": 1.3119777158774374e-05,
|
| 733 |
+
"loss": 0.0377,
|
| 734 |
"step": 1210
|
| 735 |
},
|
| 736 |
{
|
| 737 |
"epoch": 3.06,
|
| 738 |
"learning_rate": 1.2952646239554317e-05,
|
| 739 |
+
"loss": 0.1295,
|
| 740 |
"step": 1220
|
| 741 |
},
|
| 742 |
{
|
| 743 |
"epoch": 3.08,
|
| 744 |
"learning_rate": 1.2785515320334262e-05,
|
| 745 |
+
"loss": 0.1053,
|
| 746 |
"step": 1230
|
| 747 |
},
|
| 748 |
{
|
| 749 |
"epoch": 3.11,
|
| 750 |
"learning_rate": 1.2618384401114206e-05,
|
| 751 |
+
"loss": 0.0191,
|
| 752 |
"step": 1240
|
| 753 |
},
|
| 754 |
{
|
| 755 |
"epoch": 3.13,
|
| 756 |
"learning_rate": 1.245125348189415e-05,
|
| 757 |
+
"loss": 0.1647,
|
| 758 |
"step": 1250
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 3.16,
|
| 762 |
"learning_rate": 1.2284122562674095e-05,
|
| 763 |
+
"loss": 0.0731,
|
| 764 |
"step": 1260
|
| 765 |
},
|
| 766 |
{
|
| 767 |
"epoch": 3.18,
|
| 768 |
"learning_rate": 1.211699164345404e-05,
|
| 769 |
+
"loss": 0.0238,
|
| 770 |
"step": 1270
|
| 771 |
},
|
| 772 |
{
|
| 773 |
"epoch": 3.21,
|
| 774 |
"learning_rate": 1.1949860724233983e-05,
|
| 775 |
+
"loss": 0.0133,
|
| 776 |
"step": 1280
|
| 777 |
},
|
| 778 |
{
|
| 779 |
"epoch": 3.23,
|
| 780 |
"learning_rate": 1.1782729805013929e-05,
|
| 781 |
+
"loss": 0.167,
|
| 782 |
"step": 1290
|
| 783 |
},
|
| 784 |
{
|
| 785 |
"epoch": 3.26,
|
| 786 |
"learning_rate": 1.1615598885793872e-05,
|
| 787 |
+
"loss": 0.0532,
|
| 788 |
"step": 1300
|
| 789 |
},
|
| 790 |
{
|
| 791 |
"epoch": 3.28,
|
| 792 |
"learning_rate": 1.1448467966573816e-05,
|
| 793 |
+
"loss": 0.0524,
|
| 794 |
"step": 1310
|
| 795 |
},
|
| 796 |
{
|
| 797 |
"epoch": 3.31,
|
| 798 |
"learning_rate": 1.1281337047353761e-05,
|
| 799 |
+
"loss": 0.0206,
|
| 800 |
"step": 1320
|
| 801 |
},
|
| 802 |
{
|
|
|
|
| 808 |
{
|
| 809 |
"epoch": 3.36,
|
| 810 |
"learning_rate": 1.0947075208913648e-05,
|
| 811 |
+
"loss": 0.0127,
|
| 812 |
"step": 1340
|
| 813 |
},
|
| 814 |
{
|
| 815 |
"epoch": 3.38,
|
| 816 |
"learning_rate": 1.0779944289693595e-05,
|
| 817 |
+
"loss": 0.0625,
|
| 818 |
"step": 1350
|
| 819 |
},
|
| 820 |
{
|
| 821 |
"epoch": 3.41,
|
| 822 |
"learning_rate": 1.0612813370473537e-05,
|
| 823 |
+
"loss": 0.016,
|
| 824 |
"step": 1360
|
| 825 |
},
|
| 826 |
{
|
| 827 |
"epoch": 3.43,
|
| 828 |
"learning_rate": 1.0445682451253482e-05,
|
| 829 |
+
"loss": 0.0521,
|
| 830 |
"step": 1370
|
| 831 |
},
|
| 832 |
{
|
| 833 |
"epoch": 3.46,
|
| 834 |
"learning_rate": 1.0278551532033427e-05,
|
| 835 |
+
"loss": 0.0317,
|
| 836 |
"step": 1380
|
| 837 |
},
|
| 838 |
{
|
| 839 |
"epoch": 3.48,
|
| 840 |
"learning_rate": 1.0111420612813371e-05,
|
| 841 |
+
"loss": 0.0154,
|
| 842 |
"step": 1390
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 3.51,
|
| 846 |
"learning_rate": 9.944289693593314e-06,
|
| 847 |
+
"loss": 0.0907,
|
| 848 |
"step": 1400
|
| 849 |
},
|
| 850 |
{
|
| 851 |
"epoch": 3.53,
|
| 852 |
"learning_rate": 9.77715877437326e-06,
|
| 853 |
+
"loss": 0.0105,
|
| 854 |
"step": 1410
|
| 855 |
},
|
| 856 |
{
|
| 857 |
"epoch": 3.56,
|
| 858 |
"learning_rate": 9.610027855153203e-06,
|
| 859 |
+
"loss": 0.0122,
|
| 860 |
"step": 1420
|
| 861 |
},
|
| 862 |
{
|
| 863 |
"epoch": 3.58,
|
| 864 |
"learning_rate": 9.44289693593315e-06,
|
| 865 |
+
"loss": 0.0247,
|
| 866 |
"step": 1430
|
| 867 |
},
|
| 868 |
{
|
| 869 |
"epoch": 3.61,
|
| 870 |
"learning_rate": 9.275766016713092e-06,
|
| 871 |
+
"loss": 0.0598,
|
| 872 |
"step": 1440
|
| 873 |
},
|
| 874 |
{
|
| 875 |
"epoch": 3.63,
|
| 876 |
"learning_rate": 9.108635097493037e-06,
|
| 877 |
+
"loss": 0.0195,
|
| 878 |
"step": 1450
|
| 879 |
},
|
| 880 |
{
|
| 881 |
"epoch": 3.66,
|
| 882 |
"learning_rate": 8.941504178272981e-06,
|
| 883 |
+
"loss": 0.0195,
|
| 884 |
"step": 1460
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 3.68,
|
| 888 |
"learning_rate": 8.774373259052926e-06,
|
| 889 |
+
"loss": 0.0211,
|
| 890 |
"step": 1470
|
| 891 |
},
|
| 892 |
{
|
| 893 |
"epoch": 3.71,
|
| 894 |
"learning_rate": 8.607242339832869e-06,
|
| 895 |
+
"loss": 0.1396,
|
| 896 |
"step": 1480
|
| 897 |
},
|
| 898 |
{
|
| 899 |
"epoch": 3.73,
|
| 900 |
"learning_rate": 8.440111420612815e-06,
|
| 901 |
+
"loss": 0.0316,
|
| 902 |
"step": 1490
|
| 903 |
},
|
| 904 |
{
|
| 905 |
"epoch": 3.76,
|
| 906 |
"learning_rate": 8.272980501392758e-06,
|
| 907 |
+
"loss": 0.0208,
|
| 908 |
"step": 1500
|
| 909 |
},
|
| 910 |
{
|
| 911 |
"epoch": 3.78,
|
| 912 |
"learning_rate": 8.1058495821727e-06,
|
| 913 |
+
"loss": 0.1344,
|
| 914 |
"step": 1510
|
| 915 |
},
|
| 916 |
{
|
| 917 |
"epoch": 3.81,
|
| 918 |
"learning_rate": 7.938718662952647e-06,
|
| 919 |
+
"loss": 0.0711,
|
| 920 |
"step": 1520
|
| 921 |
},
|
| 922 |
{
|
| 923 |
"epoch": 3.83,
|
| 924 |
"learning_rate": 7.77158774373259e-06,
|
| 925 |
+
"loss": 0.0596,
|
| 926 |
"step": 1530
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 3.86,
|
| 930 |
"learning_rate": 7.604456824512535e-06,
|
| 931 |
+
"loss": 0.0837,
|
| 932 |
"step": 1540
|
| 933 |
},
|
| 934 |
{
|
| 935 |
"epoch": 3.88,
|
| 936 |
"learning_rate": 7.43732590529248e-06,
|
| 937 |
+
"loss": 0.0242,
|
| 938 |
"step": 1550
|
| 939 |
},
|
| 940 |
{
|
| 941 |
"epoch": 3.91,
|
| 942 |
"learning_rate": 7.2701949860724235e-06,
|
| 943 |
+
"loss": 0.06,
|
| 944 |
"step": 1560
|
| 945 |
},
|
| 946 |
{
|
| 947 |
"epoch": 3.93,
|
| 948 |
"learning_rate": 7.103064066852368e-06,
|
| 949 |
+
"loss": 0.1006,
|
| 950 |
"step": 1570
|
| 951 |
},
|
| 952 |
{
|
| 953 |
"epoch": 3.96,
|
| 954 |
"learning_rate": 6.935933147632313e-06,
|
| 955 |
+
"loss": 0.0972,
|
| 956 |
"step": 1580
|
| 957 |
},
|
| 958 |
{
|
| 959 |
"epoch": 3.98,
|
| 960 |
"learning_rate": 6.768802228412256e-06,
|
| 961 |
+
"loss": 0.0627,
|
| 962 |
"step": 1590
|
| 963 |
},
|
| 964 |
{
|
| 965 |
"epoch": 4.01,
|
| 966 |
"learning_rate": 6.601671309192201e-06,
|
| 967 |
+
"loss": 0.01,
|
| 968 |
"step": 1600
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 4.04,
|
| 972 |
"learning_rate": 6.4345403899721455e-06,
|
| 973 |
+
"loss": 0.0734,
|
| 974 |
"step": 1610
|
| 975 |
},
|
| 976 |
{
|
| 977 |
"epoch": 4.06,
|
| 978 |
"learning_rate": 6.267409470752089e-06,
|
| 979 |
+
"loss": 0.0889,
|
| 980 |
"step": 1620
|
| 981 |
},
|
| 982 |
{
|
| 983 |
"epoch": 4.09,
|
| 984 |
"learning_rate": 6.100278551532034e-06,
|
| 985 |
+
"loss": 0.0109,
|
| 986 |
"step": 1630
|
| 987 |
},
|
| 988 |
{
|
| 989 |
"epoch": 4.11,
|
| 990 |
"learning_rate": 5.933147632311978e-06,
|
| 991 |
+
"loss": 0.0777,
|
| 992 |
"step": 1640
|
| 993 |
},
|
| 994 |
{
|
| 995 |
"epoch": 4.14,
|
| 996 |
"learning_rate": 5.766016713091923e-06,
|
| 997 |
+
"loss": 0.0475,
|
| 998 |
"step": 1650
|
| 999 |
},
|
| 1000 |
{
|
| 1001 |
"epoch": 4.16,
|
| 1002 |
"learning_rate": 5.598885793871867e-06,
|
| 1003 |
+
"loss": 0.0508,
|
| 1004 |
"step": 1660
|
| 1005 |
},
|
| 1006 |
{
|
| 1007 |
"epoch": 4.19,
|
| 1008 |
"learning_rate": 5.43175487465181e-06,
|
| 1009 |
+
"loss": 0.01,
|
| 1010 |
"step": 1670
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 4.21,
|
| 1014 |
"learning_rate": 5.264623955431755e-06,
|
| 1015 |
+
"loss": 0.0317,
|
| 1016 |
"step": 1680
|
| 1017 |
},
|
| 1018 |
{
|
| 1019 |
"epoch": 4.24,
|
| 1020 |
"learning_rate": 5.0974930362116986e-06,
|
| 1021 |
+
"loss": 0.0109,
|
| 1022 |
"step": 1690
|
| 1023 |
},
|
| 1024 |
{
|
| 1025 |
"epoch": 4.26,
|
| 1026 |
"learning_rate": 4.930362116991643e-06,
|
| 1027 |
+
"loss": 0.0466,
|
| 1028 |
"step": 1700
|
| 1029 |
},
|
| 1030 |
{
|
| 1031 |
"epoch": 4.29,
|
| 1032 |
"learning_rate": 4.763231197771588e-06,
|
| 1033 |
+
"loss": 0.0477,
|
| 1034 |
"step": 1710
|
| 1035 |
},
|
| 1036 |
{
|
| 1037 |
"epoch": 4.31,
|
| 1038 |
"learning_rate": 4.596100278551532e-06,
|
| 1039 |
+
"loss": 0.1175,
|
| 1040 |
"step": 1720
|
| 1041 |
},
|
| 1042 |
{
|
| 1043 |
"epoch": 4.34,
|
| 1044 |
"learning_rate": 4.428969359331476e-06,
|
| 1045 |
+
"loss": 0.0198,
|
| 1046 |
"step": 1730
|
| 1047 |
},
|
| 1048 |
{
|
| 1049 |
"epoch": 4.36,
|
| 1050 |
"learning_rate": 4.2618384401114205e-06,
|
| 1051 |
+
"loss": 0.0833,
|
| 1052 |
"step": 1740
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 4.39,
|
| 1056 |
"learning_rate": 4.094707520891365e-06,
|
| 1057 |
+
"loss": 0.1487,
|
| 1058 |
"step": 1750
|
| 1059 |
},
|
| 1060 |
{
|
| 1061 |
"epoch": 4.41,
|
| 1062 |
"learning_rate": 3.927576601671309e-06,
|
| 1063 |
+
"loss": 0.0096,
|
| 1064 |
"step": 1760
|
| 1065 |
},
|
| 1066 |
{
|
| 1067 |
"epoch": 4.44,
|
| 1068 |
"learning_rate": 3.7604456824512533e-06,
|
| 1069 |
+
"loss": 0.0101,
|
| 1070 |
"step": 1770
|
| 1071 |
},
|
| 1072 |
{
|
| 1073 |
"epoch": 4.46,
|
| 1074 |
"learning_rate": 3.593314763231198e-06,
|
| 1075 |
+
"loss": 0.01,
|
| 1076 |
"step": 1780
|
| 1077 |
},
|
| 1078 |
{
|
| 1079 |
"epoch": 4.49,
|
| 1080 |
"learning_rate": 3.426183844011142e-06,
|
| 1081 |
+
"loss": 0.0578,
|
| 1082 |
"step": 1790
|
| 1083 |
},
|
| 1084 |
{
|
| 1085 |
"epoch": 4.51,
|
| 1086 |
"learning_rate": 3.259052924791086e-06,
|
| 1087 |
+
"loss": 0.047,
|
| 1088 |
"step": 1800
|
| 1089 |
},
|
| 1090 |
{
|
| 1091 |
"epoch": 4.54,
|
| 1092 |
"learning_rate": 3.0919220055710307e-06,
|
| 1093 |
+
"loss": 0.08,
|
| 1094 |
"step": 1810
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 4.56,
|
| 1098 |
"learning_rate": 2.924791086350975e-06,
|
| 1099 |
+
"loss": 0.013,
|
| 1100 |
"step": 1820
|
| 1101 |
},
|
| 1102 |
{
|
| 1103 |
"epoch": 4.59,
|
| 1104 |
"learning_rate": 2.7576601671309194e-06,
|
| 1105 |
+
"loss": 0.0812,
|
| 1106 |
"step": 1830
|
| 1107 |
},
|
| 1108 |
{
|
| 1109 |
"epoch": 4.61,
|
| 1110 |
"learning_rate": 2.5905292479108636e-06,
|
| 1111 |
+
"loss": 0.037,
|
| 1112 |
"step": 1840
|
| 1113 |
},
|
| 1114 |
{
|
| 1115 |
"epoch": 4.64,
|
| 1116 |
"learning_rate": 2.4233983286908077e-06,
|
| 1117 |
+
"loss": 0.0472,
|
| 1118 |
"step": 1850
|
| 1119 |
},
|
| 1120 |
{
|
| 1121 |
"epoch": 4.66,
|
| 1122 |
"learning_rate": 2.2562674094707523e-06,
|
| 1123 |
+
"loss": 0.0905,
|
| 1124 |
"step": 1860
|
| 1125 |
},
|
| 1126 |
{
|
| 1127 |
"epoch": 4.69,
|
| 1128 |
"learning_rate": 2.0891364902506964e-06,
|
| 1129 |
+
"loss": 0.0469,
|
| 1130 |
"step": 1870
|
| 1131 |
},
|
| 1132 |
{
|
| 1133 |
"epoch": 4.71,
|
| 1134 |
"learning_rate": 1.922005571030641e-06,
|
| 1135 |
+
"loss": 0.0297,
|
| 1136 |
"step": 1880
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 4.74,
|
| 1140 |
"learning_rate": 1.7548746518105849e-06,
|
| 1141 |
+
"loss": 0.0126,
|
| 1142 |
"step": 1890
|
| 1143 |
},
|
| 1144 |
{
|
| 1145 |
"epoch": 4.76,
|
| 1146 |
"learning_rate": 1.5877437325905292e-06,
|
| 1147 |
+
"loss": 0.094,
|
| 1148 |
"step": 1900
|
| 1149 |
},
|
| 1150 |
{
|
| 1151 |
"epoch": 4.79,
|
| 1152 |
"learning_rate": 1.4206128133704736e-06,
|
| 1153 |
+
"loss": 0.0132,
|
| 1154 |
"step": 1910
|
| 1155 |
},
|
| 1156 |
{
|
| 1157 |
"epoch": 4.81,
|
| 1158 |
"learning_rate": 1.253481894150418e-06,
|
| 1159 |
+
"loss": 0.1048,
|
| 1160 |
"step": 1920
|
| 1161 |
},
|
| 1162 |
{
|
| 1163 |
"epoch": 4.84,
|
| 1164 |
"learning_rate": 1.0863509749303623e-06,
|
| 1165 |
+
"loss": 0.0526,
|
| 1166 |
"step": 1930
|
| 1167 |
},
|
| 1168 |
{
|
| 1169 |
"epoch": 4.86,
|
| 1170 |
"learning_rate": 9.192200557103064e-07,
|
| 1171 |
+
"loss": 0.0498,
|
| 1172 |
"step": 1940
|
| 1173 |
},
|
| 1174 |
{
|
|
|
|
| 1186 |
{
|
| 1187 |
"epoch": 4.94,
|
| 1188 |
"learning_rate": 4.178272980501393e-07,
|
| 1189 |
+
"loss": 0.0078,
|
| 1190 |
"step": 1970
|
| 1191 |
},
|
| 1192 |
{
|
| 1193 |
"epoch": 4.96,
|
| 1194 |
"learning_rate": 2.506963788300836e-07,
|
| 1195 |
+
"loss": 0.0161,
|
| 1196 |
"step": 1980
|
| 1197 |
},
|
| 1198 |
{
|
| 1199 |
"epoch": 4.99,
|
| 1200 |
"learning_rate": 8.356545961002785e-08,
|
| 1201 |
+
"loss": 0.0226,
|
| 1202 |
"step": 1990
|
| 1203 |
},
|
| 1204 |
{
|
| 1205 |
"epoch": 5.0,
|
| 1206 |
"step": 1995,
|
| 1207 |
"total_flos": 1.4837528359600128e+20,
|
| 1208 |
+
"train_loss": 0.32014525516290115,
|
| 1209 |
+
"train_runtime": 392.8902,
|
| 1210 |
+
"train_samples_per_second": 650.233,
|
| 1211 |
+
"train_steps_per_second": 5.078
|
| 1212 |
}
|
| 1213 |
],
|
| 1214 |
"max_steps": 1995,
|