Commit
•
70b60b4
1
Parent(s):
5b51cf0
Upload folder using huggingface_hub
Browse files- latest +1 -1
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- tokenizer.json +1 -6
- trainer_state.json +85 -1893
- training_args.bin +1 -1
latest
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
|
|
|
|
| 1 |
+
global_step200
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4944210912
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2565a7a2fc4ce63653c502762e327192c7d18c755f49a02f04260ff18dc04a85
|
| 3 |
size 4944210912
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4883292ca7b1cccf79cda6a921979df617dce655669564a8668ce1d9c573b4d
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4541564920
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d7bf9a6a50086d466d81a951fde610b3fda6e2eff1d7e5b447d7a6d1d854bff
|
| 3 |
size 4541564920
|
tokenizer.json
CHANGED
|
@@ -1,11 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
-
"truncation":
|
| 4 |
-
"direction": "Right",
|
| 5 |
-
"max_length": 4096,
|
| 6 |
-
"strategy": "LongestFirst",
|
| 7 |
-
"stride": 0
|
| 8 |
-
},
|
| 9 |
"padding": null,
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
|
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"padding": null,
|
| 5 |
"added_tokens": [
|
| 6 |
{
|
trainer_state.json
CHANGED
|
@@ -1,2066 +1,258 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 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": 0.0,
|
| 13 |
-
"learning_rate": 8.
|
| 14 |
"loss": 3.5463,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.0,
|
| 19 |
-
"learning_rate": 4.048582995951417e-
|
| 20 |
-
"loss": 3.
|
| 21 |
"step": 5
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
-
"learning_rate": 8.097165991902834e-
|
| 26 |
-
"loss":
|
| 27 |
"step": 10
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
-
"learning_rate": 1.214574898785425e-
|
| 32 |
-
"loss":
|
| 33 |
"step": 15
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
-
"learning_rate": 1.
|
| 38 |
-
"loss":
|
| 39 |
"step": 20
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
-
"learning_rate": 2.
|
| 44 |
-
"loss":
|
| 45 |
"step": 25
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
-
"learning_rate": 2.42914979757085e-
|
| 50 |
-
"loss":
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
-
"learning_rate": 2.
|
| 56 |
-
"loss":
|
| 57 |
"step": 35
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.0,
|
| 61 |
-
"learning_rate": 3.
|
| 62 |
-
"loss":
|
| 63 |
"step": 40
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
-
"learning_rate": 3.
|
| 68 |
-
"loss":
|
| 69 |
"step": 45
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.0,
|
| 73 |
-
"learning_rate": 4.
|
| 74 |
-
"loss":
|
| 75 |
"step": 50
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.0,
|
| 79 |
-
"learning_rate": 4.
|
| 80 |
-
"loss":
|
| 81 |
"step": 55
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.0,
|
| 85 |
-
"learning_rate": 4.8582995951417e-
|
| 86 |
-
"loss": 0.
|
| 87 |
"step": 60
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.0,
|
| 91 |
-
"learning_rate": 5.263157894736842e-
|
| 92 |
-
"loss": 0.
|
| 93 |
"step": 65
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.0,
|
| 97 |
-
"learning_rate": 5.
|
| 98 |
-
"loss": 0.
|
| 99 |
"step": 70
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.0,
|
| 103 |
-
"learning_rate": 6.
|
| 104 |
-
"loss": 0.
|
| 105 |
"step": 75
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.0,
|
| 109 |
-
"learning_rate": 6.
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 80
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.0,
|
| 115 |
-
"learning_rate": 6.
|
| 116 |
-
"loss": 0.
|
| 117 |
"step": 85
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.0,
|
| 121 |
-
"learning_rate": 7.
|
| 122 |
-
"loss": 0.
|
| 123 |
"step": 90
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.0,
|
| 127 |
-
"learning_rate": 7.
|
| 128 |
-
"loss": 0.
|
| 129 |
"step": 95
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.0,
|
| 133 |
-
"learning_rate": 8.
|
| 134 |
-
"loss": 0.
|
| 135 |
-
"step": 100
|
| 136 |
-
},
|
| 137 |
-
{
|
| 138 |
-
"epoch": 0.0,
|
| 139 |
-
"eval_loss": 0.7469579577445984,
|
| 140 |
-
"eval_runtime": 138.4332,
|
| 141 |
-
"eval_samples_per_second": 17.091,
|
| 142 |
-
"eval_steps_per_second": 2.853,
|
| 143 |
"step": 100
|
| 144 |
},
|
| 145 |
{
|
| 146 |
"epoch": 0.0,
|
| 147 |
-
"learning_rate": 8.
|
| 148 |
-
"loss": 0.
|
| 149 |
"step": 105
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.0,
|
| 153 |
-
"learning_rate": 8.
|
| 154 |
-
"loss": 0.
|
| 155 |
"step": 110
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 0.0,
|
| 159 |
-
"learning_rate": 9.
|
| 160 |
-
"loss": 0.
|
| 161 |
"step": 115
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"epoch": 0.0,
|
| 165 |
-
"learning_rate": 9.7165991902834e-
|
| 166 |
-
"loss": 0.
|
| 167 |
"step": 120
|
| 168 |
},
|
| 169 |
{
|
| 170 |
"epoch": 0.01,
|
| 171 |
-
"learning_rate": 1.
|
| 172 |
-
"loss": 0.
|
| 173 |
"step": 125
|
| 174 |
},
|
| 175 |
{
|
| 176 |
"epoch": 0.01,
|
| 177 |
-
"learning_rate": 1.
|
| 178 |
-
"loss": 0.
|
| 179 |
"step": 130
|
| 180 |
},
|
| 181 |
{
|
| 182 |
"epoch": 0.01,
|
| 183 |
-
"learning_rate": 1.
|
| 184 |
-
"loss": 0.
|
| 185 |
"step": 135
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 0.01,
|
| 189 |
-
"learning_rate": 1.
|
| 190 |
-
"loss": 0.
|
| 191 |
"step": 140
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.01,
|
| 195 |
-
"learning_rate": 1.
|
| 196 |
-
"loss": 0.
|
| 197 |
"step": 145
|
| 198 |
},
|
| 199 |
{
|
| 200 |
"epoch": 0.01,
|
| 201 |
-
"learning_rate": 1.
|
| 202 |
-
"loss": 0.
|
| 203 |
"step": 150
|
| 204 |
},
|
| 205 |
{
|
| 206 |
"epoch": 0.01,
|
| 207 |
-
"learning_rate": 1.
|
| 208 |
-
"loss": 0.
|
| 209 |
"step": 155
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 0.01,
|
| 213 |
-
"learning_rate": 1.
|
| 214 |
-
"loss": 0.
|
| 215 |
"step": 160
|
| 216 |
},
|
| 217 |
{
|
| 218 |
"epoch": 0.01,
|
| 219 |
-
"learning_rate": 1.
|
| 220 |
-
"loss": 0.
|
| 221 |
"step": 165
|
| 222 |
},
|
| 223 |
{
|
| 224 |
"epoch": 0.01,
|
| 225 |
-
"learning_rate": 1.
|
| 226 |
-
"loss": 0.
|
| 227 |
"step": 170
|
| 228 |
},
|
| 229 |
{
|
| 230 |
"epoch": 0.01,
|
| 231 |
-
"learning_rate": 1.
|
| 232 |
-
"loss": 0.
|
| 233 |
"step": 175
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.01,
|
| 237 |
-
"learning_rate": 1.
|
| 238 |
-
"loss": 0.
|
| 239 |
"step": 180
|
| 240 |
},
|
| 241 |
{
|
| 242 |
"epoch": 0.01,
|
| 243 |
-
"learning_rate": 1.
|
| 244 |
-
"loss": 0.
|
| 245 |
"step": 185
|
| 246 |
},
|
| 247 |
{
|
| 248 |
"epoch": 0.01,
|
| 249 |
-
"learning_rate": 1.
|
| 250 |
-
"loss": 0.
|
| 251 |
"step": 190
|
| 252 |
},
|
| 253 |
{
|
| 254 |
"epoch": 0.01,
|
| 255 |
-
"learning_rate": 1.
|
| 256 |
-
"loss": 0.
|
| 257 |
"step": 195
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 0.01,
|
| 261 |
-
"learning_rate": 1.
|
| 262 |
-
"loss": 0.
|
| 263 |
-
"step": 200
|
| 264 |
-
},
|
| 265 |
-
{
|
| 266 |
-
"epoch": 0.01,
|
| 267 |
-
"eval_loss": 0.8324583172798157,
|
| 268 |
-
"eval_runtime": 138.2786,
|
| 269 |
-
"eval_samples_per_second": 17.11,
|
| 270 |
-
"eval_steps_per_second": 2.857,
|
| 271 |
"step": 200
|
| 272 |
-
},
|
| 273 |
-
{
|
| 274 |
-
"epoch": 0.01,
|
| 275 |
-
"learning_rate": 1.659919028340081e-05,
|
| 276 |
-
"loss": 0.8627,
|
| 277 |
-
"step": 205
|
| 278 |
-
},
|
| 279 |
-
{
|
| 280 |
-
"epoch": 0.01,
|
| 281 |
-
"learning_rate": 1.7004048582995952e-05,
|
| 282 |
-
"loss": 0.8991,
|
| 283 |
-
"step": 210
|
| 284 |
-
},
|
| 285 |
-
{
|
| 286 |
-
"epoch": 0.01,
|
| 287 |
-
"learning_rate": 1.7408906882591094e-05,
|
| 288 |
-
"loss": 0.8706,
|
| 289 |
-
"step": 215
|
| 290 |
-
},
|
| 291 |
-
{
|
| 292 |
-
"epoch": 0.01,
|
| 293 |
-
"learning_rate": 1.7813765182186236e-05,
|
| 294 |
-
"loss": 0.8856,
|
| 295 |
-
"step": 220
|
| 296 |
-
},
|
| 297 |
-
{
|
| 298 |
-
"epoch": 0.01,
|
| 299 |
-
"learning_rate": 1.8218623481781378e-05,
|
| 300 |
-
"loss": 0.8669,
|
| 301 |
-
"step": 225
|
| 302 |
-
},
|
| 303 |
-
{
|
| 304 |
-
"epoch": 0.01,
|
| 305 |
-
"learning_rate": 1.862348178137652e-05,
|
| 306 |
-
"loss": 0.8914,
|
| 307 |
-
"step": 230
|
| 308 |
-
},
|
| 309 |
-
{
|
| 310 |
-
"epoch": 0.01,
|
| 311 |
-
"learning_rate": 1.9028340080971662e-05,
|
| 312 |
-
"loss": 0.8975,
|
| 313 |
-
"step": 235
|
| 314 |
-
},
|
| 315 |
-
{
|
| 316 |
-
"epoch": 0.01,
|
| 317 |
-
"learning_rate": 1.94331983805668e-05,
|
| 318 |
-
"loss": 0.9088,
|
| 319 |
-
"step": 240
|
| 320 |
-
},
|
| 321 |
-
{
|
| 322 |
-
"epoch": 0.01,
|
| 323 |
-
"learning_rate": 1.9838056680161946e-05,
|
| 324 |
-
"loss": 0.8997,
|
| 325 |
-
"step": 245
|
| 326 |
-
},
|
| 327 |
-
{
|
| 328 |
-
"epoch": 0.01,
|
| 329 |
-
"learning_rate": 1.999999925229564e-05,
|
| 330 |
-
"loss": 0.9735,
|
| 331 |
-
"step": 250
|
| 332 |
-
},
|
| 333 |
-
{
|
| 334 |
-
"epoch": 0.01,
|
| 335 |
-
"learning_rate": 1.9999994682991605e-05,
|
| 336 |
-
"loss": 0.9171,
|
| 337 |
-
"step": 255
|
| 338 |
-
},
|
| 339 |
-
{
|
| 340 |
-
"epoch": 0.01,
|
| 341 |
-
"learning_rate": 1.999998595977674e-05,
|
| 342 |
-
"loss": 0.9059,
|
| 343 |
-
"step": 260
|
| 344 |
-
},
|
| 345 |
-
{
|
| 346 |
-
"epoch": 0.01,
|
| 347 |
-
"learning_rate": 1.9999973082654672e-05,
|
| 348 |
-
"loss": 0.9165,
|
| 349 |
-
"step": 265
|
| 350 |
-
},
|
| 351 |
-
{
|
| 352 |
-
"epoch": 0.01,
|
| 353 |
-
"learning_rate": 1.9999956051630752e-05,
|
| 354 |
-
"loss": 0.9258,
|
| 355 |
-
"step": 270
|
| 356 |
-
},
|
| 357 |
-
{
|
| 358 |
-
"epoch": 0.01,
|
| 359 |
-
"learning_rate": 1.999993486671205e-05,
|
| 360 |
-
"loss": 0.9463,
|
| 361 |
-
"step": 275
|
| 362 |
-
},
|
| 363 |
-
{
|
| 364 |
-
"epoch": 0.01,
|
| 365 |
-
"learning_rate": 1.999990952790737e-05,
|
| 366 |
-
"loss": 0.937,
|
| 367 |
-
"step": 280
|
| 368 |
-
},
|
| 369 |
-
{
|
| 370 |
-
"epoch": 0.01,
|
| 371 |
-
"learning_rate": 1.9999880035227236e-05,
|
| 372 |
-
"loss": 0.9516,
|
| 373 |
-
"step": 285
|
| 374 |
-
},
|
| 375 |
-
{
|
| 376 |
-
"epoch": 0.01,
|
| 377 |
-
"learning_rate": 1.9999846388683897e-05,
|
| 378 |
-
"loss": 0.9477,
|
| 379 |
-
"step": 290
|
| 380 |
-
},
|
| 381 |
-
{
|
| 382 |
-
"epoch": 0.01,
|
| 383 |
-
"learning_rate": 1.999980858829133e-05,
|
| 384 |
-
"loss": 0.9517,
|
| 385 |
-
"step": 295
|
| 386 |
-
},
|
| 387 |
-
{
|
| 388 |
-
"epoch": 0.01,
|
| 389 |
-
"learning_rate": 1.9999766634065244e-05,
|
| 390 |
-
"loss": 0.9497,
|
| 391 |
-
"step": 300
|
| 392 |
-
},
|
| 393 |
-
{
|
| 394 |
-
"epoch": 0.01,
|
| 395 |
-
"eval_loss": 0.9005245566368103,
|
| 396 |
-
"eval_runtime": 138.1539,
|
| 397 |
-
"eval_samples_per_second": 17.126,
|
| 398 |
-
"eval_steps_per_second": 2.859,
|
| 399 |
-
"step": 300
|
| 400 |
-
},
|
| 401 |
-
{
|
| 402 |
-
"epoch": 0.01,
|
| 403 |
-
"learning_rate": 1.9999720526023055e-05,
|
| 404 |
-
"loss": 0.9225,
|
| 405 |
-
"step": 305
|
| 406 |
-
},
|
| 407 |
-
{
|
| 408 |
-
"epoch": 0.01,
|
| 409 |
-
"learning_rate": 1.999967026418392e-05,
|
| 410 |
-
"loss": 0.9641,
|
| 411 |
-
"step": 310
|
| 412 |
-
},
|
| 413 |
-
{
|
| 414 |
-
"epoch": 0.01,
|
| 415 |
-
"learning_rate": 1.999961584856872e-05,
|
| 416 |
-
"loss": 0.9182,
|
| 417 |
-
"step": 315
|
| 418 |
-
},
|
| 419 |
-
{
|
| 420 |
-
"epoch": 0.01,
|
| 421 |
-
"learning_rate": 1.999955727920006e-05,
|
| 422 |
-
"loss": 0.9285,
|
| 423 |
-
"step": 320
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.01,
|
| 427 |
-
"learning_rate": 1.9999494556102267e-05,
|
| 428 |
-
"loss": 0.9027,
|
| 429 |
-
"step": 325
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 0.01,
|
| 433 |
-
"learning_rate": 1.999942767930139e-05,
|
| 434 |
-
"loss": 0.964,
|
| 435 |
-
"step": 330
|
| 436 |
-
},
|
| 437 |
-
{
|
| 438 |
-
"epoch": 0.01,
|
| 439 |
-
"learning_rate": 1.9999356648825218e-05,
|
| 440 |
-
"loss": 0.9156,
|
| 441 |
-
"step": 335
|
| 442 |
-
},
|
| 443 |
-
{
|
| 444 |
-
"epoch": 0.01,
|
| 445 |
-
"learning_rate": 1.9999281464703252e-05,
|
| 446 |
-
"loss": 0.9136,
|
| 447 |
-
"step": 340
|
| 448 |
-
},
|
| 449 |
-
{
|
| 450 |
-
"epoch": 0.01,
|
| 451 |
-
"learning_rate": 1.9999202126966724e-05,
|
| 452 |
-
"loss": 0.9605,
|
| 453 |
-
"step": 345
|
| 454 |
-
},
|
| 455 |
-
{
|
| 456 |
-
"epoch": 0.01,
|
| 457 |
-
"learning_rate": 1.999911863564859e-05,
|
| 458 |
-
"loss": 0.9233,
|
| 459 |
-
"step": 350
|
| 460 |
-
},
|
| 461 |
-
{
|
| 462 |
-
"epoch": 0.01,
|
| 463 |
-
"learning_rate": 1.999903099078353e-05,
|
| 464 |
-
"loss": 0.9181,
|
| 465 |
-
"step": 355
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.01,
|
| 469 |
-
"learning_rate": 1.9998939192407955e-05,
|
| 470 |
-
"loss": 0.9082,
|
| 471 |
-
"step": 360
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 0.01,
|
| 475 |
-
"learning_rate": 1.999884324055999e-05,
|
| 476 |
-
"loss": 0.9724,
|
| 477 |
-
"step": 365
|
| 478 |
-
},
|
| 479 |
-
{
|
| 480 |
-
"epoch": 0.02,
|
| 481 |
-
"learning_rate": 1.9998743135279503e-05,
|
| 482 |
-
"loss": 0.9459,
|
| 483 |
-
"step": 370
|
| 484 |
-
},
|
| 485 |
-
{
|
| 486 |
-
"epoch": 0.02,
|
| 487 |
-
"learning_rate": 1.9998638876608064e-05,
|
| 488 |
-
"loss": 0.9451,
|
| 489 |
-
"step": 375
|
| 490 |
-
},
|
| 491 |
-
{
|
| 492 |
-
"epoch": 0.02,
|
| 493 |
-
"learning_rate": 1.9998530464588993e-05,
|
| 494 |
-
"loss": 0.9103,
|
| 495 |
-
"step": 380
|
| 496 |
-
},
|
| 497 |
-
{
|
| 498 |
-
"epoch": 0.02,
|
| 499 |
-
"learning_rate": 1.9998417899267316e-05,
|
| 500 |
-
"loss": 0.9728,
|
| 501 |
-
"step": 385
|
| 502 |
-
},
|
| 503 |
-
{
|
| 504 |
-
"epoch": 0.02,
|
| 505 |
-
"learning_rate": 1.9998301180689794e-05,
|
| 506 |
-
"loss": 0.9346,
|
| 507 |
-
"step": 390
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.02,
|
| 511 |
-
"learning_rate": 1.9998180308904913e-05,
|
| 512 |
-
"loss": 0.9321,
|
| 513 |
-
"step": 395
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 0.02,
|
| 517 |
-
"learning_rate": 1.999805528396288e-05,
|
| 518 |
-
"loss": 0.9671,
|
| 519 |
-
"step": 400
|
| 520 |
-
},
|
| 521 |
-
{
|
| 522 |
-
"epoch": 0.02,
|
| 523 |
-
"eval_loss": 0.9101449251174927,
|
| 524 |
-
"eval_runtime": 138.1297,
|
| 525 |
-
"eval_samples_per_second": 17.129,
|
| 526 |
-
"eval_steps_per_second": 2.86,
|
| 527 |
-
"step": 400
|
| 528 |
-
},
|
| 529 |
-
{
|
| 530 |
-
"epoch": 0.02,
|
| 531 |
-
"learning_rate": 1.9997926105915627e-05,
|
| 532 |
-
"loss": 0.9217,
|
| 533 |
-
"step": 405
|
| 534 |
-
},
|
| 535 |
-
{
|
| 536 |
-
"epoch": 0.02,
|
| 537 |
-
"learning_rate": 1.999779277481682e-05,
|
| 538 |
-
"loss": 0.9484,
|
| 539 |
-
"step": 410
|
| 540 |
-
},
|
| 541 |
-
{
|
| 542 |
-
"epoch": 0.02,
|
| 543 |
-
"learning_rate": 1.9997655290721837e-05,
|
| 544 |
-
"loss": 0.9619,
|
| 545 |
-
"step": 415
|
| 546 |
-
},
|
| 547 |
-
{
|
| 548 |
-
"epoch": 0.02,
|
| 549 |
-
"learning_rate": 1.999751365368779e-05,
|
| 550 |
-
"loss": 0.945,
|
| 551 |
-
"step": 420
|
| 552 |
-
},
|
| 553 |
-
{
|
| 554 |
-
"epoch": 0.02,
|
| 555 |
-
"learning_rate": 1.9997367863773512e-05,
|
| 556 |
-
"loss": 0.9864,
|
| 557 |
-
"step": 425
|
| 558 |
-
},
|
| 559 |
-
{
|
| 560 |
-
"epoch": 0.02,
|
| 561 |
-
"learning_rate": 1.999721792103957e-05,
|
| 562 |
-
"loss": 0.9584,
|
| 563 |
-
"step": 430
|
| 564 |
-
},
|
| 565 |
-
{
|
| 566 |
-
"epoch": 0.02,
|
| 567 |
-
"learning_rate": 1.999706382554824e-05,
|
| 568 |
-
"loss": 0.9184,
|
| 569 |
-
"step": 435
|
| 570 |
-
},
|
| 571 |
-
{
|
| 572 |
-
"epoch": 0.02,
|
| 573 |
-
"learning_rate": 1.9996905577363534e-05,
|
| 574 |
-
"loss": 0.9764,
|
| 575 |
-
"step": 440
|
| 576 |
-
},
|
| 577 |
-
{
|
| 578 |
-
"epoch": 0.02,
|
| 579 |
-
"learning_rate": 1.999674317655119e-05,
|
| 580 |
-
"loss": 0.9649,
|
| 581 |
-
"step": 445
|
| 582 |
-
},
|
| 583 |
-
{
|
| 584 |
-
"epoch": 0.02,
|
| 585 |
-
"learning_rate": 1.9996576623178665e-05,
|
| 586 |
-
"loss": 0.9567,
|
| 587 |
-
"step": 450
|
| 588 |
-
},
|
| 589 |
-
{
|
| 590 |
-
"epoch": 0.02,
|
| 591 |
-
"learning_rate": 1.9996405917315148e-05,
|
| 592 |
-
"loss": 0.9309,
|
| 593 |
-
"step": 455
|
| 594 |
-
},
|
| 595 |
-
{
|
| 596 |
-
"epoch": 0.02,
|
| 597 |
-
"learning_rate": 1.9996231059031547e-05,
|
| 598 |
-
"loss": 0.9588,
|
| 599 |
-
"step": 460
|
| 600 |
-
},
|
| 601 |
-
{
|
| 602 |
-
"epoch": 0.02,
|
| 603 |
-
"learning_rate": 1.999605204840049e-05,
|
| 604 |
-
"loss": 0.9717,
|
| 605 |
-
"step": 465
|
| 606 |
-
},
|
| 607 |
-
{
|
| 608 |
-
"epoch": 0.02,
|
| 609 |
-
"learning_rate": 1.999586888549635e-05,
|
| 610 |
-
"loss": 0.9574,
|
| 611 |
-
"step": 470
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 0.02,
|
| 615 |
-
"learning_rate": 1.9995681570395196e-05,
|
| 616 |
-
"loss": 1.01,
|
| 617 |
-
"step": 475
|
| 618 |
-
},
|
| 619 |
-
{
|
| 620 |
-
"epoch": 0.02,
|
| 621 |
-
"learning_rate": 1.999549010317485e-05,
|
| 622 |
-
"loss": 0.955,
|
| 623 |
-
"step": 480
|
| 624 |
-
},
|
| 625 |
-
{
|
| 626 |
-
"epoch": 0.02,
|
| 627 |
-
"learning_rate": 1.9995294483914834e-05,
|
| 628 |
-
"loss": 0.981,
|
| 629 |
-
"step": 485
|
| 630 |
-
},
|
| 631 |
-
{
|
| 632 |
-
"epoch": 0.02,
|
| 633 |
-
"learning_rate": 1.9995094712696413e-05,
|
| 634 |
-
"loss": 0.9827,
|
| 635 |
-
"step": 490
|
| 636 |
-
},
|
| 637 |
-
{
|
| 638 |
-
"epoch": 0.02,
|
| 639 |
-
"learning_rate": 1.9994890789602575e-05,
|
| 640 |
-
"loss": 0.9377,
|
| 641 |
-
"step": 495
|
| 642 |
-
},
|
| 643 |
-
{
|
| 644 |
-
"epoch": 0.02,
|
| 645 |
-
"learning_rate": 1.9994682714718023e-05,
|
| 646 |
-
"loss": 0.9555,
|
| 647 |
-
"step": 500
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 0.02,
|
| 651 |
-
"eval_loss": 0.9095268249511719,
|
| 652 |
-
"eval_runtime": 138.1033,
|
| 653 |
-
"eval_samples_per_second": 17.132,
|
| 654 |
-
"eval_steps_per_second": 2.86,
|
| 655 |
-
"step": 500
|
| 656 |
-
},
|
| 657 |
-
{
|
| 658 |
-
"epoch": 0.02,
|
| 659 |
-
"learning_rate": 1.9994470488129188e-05,
|
| 660 |
-
"loss": 0.9694,
|
| 661 |
-
"step": 505
|
| 662 |
-
},
|
| 663 |
-
{
|
| 664 |
-
"epoch": 0.02,
|
| 665 |
-
"learning_rate": 1.9994254109924226e-05,
|
| 666 |
-
"loss": 0.9381,
|
| 667 |
-
"step": 510
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 0.02,
|
| 671 |
-
"learning_rate": 1.999403358019302e-05,
|
| 672 |
-
"loss": 0.9496,
|
| 673 |
-
"step": 515
|
| 674 |
-
},
|
| 675 |
-
{
|
| 676 |
-
"epoch": 0.02,
|
| 677 |
-
"learning_rate": 1.9993808899027184e-05,
|
| 678 |
-
"loss": 0.9329,
|
| 679 |
-
"step": 520
|
| 680 |
-
},
|
| 681 |
-
{
|
| 682 |
-
"epoch": 0.02,
|
| 683 |
-
"learning_rate": 1.9993580066520035e-05,
|
| 684 |
-
"loss": 0.9713,
|
| 685 |
-
"step": 525
|
| 686 |
-
},
|
| 687 |
-
{
|
| 688 |
-
"epoch": 0.02,
|
| 689 |
-
"learning_rate": 1.9993347082766638e-05,
|
| 690 |
-
"loss": 0.9702,
|
| 691 |
-
"step": 530
|
| 692 |
-
},
|
| 693 |
-
{
|
| 694 |
-
"epoch": 0.02,
|
| 695 |
-
"learning_rate": 1.9993109947863768e-05,
|
| 696 |
-
"loss": 0.9277,
|
| 697 |
-
"step": 535
|
| 698 |
-
},
|
| 699 |
-
{
|
| 700 |
-
"epoch": 0.02,
|
| 701 |
-
"learning_rate": 1.999286866190993e-05,
|
| 702 |
-
"loss": 0.9512,
|
| 703 |
-
"step": 540
|
| 704 |
-
},
|
| 705 |
-
{
|
| 706 |
-
"epoch": 0.02,
|
| 707 |
-
"learning_rate": 1.9992623225005353e-05,
|
| 708 |
-
"loss": 0.9268,
|
| 709 |
-
"step": 545
|
| 710 |
-
},
|
| 711 |
-
{
|
| 712 |
-
"epoch": 0.02,
|
| 713 |
-
"learning_rate": 1.9992373637251985e-05,
|
| 714 |
-
"loss": 0.9178,
|
| 715 |
-
"step": 550
|
| 716 |
-
},
|
| 717 |
-
{
|
| 718 |
-
"epoch": 0.02,
|
| 719 |
-
"learning_rate": 1.999211989875351e-05,
|
| 720 |
-
"loss": 0.9274,
|
| 721 |
-
"step": 555
|
| 722 |
-
},
|
| 723 |
-
{
|
| 724 |
-
"epoch": 0.02,
|
| 725 |
-
"learning_rate": 1.9991862009615325e-05,
|
| 726 |
-
"loss": 0.9679,
|
| 727 |
-
"step": 560
|
| 728 |
-
},
|
| 729 |
-
{
|
| 730 |
-
"epoch": 0.02,
|
| 731 |
-
"learning_rate": 1.9991599969944554e-05,
|
| 732 |
-
"loss": 0.9737,
|
| 733 |
-
"step": 565
|
| 734 |
-
},
|
| 735 |
-
{
|
| 736 |
-
"epoch": 0.02,
|
| 737 |
-
"learning_rate": 1.9991333779850044e-05,
|
| 738 |
-
"loss": 0.9297,
|
| 739 |
-
"step": 570
|
| 740 |
-
},
|
| 741 |
-
{
|
| 742 |
-
"epoch": 0.02,
|
| 743 |
-
"learning_rate": 1.999106343944237e-05,
|
| 744 |
-
"loss": 0.967,
|
| 745 |
-
"step": 575
|
| 746 |
-
},
|
| 747 |
-
{
|
| 748 |
-
"epoch": 0.02,
|
| 749 |
-
"learning_rate": 1.9990788948833835e-05,
|
| 750 |
-
"loss": 0.9758,
|
| 751 |
-
"step": 580
|
| 752 |
-
},
|
| 753 |
-
{
|
| 754 |
-
"epoch": 0.02,
|
| 755 |
-
"learning_rate": 1.9990510308138452e-05,
|
| 756 |
-
"loss": 0.9363,
|
| 757 |
-
"step": 585
|
| 758 |
-
},
|
| 759 |
-
{
|
| 760 |
-
"epoch": 0.02,
|
| 761 |
-
"learning_rate": 1.999022751747197e-05,
|
| 762 |
-
"loss": 0.9524,
|
| 763 |
-
"step": 590
|
| 764 |
-
},
|
| 765 |
-
{
|
| 766 |
-
"epoch": 0.02,
|
| 767 |
-
"learning_rate": 1.9989940576951853e-05,
|
| 768 |
-
"loss": 0.945,
|
| 769 |
-
"step": 595
|
| 770 |
-
},
|
| 771 |
-
{
|
| 772 |
-
"epoch": 0.02,
|
| 773 |
-
"learning_rate": 1.99896494866973e-05,
|
| 774 |
-
"loss": 0.8981,
|
| 775 |
-
"step": 600
|
| 776 |
-
},
|
| 777 |
-
{
|
| 778 |
-
"epoch": 0.02,
|
| 779 |
-
"eval_loss": 0.917118489742279,
|
| 780 |
-
"eval_runtime": 138.0324,
|
| 781 |
-
"eval_samples_per_second": 17.141,
|
| 782 |
-
"eval_steps_per_second": 2.862,
|
| 783 |
-
"step": 600
|
| 784 |
-
},
|
| 785 |
-
{
|
| 786 |
-
"epoch": 0.02,
|
| 787 |
-
"learning_rate": 1.9989354246829226e-05,
|
| 788 |
-
"loss": 0.9747,
|
| 789 |
-
"step": 605
|
| 790 |
-
},
|
| 791 |
-
{
|
| 792 |
-
"epoch": 0.02,
|
| 793 |
-
"learning_rate": 1.998905485747027e-05,
|
| 794 |
-
"loss": 0.9696,
|
| 795 |
-
"step": 610
|
| 796 |
-
},
|
| 797 |
-
{
|
| 798 |
-
"epoch": 0.02,
|
| 799 |
-
"learning_rate": 1.998875131874479e-05,
|
| 800 |
-
"loss": 1.0225,
|
| 801 |
-
"step": 615
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 0.03,
|
| 805 |
-
"learning_rate": 1.998844363077888e-05,
|
| 806 |
-
"loss": 0.9506,
|
| 807 |
-
"step": 620
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 0.03,
|
| 811 |
-
"learning_rate": 1.9988131793700352e-05,
|
| 812 |
-
"loss": 0.9706,
|
| 813 |
-
"step": 625
|
| 814 |
-
},
|
| 815 |
-
{
|
| 816 |
-
"epoch": 0.03,
|
| 817 |
-
"learning_rate": 1.998781580763874e-05,
|
| 818 |
-
"loss": 0.9166,
|
| 819 |
-
"step": 630
|
| 820 |
-
},
|
| 821 |
-
{
|
| 822 |
-
"epoch": 0.03,
|
| 823 |
-
"learning_rate": 1.9987495672725294e-05,
|
| 824 |
-
"loss": 0.9262,
|
| 825 |
-
"step": 635
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 0.03,
|
| 829 |
-
"learning_rate": 1.9987171389093002e-05,
|
| 830 |
-
"loss": 0.9994,
|
| 831 |
-
"step": 640
|
| 832 |
-
},
|
| 833 |
-
{
|
| 834 |
-
"epoch": 0.03,
|
| 835 |
-
"learning_rate": 1.998684295687657e-05,
|
| 836 |
-
"loss": 0.9241,
|
| 837 |
-
"step": 645
|
| 838 |
-
},
|
| 839 |
-
{
|
| 840 |
-
"epoch": 0.03,
|
| 841 |
-
"learning_rate": 1.9986510376212424e-05,
|
| 842 |
-
"loss": 1.0308,
|
| 843 |
-
"step": 650
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 0.03,
|
| 847 |
-
"learning_rate": 1.9986173647238717e-05,
|
| 848 |
-
"loss": 1.0005,
|
| 849 |
-
"step": 655
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 0.03,
|
| 853 |
-
"learning_rate": 1.9985832770095313e-05,
|
| 854 |
-
"loss": 0.9595,
|
| 855 |
-
"step": 660
|
| 856 |
-
},
|
| 857 |
-
{
|
| 858 |
-
"epoch": 0.03,
|
| 859 |
-
"learning_rate": 1.9985487744923825e-05,
|
| 860 |
-
"loss": 0.9606,
|
| 861 |
-
"step": 665
|
| 862 |
-
},
|
| 863 |
-
{
|
| 864 |
-
"epoch": 0.03,
|
| 865 |
-
"learning_rate": 1.9985138571867562e-05,
|
| 866 |
-
"loss": 0.9722,
|
| 867 |
-
"step": 670
|
| 868 |
-
},
|
| 869 |
-
{
|
| 870 |
-
"epoch": 0.03,
|
| 871 |
-
"learning_rate": 1.998478525107157e-05,
|
| 872 |
-
"loss": 0.9772,
|
| 873 |
-
"step": 675
|
| 874 |
-
},
|
| 875 |
-
{
|
| 876 |
-
"epoch": 0.03,
|
| 877 |
-
"learning_rate": 1.998442778268262e-05,
|
| 878 |
-
"loss": 0.9536,
|
| 879 |
-
"step": 680
|
| 880 |
-
},
|
| 881 |
-
{
|
| 882 |
-
"epoch": 0.03,
|
| 883 |
-
"learning_rate": 1.9984066166849198e-05,
|
| 884 |
-
"loss": 0.9852,
|
| 885 |
-
"step": 685
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 0.03,
|
| 889 |
-
"learning_rate": 1.9983700403721517e-05,
|
| 890 |
-
"loss": 1.0,
|
| 891 |
-
"step": 690
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 0.03,
|
| 895 |
-
"learning_rate": 1.9983330493451506e-05,
|
| 896 |
-
"loss": 0.9386,
|
| 897 |
-
"step": 695
|
| 898 |
-
},
|
| 899 |
-
{
|
| 900 |
-
"epoch": 0.03,
|
| 901 |
-
"learning_rate": 1.9982956436192828e-05,
|
| 902 |
-
"loss": 1.0199,
|
| 903 |
-
"step": 700
|
| 904 |
-
},
|
| 905 |
-
{
|
| 906 |
-
"epoch": 0.03,
|
| 907 |
-
"eval_loss": 0.9225443005561829,
|
| 908 |
-
"eval_runtime": 138.1507,
|
| 909 |
-
"eval_samples_per_second": 17.126,
|
| 910 |
-
"eval_steps_per_second": 2.859,
|
| 911 |
-
"step": 700
|
| 912 |
-
},
|
| 913 |
-
{
|
| 914 |
-
"epoch": 0.03,
|
| 915 |
-
"learning_rate": 1.9982578232100865e-05,
|
| 916 |
-
"loss": 0.9655,
|
| 917 |
-
"step": 705
|
| 918 |
-
},
|
| 919 |
-
{
|
| 920 |
-
"epoch": 0.03,
|
| 921 |
-
"learning_rate": 1.998219588133272e-05,
|
| 922 |
-
"loss": 1.0133,
|
| 923 |
-
"step": 710
|
| 924 |
-
},
|
| 925 |
-
{
|
| 926 |
-
"epoch": 0.03,
|
| 927 |
-
"learning_rate": 1.998180938404721e-05,
|
| 928 |
-
"loss": 0.9805,
|
| 929 |
-
"step": 715
|
| 930 |
-
},
|
| 931 |
-
{
|
| 932 |
-
"epoch": 0.03,
|
| 933 |
-
"learning_rate": 1.9981418740404886e-05,
|
| 934 |
-
"loss": 1.0031,
|
| 935 |
-
"step": 720
|
| 936 |
-
},
|
| 937 |
-
{
|
| 938 |
-
"epoch": 0.03,
|
| 939 |
-
"learning_rate": 1.9981023950568024e-05,
|
| 940 |
-
"loss": 0.9759,
|
| 941 |
-
"step": 725
|
| 942 |
-
},
|
| 943 |
-
{
|
| 944 |
-
"epoch": 0.03,
|
| 945 |
-
"learning_rate": 1.998062501470061e-05,
|
| 946 |
-
"loss": 0.9529,
|
| 947 |
-
"step": 730
|
| 948 |
-
},
|
| 949 |
-
{
|
| 950 |
-
"epoch": 0.03,
|
| 951 |
-
"learning_rate": 1.9980221932968363e-05,
|
| 952 |
-
"loss": 0.9865,
|
| 953 |
-
"step": 735
|
| 954 |
-
},
|
| 955 |
-
{
|
| 956 |
-
"epoch": 0.03,
|
| 957 |
-
"learning_rate": 1.997981470553872e-05,
|
| 958 |
-
"loss": 0.9195,
|
| 959 |
-
"step": 740
|
| 960 |
-
},
|
| 961 |
-
{
|
| 962 |
-
"epoch": 0.03,
|
| 963 |
-
"learning_rate": 1.997940333258083e-05,
|
| 964 |
-
"loss": 0.9669,
|
| 965 |
-
"step": 745
|
| 966 |
-
},
|
| 967 |
-
{
|
| 968 |
-
"epoch": 0.03,
|
| 969 |
-
"learning_rate": 1.9978987814265583e-05,
|
| 970 |
-
"loss": 0.9341,
|
| 971 |
-
"step": 750
|
| 972 |
-
},
|
| 973 |
-
{
|
| 974 |
-
"epoch": 0.03,
|
| 975 |
-
"learning_rate": 1.9978568150765584e-05,
|
| 976 |
-
"loss": 0.9229,
|
| 977 |
-
"step": 755
|
| 978 |
-
},
|
| 979 |
-
{
|
| 980 |
-
"epoch": 0.03,
|
| 981 |
-
"learning_rate": 1.997814434225515e-05,
|
| 982 |
-
"loss": 0.9183,
|
| 983 |
-
"step": 760
|
| 984 |
-
},
|
| 985 |
-
{
|
| 986 |
-
"epoch": 0.03,
|
| 987 |
-
"learning_rate": 1.997771638891033e-05,
|
| 988 |
-
"loss": 0.9622,
|
| 989 |
-
"step": 765
|
| 990 |
-
},
|
| 991 |
-
{
|
| 992 |
-
"epoch": 0.03,
|
| 993 |
-
"learning_rate": 1.997728429090889e-05,
|
| 994 |
-
"loss": 1.0141,
|
| 995 |
-
"step": 770
|
| 996 |
-
},
|
| 997 |
-
{
|
| 998 |
-
"epoch": 0.03,
|
| 999 |
-
"learning_rate": 1.9976848048430325e-05,
|
| 1000 |
-
"loss": 0.983,
|
| 1001 |
-
"step": 775
|
| 1002 |
-
},
|
| 1003 |
-
{
|
| 1004 |
-
"epoch": 0.03,
|
| 1005 |
-
"learning_rate": 1.9976407661655846e-05,
|
| 1006 |
-
"loss": 0.9445,
|
| 1007 |
-
"step": 780
|
| 1008 |
-
},
|
| 1009 |
-
{
|
| 1010 |
-
"epoch": 0.03,
|
| 1011 |
-
"learning_rate": 1.997596313076838e-05,
|
| 1012 |
-
"loss": 0.9103,
|
| 1013 |
-
"step": 785
|
| 1014 |
-
},
|
| 1015 |
-
{
|
| 1016 |
-
"epoch": 0.03,
|
| 1017 |
-
"learning_rate": 1.9975514455952584e-05,
|
| 1018 |
-
"loss": 0.9715,
|
| 1019 |
-
"step": 790
|
| 1020 |
-
},
|
| 1021 |
-
{
|
| 1022 |
-
"epoch": 0.03,
|
| 1023 |
-
"learning_rate": 1.997506163739484e-05,
|
| 1024 |
-
"loss": 0.9485,
|
| 1025 |
-
"step": 795
|
| 1026 |
-
},
|
| 1027 |
-
{
|
| 1028 |
-
"epoch": 0.03,
|
| 1029 |
-
"learning_rate": 1.9974604675283232e-05,
|
| 1030 |
-
"loss": 0.9462,
|
| 1031 |
-
"step": 800
|
| 1032 |
-
},
|
| 1033 |
-
{
|
| 1034 |
-
"epoch": 0.03,
|
| 1035 |
-
"eval_loss": 0.9193007946014404,
|
| 1036 |
-
"eval_runtime": 138.1814,
|
| 1037 |
-
"eval_samples_per_second": 17.122,
|
| 1038 |
-
"eval_steps_per_second": 2.859,
|
| 1039 |
-
"step": 800
|
| 1040 |
-
},
|
| 1041 |
-
{
|
| 1042 |
-
"epoch": 0.03,
|
| 1043 |
-
"learning_rate": 1.9974143569807588e-05,
|
| 1044 |
-
"loss": 1.0423,
|
| 1045 |
-
"step": 805
|
| 1046 |
-
},
|
| 1047 |
-
{
|
| 1048 |
-
"epoch": 0.03,
|
| 1049 |
-
"learning_rate": 1.9973678321159445e-05,
|
| 1050 |
-
"loss": 0.9362,
|
| 1051 |
-
"step": 810
|
| 1052 |
-
},
|
| 1053 |
-
{
|
| 1054 |
-
"epoch": 0.03,
|
| 1055 |
-
"learning_rate": 1.9973208929532065e-05,
|
| 1056 |
-
"loss": 0.9342,
|
| 1057 |
-
"step": 815
|
| 1058 |
-
},
|
| 1059 |
-
{
|
| 1060 |
-
"epoch": 0.03,
|
| 1061 |
-
"learning_rate": 1.997273539512042e-05,
|
| 1062 |
-
"loss": 0.9573,
|
| 1063 |
-
"step": 820
|
| 1064 |
-
},
|
| 1065 |
-
{
|
| 1066 |
-
"epoch": 0.03,
|
| 1067 |
-
"learning_rate": 1.9972257718121225e-05,
|
| 1068 |
-
"loss": 0.9502,
|
| 1069 |
-
"step": 825
|
| 1070 |
-
},
|
| 1071 |
-
{
|
| 1072 |
-
"epoch": 0.03,
|
| 1073 |
-
"learning_rate": 1.9971775898732896e-05,
|
| 1074 |
-
"loss": 0.9559,
|
| 1075 |
-
"step": 830
|
| 1076 |
-
},
|
| 1077 |
-
{
|
| 1078 |
-
"epoch": 0.03,
|
| 1079 |
-
"learning_rate": 1.997128993715558e-05,
|
| 1080 |
-
"loss": 0.9721,
|
| 1081 |
-
"step": 835
|
| 1082 |
-
},
|
| 1083 |
-
{
|
| 1084 |
-
"epoch": 0.03,
|
| 1085 |
-
"learning_rate": 1.9970799833591133e-05,
|
| 1086 |
-
"loss": 0.9324,
|
| 1087 |
-
"step": 840
|
| 1088 |
-
},
|
| 1089 |
-
{
|
| 1090 |
-
"epoch": 0.03,
|
| 1091 |
-
"learning_rate": 1.997030558824315e-05,
|
| 1092 |
-
"loss": 0.9464,
|
| 1093 |
-
"step": 845
|
| 1094 |
-
},
|
| 1095 |
-
{
|
| 1096 |
-
"epoch": 0.03,
|
| 1097 |
-
"learning_rate": 1.996980720131693e-05,
|
| 1098 |
-
"loss": 0.9312,
|
| 1099 |
-
"step": 850
|
| 1100 |
-
},
|
| 1101 |
-
{
|
| 1102 |
-
"epoch": 0.03,
|
| 1103 |
-
"learning_rate": 1.9969304673019498e-05,
|
| 1104 |
-
"loss": 0.959,
|
| 1105 |
-
"step": 855
|
| 1106 |
-
},
|
| 1107 |
-
{
|
| 1108 |
-
"epoch": 0.03,
|
| 1109 |
-
"learning_rate": 1.9968798003559604e-05,
|
| 1110 |
-
"loss": 0.9742,
|
| 1111 |
-
"step": 860
|
| 1112 |
-
},
|
| 1113 |
-
{
|
| 1114 |
-
"epoch": 0.04,
|
| 1115 |
-
"learning_rate": 1.9968287193147714e-05,
|
| 1116 |
-
"loss": 0.9542,
|
| 1117 |
-
"step": 865
|
| 1118 |
-
},
|
| 1119 |
-
{
|
| 1120 |
-
"epoch": 0.04,
|
| 1121 |
-
"learning_rate": 1.996777224199601e-05,
|
| 1122 |
-
"loss": 0.9477,
|
| 1123 |
-
"step": 870
|
| 1124 |
-
},
|
| 1125 |
-
{
|
| 1126 |
-
"epoch": 0.04,
|
| 1127 |
-
"learning_rate": 1.99672531503184e-05,
|
| 1128 |
-
"loss": 0.9755,
|
| 1129 |
-
"step": 875
|
| 1130 |
-
},
|
| 1131 |
-
{
|
| 1132 |
-
"epoch": 0.04,
|
| 1133 |
-
"learning_rate": 1.996672991833051e-05,
|
| 1134 |
-
"loss": 0.9992,
|
| 1135 |
-
"step": 880
|
| 1136 |
-
},
|
| 1137 |
-
{
|
| 1138 |
-
"epoch": 0.04,
|
| 1139 |
-
"learning_rate": 1.9966202546249692e-05,
|
| 1140 |
-
"loss": 0.9502,
|
| 1141 |
-
"step": 885
|
| 1142 |
-
},
|
| 1143 |
-
{
|
| 1144 |
-
"epoch": 0.04,
|
| 1145 |
-
"learning_rate": 1.9965671034295004e-05,
|
| 1146 |
-
"loss": 0.984,
|
| 1147 |
-
"step": 890
|
| 1148 |
-
},
|
| 1149 |
-
{
|
| 1150 |
-
"epoch": 0.04,
|
| 1151 |
-
"learning_rate": 1.9965135382687235e-05,
|
| 1152 |
-
"loss": 1.0316,
|
| 1153 |
-
"step": 895
|
| 1154 |
-
},
|
| 1155 |
-
{
|
| 1156 |
-
"epoch": 0.04,
|
| 1157 |
-
"learning_rate": 1.9964595591648887e-05,
|
| 1158 |
-
"loss": 0.9855,
|
| 1159 |
-
"step": 900
|
| 1160 |
-
},
|
| 1161 |
-
{
|
| 1162 |
-
"epoch": 0.04,
|
| 1163 |
-
"eval_loss": 0.9192684888839722,
|
| 1164 |
-
"eval_runtime": 138.0438,
|
| 1165 |
-
"eval_samples_per_second": 17.139,
|
| 1166 |
-
"eval_steps_per_second": 2.861,
|
| 1167 |
-
"step": 900
|
| 1168 |
-
},
|
| 1169 |
-
{
|
| 1170 |
-
"epoch": 0.04,
|
| 1171 |
-
"learning_rate": 1.9964051661404186e-05,
|
| 1172 |
-
"loss": 0.9267,
|
| 1173 |
-
"step": 905
|
| 1174 |
-
},
|
| 1175 |
-
{
|
| 1176 |
-
"epoch": 0.04,
|
| 1177 |
-
"learning_rate": 1.9963503592179078e-05,
|
| 1178 |
-
"loss": 0.96,
|
| 1179 |
-
"step": 910
|
| 1180 |
-
},
|
| 1181 |
-
{
|
| 1182 |
-
"epoch": 0.04,
|
| 1183 |
-
"learning_rate": 1.9962951384201227e-05,
|
| 1184 |
-
"loss": 1.0012,
|
| 1185 |
-
"step": 915
|
| 1186 |
-
},
|
| 1187 |
-
{
|
| 1188 |
-
"epoch": 0.04,
|
| 1189 |
-
"learning_rate": 1.996239503770001e-05,
|
| 1190 |
-
"loss": 0.9984,
|
| 1191 |
-
"step": 920
|
| 1192 |
-
},
|
| 1193 |
-
{
|
| 1194 |
-
"epoch": 0.04,
|
| 1195 |
-
"learning_rate": 1.9961834552906532e-05,
|
| 1196 |
-
"loss": 0.9311,
|
| 1197 |
-
"step": 925
|
| 1198 |
-
},
|
| 1199 |
-
{
|
| 1200 |
-
"epoch": 0.04,
|
| 1201 |
-
"learning_rate": 1.9961269930053614e-05,
|
| 1202 |
-
"loss": 0.9479,
|
| 1203 |
-
"step": 930
|
| 1204 |
-
},
|
| 1205 |
-
{
|
| 1206 |
-
"epoch": 0.04,
|
| 1207 |
-
"learning_rate": 1.9960701169375794e-05,
|
| 1208 |
-
"loss": 0.9597,
|
| 1209 |
-
"step": 935
|
| 1210 |
-
},
|
| 1211 |
-
{
|
| 1212 |
-
"epoch": 0.04,
|
| 1213 |
-
"learning_rate": 1.9960128271109326e-05,
|
| 1214 |
-
"loss": 0.9303,
|
| 1215 |
-
"step": 940
|
| 1216 |
-
},
|
| 1217 |
-
{
|
| 1218 |
-
"epoch": 0.04,
|
| 1219 |
-
"learning_rate": 1.99595512354922e-05,
|
| 1220 |
-
"loss": 0.9776,
|
| 1221 |
-
"step": 945
|
| 1222 |
-
},
|
| 1223 |
-
{
|
| 1224 |
-
"epoch": 0.04,
|
| 1225 |
-
"learning_rate": 1.99589700627641e-05,
|
| 1226 |
-
"loss": 0.9779,
|
| 1227 |
-
"step": 950
|
| 1228 |
-
},
|
| 1229 |
-
{
|
| 1230 |
-
"epoch": 0.04,
|
| 1231 |
-
"learning_rate": 1.995838475316644e-05,
|
| 1232 |
-
"loss": 0.9363,
|
| 1233 |
-
"step": 955
|
| 1234 |
-
},
|
| 1235 |
-
{
|
| 1236 |
-
"epoch": 0.04,
|
| 1237 |
-
"learning_rate": 1.9957795306942362e-05,
|
| 1238 |
-
"loss": 0.9147,
|
| 1239 |
-
"step": 960
|
| 1240 |
-
},
|
| 1241 |
-
{
|
| 1242 |
-
"epoch": 0.04,
|
| 1243 |
-
"learning_rate": 1.9957201724336707e-05,
|
| 1244 |
-
"loss": 0.9506,
|
| 1245 |
-
"step": 965
|
| 1246 |
-
},
|
| 1247 |
-
{
|
| 1248 |
-
"epoch": 0.04,
|
| 1249 |
-
"learning_rate": 1.9956604005596047e-05,
|
| 1250 |
-
"loss": 0.9331,
|
| 1251 |
-
"step": 970
|
| 1252 |
-
},
|
| 1253 |
-
{
|
| 1254 |
-
"epoch": 0.04,
|
| 1255 |
-
"learning_rate": 1.9956002150968673e-05,
|
| 1256 |
-
"loss": 0.9433,
|
| 1257 |
-
"step": 975
|
| 1258 |
-
},
|
| 1259 |
-
{
|
| 1260 |
-
"epoch": 0.04,
|
| 1261 |
-
"learning_rate": 1.9955396160704584e-05,
|
| 1262 |
-
"loss": 0.9209,
|
| 1263 |
-
"step": 980
|
| 1264 |
-
},
|
| 1265 |
-
{
|
| 1266 |
-
"epoch": 0.04,
|
| 1267 |
-
"learning_rate": 1.9954786035055505e-05,
|
| 1268 |
-
"loss": 0.9152,
|
| 1269 |
-
"step": 985
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 0.04,
|
| 1273 |
-
"learning_rate": 1.995417177427488e-05,
|
| 1274 |
-
"loss": 0.9596,
|
| 1275 |
-
"step": 990
|
| 1276 |
-
},
|
| 1277 |
-
{
|
| 1278 |
-
"epoch": 0.04,
|
| 1279 |
-
"learning_rate": 1.9953553378617866e-05,
|
| 1280 |
-
"loss": 0.9222,
|
| 1281 |
-
"step": 995
|
| 1282 |
-
},
|
| 1283 |
-
{
|
| 1284 |
-
"epoch": 0.04,
|
| 1285 |
-
"learning_rate": 1.995293084834134e-05,
|
| 1286 |
-
"loss": 0.9551,
|
| 1287 |
-
"step": 1000
|
| 1288 |
-
},
|
| 1289 |
-
{
|
| 1290 |
-
"epoch": 0.04,
|
| 1291 |
-
"eval_loss": 0.9189639687538147,
|
| 1292 |
-
"eval_runtime": 137.9584,
|
| 1293 |
-
"eval_samples_per_second": 17.15,
|
| 1294 |
-
"eval_steps_per_second": 2.863,
|
| 1295 |
-
"step": 1000
|
| 1296 |
-
},
|
| 1297 |
-
{
|
| 1298 |
-
"epoch": 0.04,
|
| 1299 |
-
"learning_rate": 1.9952304183703895e-05,
|
| 1300 |
-
"loss": 0.954,
|
| 1301 |
-
"step": 1005
|
| 1302 |
-
},
|
| 1303 |
-
{
|
| 1304 |
-
"epoch": 0.04,
|
| 1305 |
-
"learning_rate": 1.995167338496584e-05,
|
| 1306 |
-
"loss": 0.9456,
|
| 1307 |
-
"step": 1010
|
| 1308 |
-
},
|
| 1309 |
-
{
|
| 1310 |
-
"epoch": 0.04,
|
| 1311 |
-
"learning_rate": 1.9951038452389204e-05,
|
| 1312 |
-
"loss": 1.0162,
|
| 1313 |
-
"step": 1015
|
| 1314 |
-
},
|
| 1315 |
-
{
|
| 1316 |
-
"epoch": 0.04,
|
| 1317 |
-
"learning_rate": 1.9950399386237732e-05,
|
| 1318 |
-
"loss": 0.9734,
|
| 1319 |
-
"step": 1020
|
| 1320 |
-
},
|
| 1321 |
-
{
|
| 1322 |
-
"epoch": 0.04,
|
| 1323 |
-
"learning_rate": 1.9949756186776894e-05,
|
| 1324 |
-
"loss": 0.9695,
|
| 1325 |
-
"step": 1025
|
| 1326 |
-
},
|
| 1327 |
-
{
|
| 1328 |
-
"epoch": 0.04,
|
| 1329 |
-
"learning_rate": 1.9949108854273856e-05,
|
| 1330 |
-
"loss": 0.9703,
|
| 1331 |
-
"step": 1030
|
| 1332 |
-
},
|
| 1333 |
-
{
|
| 1334 |
-
"epoch": 0.04,
|
| 1335 |
-
"learning_rate": 1.994845738899753e-05,
|
| 1336 |
-
"loss": 0.9394,
|
| 1337 |
-
"step": 1035
|
| 1338 |
-
},
|
| 1339 |
-
{
|
| 1340 |
-
"epoch": 0.04,
|
| 1341 |
-
"learning_rate": 1.9947801791218513e-05,
|
| 1342 |
-
"loss": 1.0069,
|
| 1343 |
-
"step": 1040
|
| 1344 |
-
},
|
| 1345 |
-
{
|
| 1346 |
-
"epoch": 0.04,
|
| 1347 |
-
"learning_rate": 1.9947142061209144e-05,
|
| 1348 |
-
"loss": 0.9777,
|
| 1349 |
-
"step": 1045
|
| 1350 |
-
},
|
| 1351 |
-
{
|
| 1352 |
-
"epoch": 0.04,
|
| 1353 |
-
"learning_rate": 1.994647819924347e-05,
|
| 1354 |
-
"loss": 0.9654,
|
| 1355 |
-
"step": 1050
|
| 1356 |
-
},
|
| 1357 |
-
{
|
| 1358 |
-
"epoch": 0.04,
|
| 1359 |
-
"learning_rate": 1.994581020559725e-05,
|
| 1360 |
-
"loss": 0.9515,
|
| 1361 |
-
"step": 1055
|
| 1362 |
-
},
|
| 1363 |
-
{
|
| 1364 |
-
"epoch": 0.04,
|
| 1365 |
-
"learning_rate": 1.994513808054796e-05,
|
| 1366 |
-
"loss": 0.9378,
|
| 1367 |
-
"step": 1060
|
| 1368 |
-
},
|
| 1369 |
-
{
|
| 1370 |
-
"epoch": 0.04,
|
| 1371 |
-
"learning_rate": 1.9944461824374804e-05,
|
| 1372 |
-
"loss": 0.9557,
|
| 1373 |
-
"step": 1065
|
| 1374 |
-
},
|
| 1375 |
-
{
|
| 1376 |
-
"epoch": 0.04,
|
| 1377 |
-
"learning_rate": 1.994378143735868e-05,
|
| 1378 |
-
"loss": 0.9409,
|
| 1379 |
-
"step": 1070
|
| 1380 |
-
},
|
| 1381 |
-
{
|
| 1382 |
-
"epoch": 0.04,
|
| 1383 |
-
"learning_rate": 1.9943096919782227e-05,
|
| 1384 |
-
"loss": 0.9353,
|
| 1385 |
-
"step": 1075
|
| 1386 |
-
},
|
| 1387 |
-
{
|
| 1388 |
-
"epoch": 0.04,
|
| 1389 |
-
"learning_rate": 1.9942408271929786e-05,
|
| 1390 |
-
"loss": 0.9319,
|
| 1391 |
-
"step": 1080
|
| 1392 |
-
},
|
| 1393 |
-
{
|
| 1394 |
-
"epoch": 0.04,
|
| 1395 |
-
"learning_rate": 1.9941715494087407e-05,
|
| 1396 |
-
"loss": 0.9763,
|
| 1397 |
-
"step": 1085
|
| 1398 |
-
},
|
| 1399 |
-
{
|
| 1400 |
-
"epoch": 0.04,
|
| 1401 |
-
"learning_rate": 1.9941018586542872e-05,
|
| 1402 |
-
"loss": 0.9323,
|
| 1403 |
-
"step": 1090
|
| 1404 |
-
},
|
| 1405 |
-
{
|
| 1406 |
-
"epoch": 0.04,
|
| 1407 |
-
"learning_rate": 1.9940317549585665e-05,
|
| 1408 |
-
"loss": 0.9712,
|
| 1409 |
-
"step": 1095
|
| 1410 |
-
},
|
| 1411 |
-
{
|
| 1412 |
-
"epoch": 0.04,
|
| 1413 |
-
"learning_rate": 1.9939612383506995e-05,
|
| 1414 |
-
"loss": 1.0219,
|
| 1415 |
-
"step": 1100
|
| 1416 |
-
},
|
| 1417 |
-
{
|
| 1418 |
-
"epoch": 0.04,
|
| 1419 |
-
"eval_loss": 0.923427939414978,
|
| 1420 |
-
"eval_runtime": 138.139,
|
| 1421 |
-
"eval_samples_per_second": 17.128,
|
| 1422 |
-
"eval_steps_per_second": 2.859,
|
| 1423 |
-
"step": 1100
|
| 1424 |
-
},
|
| 1425 |
-
{
|
| 1426 |
-
"epoch": 0.04,
|
| 1427 |
-
"learning_rate": 1.993890308859978e-05,
|
| 1428 |
-
"loss": 0.9687,
|
| 1429 |
-
"step": 1105
|
| 1430 |
-
},
|
| 1431 |
-
{
|
| 1432 |
-
"epoch": 0.05,
|
| 1433 |
-
"learning_rate": 1.9938189665158654e-05,
|
| 1434 |
-
"loss": 0.9286,
|
| 1435 |
-
"step": 1110
|
| 1436 |
-
},
|
| 1437 |
-
{
|
| 1438 |
-
"epoch": 0.05,
|
| 1439 |
-
"learning_rate": 1.993747211347997e-05,
|
| 1440 |
-
"loss": 0.9519,
|
| 1441 |
-
"step": 1115
|
| 1442 |
-
},
|
| 1443 |
-
{
|
| 1444 |
-
"epoch": 0.05,
|
| 1445 |
-
"learning_rate": 1.993675043386179e-05,
|
| 1446 |
-
"loss": 0.984,
|
| 1447 |
-
"step": 1120
|
| 1448 |
-
},
|
| 1449 |
-
{
|
| 1450 |
-
"epoch": 0.05,
|
| 1451 |
-
"learning_rate": 1.9936024626603894e-05,
|
| 1452 |
-
"loss": 0.9435,
|
| 1453 |
-
"step": 1125
|
| 1454 |
-
},
|
| 1455 |
-
{
|
| 1456 |
-
"epoch": 0.05,
|
| 1457 |
-
"learning_rate": 1.9935294692007773e-05,
|
| 1458 |
-
"loss": 0.9432,
|
| 1459 |
-
"step": 1130
|
| 1460 |
-
},
|
| 1461 |
-
{
|
| 1462 |
-
"epoch": 0.05,
|
| 1463 |
-
"learning_rate": 1.9934560630376642e-05,
|
| 1464 |
-
"loss": 0.9505,
|
| 1465 |
-
"step": 1135
|
| 1466 |
-
},
|
| 1467 |
-
{
|
| 1468 |
-
"epoch": 0.05,
|
| 1469 |
-
"learning_rate": 1.9933822442015418e-05,
|
| 1470 |
-
"loss": 0.9747,
|
| 1471 |
-
"step": 1140
|
| 1472 |
-
},
|
| 1473 |
-
{
|
| 1474 |
-
"epoch": 0.05,
|
| 1475 |
-
"learning_rate": 1.993308012723074e-05,
|
| 1476 |
-
"loss": 0.9539,
|
| 1477 |
-
"step": 1145
|
| 1478 |
-
},
|
| 1479 |
-
{
|
| 1480 |
-
"epoch": 0.05,
|
| 1481 |
-
"learning_rate": 1.9932333686330962e-05,
|
| 1482 |
-
"loss": 0.9309,
|
| 1483 |
-
"step": 1150
|
| 1484 |
-
},
|
| 1485 |
-
{
|
| 1486 |
-
"epoch": 0.05,
|
| 1487 |
-
"learning_rate": 1.9931583119626142e-05,
|
| 1488 |
-
"loss": 0.9459,
|
| 1489 |
-
"step": 1155
|
| 1490 |
-
},
|
| 1491 |
-
{
|
| 1492 |
-
"epoch": 0.05,
|
| 1493 |
-
"learning_rate": 1.993082842742807e-05,
|
| 1494 |
-
"loss": 0.9511,
|
| 1495 |
-
"step": 1160
|
| 1496 |
-
},
|
| 1497 |
-
{
|
| 1498 |
-
"epoch": 0.05,
|
| 1499 |
-
"learning_rate": 1.9930069610050228e-05,
|
| 1500 |
-
"loss": 0.9623,
|
| 1501 |
-
"step": 1165
|
| 1502 |
-
},
|
| 1503 |
-
{
|
| 1504 |
-
"epoch": 0.05,
|
| 1505 |
-
"learning_rate": 1.9929306667807825e-05,
|
| 1506 |
-
"loss": 0.9476,
|
| 1507 |
-
"step": 1170
|
| 1508 |
-
},
|
| 1509 |
-
{
|
| 1510 |
-
"epoch": 0.05,
|
| 1511 |
-
"learning_rate": 1.9928539601017785e-05,
|
| 1512 |
-
"loss": 0.9104,
|
| 1513 |
-
"step": 1175
|
| 1514 |
-
},
|
| 1515 |
-
{
|
| 1516 |
-
"epoch": 0.05,
|
| 1517 |
-
"learning_rate": 1.9927768409998734e-05,
|
| 1518 |
-
"loss": 0.9787,
|
| 1519 |
-
"step": 1180
|
| 1520 |
-
},
|
| 1521 |
-
{
|
| 1522 |
-
"epoch": 0.05,
|
| 1523 |
-
"learning_rate": 1.9926993095071024e-05,
|
| 1524 |
-
"loss": 0.9112,
|
| 1525 |
-
"step": 1185
|
| 1526 |
-
},
|
| 1527 |
-
{
|
| 1528 |
-
"epoch": 0.05,
|
| 1529 |
-
"learning_rate": 1.992621365655671e-05,
|
| 1530 |
-
"loss": 0.9624,
|
| 1531 |
-
"step": 1190
|
| 1532 |
-
},
|
| 1533 |
-
{
|
| 1534 |
-
"epoch": 0.05,
|
| 1535 |
-
"learning_rate": 1.9925430094779568e-05,
|
| 1536 |
-
"loss": 0.9269,
|
| 1537 |
-
"step": 1195
|
| 1538 |
-
},
|
| 1539 |
-
{
|
| 1540 |
-
"epoch": 0.05,
|
| 1541 |
-
"learning_rate": 1.992464241006508e-05,
|
| 1542 |
-
"loss": 0.9909,
|
| 1543 |
-
"step": 1200
|
| 1544 |
-
},
|
| 1545 |
-
{
|
| 1546 |
-
"epoch": 0.05,
|
| 1547 |
-
"eval_loss": 0.9236459136009216,
|
| 1548 |
-
"eval_runtime": 137.9974,
|
| 1549 |
-
"eval_samples_per_second": 17.145,
|
| 1550 |
-
"eval_steps_per_second": 2.862,
|
| 1551 |
-
"step": 1200
|
| 1552 |
-
},
|
| 1553 |
-
{
|
| 1554 |
-
"epoch": 0.05,
|
| 1555 |
-
"learning_rate": 1.9923850602740443e-05,
|
| 1556 |
-
"loss": 0.9497,
|
| 1557 |
-
"step": 1205
|
| 1558 |
-
},
|
| 1559 |
-
{
|
| 1560 |
-
"epoch": 0.05,
|
| 1561 |
-
"learning_rate": 1.9923054673134566e-05,
|
| 1562 |
-
"loss": 1.0494,
|
| 1563 |
-
"step": 1210
|
| 1564 |
-
},
|
| 1565 |
-
{
|
| 1566 |
-
"epoch": 0.05,
|
| 1567 |
-
"learning_rate": 1.992225462157807e-05,
|
| 1568 |
-
"loss": 0.9503,
|
| 1569 |
-
"step": 1215
|
| 1570 |
-
},
|
| 1571 |
-
{
|
| 1572 |
-
"epoch": 0.05,
|
| 1573 |
-
"learning_rate": 1.99214504484033e-05,
|
| 1574 |
-
"loss": 0.9844,
|
| 1575 |
-
"step": 1220
|
| 1576 |
-
},
|
| 1577 |
-
{
|
| 1578 |
-
"epoch": 0.05,
|
| 1579 |
-
"learning_rate": 1.9920642153944293e-05,
|
| 1580 |
-
"loss": 0.9681,
|
| 1581 |
-
"step": 1225
|
| 1582 |
-
},
|
| 1583 |
-
{
|
| 1584 |
-
"epoch": 0.05,
|
| 1585 |
-
"learning_rate": 1.9919829738536807e-05,
|
| 1586 |
-
"loss": 0.9266,
|
| 1587 |
-
"step": 1230
|
| 1588 |
-
},
|
| 1589 |
-
{
|
| 1590 |
-
"epoch": 0.05,
|
| 1591 |
-
"learning_rate": 1.9919013202518313e-05,
|
| 1592 |
-
"loss": 0.9289,
|
| 1593 |
-
"step": 1235
|
| 1594 |
-
},
|
| 1595 |
-
{
|
| 1596 |
-
"epoch": 0.05,
|
| 1597 |
-
"learning_rate": 1.9918192546228e-05,
|
| 1598 |
-
"loss": 0.9878,
|
| 1599 |
-
"step": 1240
|
| 1600 |
-
},
|
| 1601 |
-
{
|
| 1602 |
-
"epoch": 0.05,
|
| 1603 |
-
"learning_rate": 1.991736777000675e-05,
|
| 1604 |
-
"loss": 0.9889,
|
| 1605 |
-
"step": 1245
|
| 1606 |
-
},
|
| 1607 |
-
{
|
| 1608 |
-
"epoch": 0.05,
|
| 1609 |
-
"learning_rate": 1.9916538874197177e-05,
|
| 1610 |
-
"loss": 1.0111,
|
| 1611 |
-
"step": 1250
|
| 1612 |
-
},
|
| 1613 |
-
{
|
| 1614 |
-
"epoch": 0.05,
|
| 1615 |
-
"learning_rate": 1.9915705859143597e-05,
|
| 1616 |
-
"loss": 0.9014,
|
| 1617 |
-
"step": 1255
|
| 1618 |
-
},
|
| 1619 |
-
{
|
| 1620 |
-
"epoch": 0.05,
|
| 1621 |
-
"learning_rate": 1.991486872519203e-05,
|
| 1622 |
-
"loss": 0.939,
|
| 1623 |
-
"step": 1260
|
| 1624 |
-
},
|
| 1625 |
-
{
|
| 1626 |
-
"epoch": 0.05,
|
| 1627 |
-
"learning_rate": 1.991402747269022e-05,
|
| 1628 |
-
"loss": 0.9945,
|
| 1629 |
-
"step": 1265
|
| 1630 |
-
},
|
| 1631 |
-
{
|
| 1632 |
-
"epoch": 0.05,
|
| 1633 |
-
"learning_rate": 1.9913182101987614e-05,
|
| 1634 |
-
"loss": 0.9495,
|
| 1635 |
-
"step": 1270
|
| 1636 |
-
},
|
| 1637 |
-
{
|
| 1638 |
-
"epoch": 0.05,
|
| 1639 |
-
"learning_rate": 1.9912332613435375e-05,
|
| 1640 |
-
"loss": 0.9089,
|
| 1641 |
-
"step": 1275
|
| 1642 |
-
},
|
| 1643 |
-
{
|
| 1644 |
-
"epoch": 0.05,
|
| 1645 |
-
"learning_rate": 1.9911479007386364e-05,
|
| 1646 |
-
"loss": 0.9804,
|
| 1647 |
-
"step": 1280
|
| 1648 |
-
},
|
| 1649 |
-
{
|
| 1650 |
-
"epoch": 0.05,
|
| 1651 |
-
"learning_rate": 1.9910621284195173e-05,
|
| 1652 |
-
"loss": 0.9973,
|
| 1653 |
-
"step": 1285
|
| 1654 |
-
},
|
| 1655 |
-
{
|
| 1656 |
-
"epoch": 0.05,
|
| 1657 |
-
"learning_rate": 1.9909759444218086e-05,
|
| 1658 |
-
"loss": 0.9714,
|
| 1659 |
-
"step": 1290
|
| 1660 |
-
},
|
| 1661 |
-
{
|
| 1662 |
-
"epoch": 0.05,
|
| 1663 |
-
"learning_rate": 1.9908893487813107e-05,
|
| 1664 |
-
"loss": 0.9559,
|
| 1665 |
-
"step": 1295
|
| 1666 |
-
},
|
| 1667 |
-
{
|
| 1668 |
-
"epoch": 0.05,
|
| 1669 |
-
"learning_rate": 1.9908023415339942e-05,
|
| 1670 |
-
"loss": 0.9792,
|
| 1671 |
-
"step": 1300
|
| 1672 |
-
},
|
| 1673 |
-
{
|
| 1674 |
-
"epoch": 0.05,
|
| 1675 |
-
"eval_loss": 0.9266601800918579,
|
| 1676 |
-
"eval_runtime": 138.2041,
|
| 1677 |
-
"eval_samples_per_second": 17.12,
|
| 1678 |
-
"eval_steps_per_second": 2.858,
|
| 1679 |
-
"step": 1300
|
| 1680 |
-
},
|
| 1681 |
-
{
|
| 1682 |
-
"epoch": 0.05,
|
| 1683 |
-
"learning_rate": 1.9907149227160015e-05,
|
| 1684 |
-
"loss": 0.9514,
|
| 1685 |
-
"step": 1305
|
| 1686 |
-
},
|
| 1687 |
-
{
|
| 1688 |
-
"epoch": 0.05,
|
| 1689 |
-
"learning_rate": 1.990627092363646e-05,
|
| 1690 |
-
"loss": 0.9752,
|
| 1691 |
-
"step": 1310
|
| 1692 |
-
},
|
| 1693 |
-
{
|
| 1694 |
-
"epoch": 0.05,
|
| 1695 |
-
"learning_rate": 1.990538850513411e-05,
|
| 1696 |
-
"loss": 0.9322,
|
| 1697 |
-
"step": 1315
|
| 1698 |
-
},
|
| 1699 |
-
{
|
| 1700 |
-
"epoch": 0.05,
|
| 1701 |
-
"learning_rate": 1.9904501972019515e-05,
|
| 1702 |
-
"loss": 0.9474,
|
| 1703 |
-
"step": 1320
|
| 1704 |
-
},
|
| 1705 |
-
{
|
| 1706 |
-
"epoch": 0.05,
|
| 1707 |
-
"learning_rate": 1.9903611324660933e-05,
|
| 1708 |
-
"loss": 0.9509,
|
| 1709 |
-
"step": 1325
|
| 1710 |
-
},
|
| 1711 |
-
{
|
| 1712 |
-
"epoch": 0.05,
|
| 1713 |
-
"learning_rate": 1.9902716563428338e-05,
|
| 1714 |
-
"loss": 0.9716,
|
| 1715 |
-
"step": 1330
|
| 1716 |
-
},
|
| 1717 |
-
{
|
| 1718 |
-
"epoch": 0.05,
|
| 1719 |
-
"learning_rate": 1.9901817688693395e-05,
|
| 1720 |
-
"loss": 0.95,
|
| 1721 |
-
"step": 1335
|
| 1722 |
-
},
|
| 1723 |
-
{
|
| 1724 |
-
"epoch": 0.05,
|
| 1725 |
-
"learning_rate": 1.99009147008295e-05,
|
| 1726 |
-
"loss": 0.9852,
|
| 1727 |
-
"step": 1340
|
| 1728 |
-
},
|
| 1729 |
-
{
|
| 1730 |
-
"epoch": 0.05,
|
| 1731 |
-
"learning_rate": 1.990000760021174e-05,
|
| 1732 |
-
"loss": 0.8945,
|
| 1733 |
-
"step": 1345
|
| 1734 |
-
},
|
| 1735 |
-
{
|
| 1736 |
-
"epoch": 0.05,
|
| 1737 |
-
"learning_rate": 1.9899096387216914e-05,
|
| 1738 |
-
"loss": 0.998,
|
| 1739 |
-
"step": 1350
|
| 1740 |
-
},
|
| 1741 |
-
{
|
| 1742 |
-
"epoch": 0.06,
|
| 1743 |
-
"learning_rate": 1.989818106222354e-05,
|
| 1744 |
-
"loss": 0.9691,
|
| 1745 |
-
"step": 1355
|
| 1746 |
-
},
|
| 1747 |
-
{
|
| 1748 |
-
"epoch": 0.06,
|
| 1749 |
-
"learning_rate": 1.9897261625611826e-05,
|
| 1750 |
-
"loss": 0.9284,
|
| 1751 |
-
"step": 1360
|
| 1752 |
-
},
|
| 1753 |
-
{
|
| 1754 |
-
"epoch": 0.06,
|
| 1755 |
-
"learning_rate": 1.9896338077763708e-05,
|
| 1756 |
-
"loss": 0.942,
|
| 1757 |
-
"step": 1365
|
| 1758 |
-
},
|
| 1759 |
-
{
|
| 1760 |
-
"epoch": 0.06,
|
| 1761 |
-
"learning_rate": 1.989541041906281e-05,
|
| 1762 |
-
"loss": 0.9607,
|
| 1763 |
-
"step": 1370
|
| 1764 |
-
},
|
| 1765 |
-
{
|
| 1766 |
-
"epoch": 0.06,
|
| 1767 |
-
"learning_rate": 1.9894478649894483e-05,
|
| 1768 |
-
"loss": 0.9424,
|
| 1769 |
-
"step": 1375
|
| 1770 |
-
},
|
| 1771 |
-
{
|
| 1772 |
-
"epoch": 0.06,
|
| 1773 |
-
"learning_rate": 1.989354277064577e-05,
|
| 1774 |
-
"loss": 0.9869,
|
| 1775 |
-
"step": 1380
|
| 1776 |
-
},
|
| 1777 |
-
{
|
| 1778 |
-
"epoch": 0.06,
|
| 1779 |
-
"learning_rate": 1.989260278170543e-05,
|
| 1780 |
-
"loss": 0.9437,
|
| 1781 |
-
"step": 1385
|
| 1782 |
-
},
|
| 1783 |
-
{
|
| 1784 |
-
"epoch": 0.06,
|
| 1785 |
-
"learning_rate": 1.9891658683463926e-05,
|
| 1786 |
-
"loss": 0.9921,
|
| 1787 |
-
"step": 1390
|
| 1788 |
-
},
|
| 1789 |
-
{
|
| 1790 |
-
"epoch": 0.06,
|
| 1791 |
-
"learning_rate": 1.9890710476313426e-05,
|
| 1792 |
-
"loss": 0.9039,
|
| 1793 |
-
"step": 1395
|
| 1794 |
-
},
|
| 1795 |
-
{
|
| 1796 |
-
"epoch": 0.06,
|
| 1797 |
-
"learning_rate": 1.9889758160647804e-05,
|
| 1798 |
-
"loss": 0.9356,
|
| 1799 |
-
"step": 1400
|
| 1800 |
-
},
|
| 1801 |
-
{
|
| 1802 |
-
"epoch": 0.06,
|
| 1803 |
-
"eval_loss": 0.9211186170578003,
|
| 1804 |
-
"eval_runtime": 137.9724,
|
| 1805 |
-
"eval_samples_per_second": 17.148,
|
| 1806 |
-
"eval_steps_per_second": 2.863,
|
| 1807 |
-
"step": 1400
|
| 1808 |
-
},
|
| 1809 |
-
{
|
| 1810 |
-
"epoch": 0.06,
|
| 1811 |
-
"learning_rate": 1.988880173686265e-05,
|
| 1812 |
-
"loss": 1.0002,
|
| 1813 |
-
"step": 1405
|
| 1814 |
-
},
|
| 1815 |
-
{
|
| 1816 |
-
"epoch": 0.06,
|
| 1817 |
-
"learning_rate": 1.9887841205355252e-05,
|
| 1818 |
-
"loss": 0.9917,
|
| 1819 |
-
"step": 1410
|
| 1820 |
-
},
|
| 1821 |
-
{
|
| 1822 |
-
"epoch": 0.06,
|
| 1823 |
-
"learning_rate": 1.988687656652461e-05,
|
| 1824 |
-
"loss": 0.9554,
|
| 1825 |
-
"step": 1415
|
| 1826 |
-
},
|
| 1827 |
-
{
|
| 1828 |
-
"epoch": 0.06,
|
| 1829 |
-
"learning_rate": 1.9885907820771417e-05,
|
| 1830 |
-
"loss": 0.9751,
|
| 1831 |
-
"step": 1420
|
| 1832 |
-
},
|
| 1833 |
-
{
|
| 1834 |
-
"epoch": 0.06,
|
| 1835 |
-
"learning_rate": 1.988493496849809e-05,
|
| 1836 |
-
"loss": 0.987,
|
| 1837 |
-
"step": 1425
|
| 1838 |
-
},
|
| 1839 |
-
{
|
| 1840 |
-
"epoch": 0.06,
|
| 1841 |
-
"learning_rate": 1.988395801010874e-05,
|
| 1842 |
-
"loss": 0.9501,
|
| 1843 |
-
"step": 1430
|
| 1844 |
-
},
|
| 1845 |
-
{
|
| 1846 |
-
"epoch": 0.06,
|
| 1847 |
-
"learning_rate": 1.9882976946009188e-05,
|
| 1848 |
-
"loss": 0.9782,
|
| 1849 |
-
"step": 1435
|
| 1850 |
-
},
|
| 1851 |
-
{
|
| 1852 |
-
"epoch": 0.06,
|
| 1853 |
-
"learning_rate": 1.9881991776606958e-05,
|
| 1854 |
-
"loss": 0.8899,
|
| 1855 |
-
"step": 1440
|
| 1856 |
-
},
|
| 1857 |
-
{
|
| 1858 |
-
"epoch": 0.06,
|
| 1859 |
-
"learning_rate": 1.9881002502311287e-05,
|
| 1860 |
-
"loss": 0.8897,
|
| 1861 |
-
"step": 1445
|
| 1862 |
-
},
|
| 1863 |
-
{
|
| 1864 |
-
"epoch": 0.06,
|
| 1865 |
-
"learning_rate": 1.98800091235331e-05,
|
| 1866 |
-
"loss": 0.9558,
|
| 1867 |
-
"step": 1450
|
| 1868 |
-
},
|
| 1869 |
-
{
|
| 1870 |
-
"epoch": 0.06,
|
| 1871 |
-
"learning_rate": 1.9879011640685046e-05,
|
| 1872 |
-
"loss": 0.9678,
|
| 1873 |
-
"step": 1455
|
| 1874 |
-
},
|
| 1875 |
-
{
|
| 1876 |
-
"epoch": 0.06,
|
| 1877 |
-
"learning_rate": 1.9878010054181468e-05,
|
| 1878 |
-
"loss": 0.9887,
|
| 1879 |
-
"step": 1460
|
| 1880 |
-
},
|
| 1881 |
-
{
|
| 1882 |
-
"epoch": 0.06,
|
| 1883 |
-
"learning_rate": 1.9877004364438416e-05,
|
| 1884 |
-
"loss": 0.9494,
|
| 1885 |
-
"step": 1465
|
| 1886 |
-
},
|
| 1887 |
-
{
|
| 1888 |
-
"epoch": 0.06,
|
| 1889 |
-
"learning_rate": 1.9875994571873648e-05,
|
| 1890 |
-
"loss": 0.9784,
|
| 1891 |
-
"step": 1470
|
| 1892 |
-
},
|
| 1893 |
-
{
|
| 1894 |
-
"epoch": 0.06,
|
| 1895 |
-
"learning_rate": 1.9874980676906618e-05,
|
| 1896 |
-
"loss": 0.9538,
|
| 1897 |
-
"step": 1475
|
| 1898 |
-
},
|
| 1899 |
-
{
|
| 1900 |
-
"epoch": 0.06,
|
| 1901 |
-
"learning_rate": 1.9873962679958496e-05,
|
| 1902 |
-
"loss": 0.9269,
|
| 1903 |
-
"step": 1480
|
| 1904 |
-
},
|
| 1905 |
-
{
|
| 1906 |
-
"epoch": 0.06,
|
| 1907 |
-
"learning_rate": 1.987294058145214e-05,
|
| 1908 |
-
"loss": 0.9983,
|
| 1909 |
-
"step": 1485
|
| 1910 |
-
},
|
| 1911 |
-
{
|
| 1912 |
-
"epoch": 0.06,
|
| 1913 |
-
"learning_rate": 1.987191438181213e-05,
|
| 1914 |
-
"loss": 0.994,
|
| 1915 |
-
"step": 1490
|
| 1916 |
-
},
|
| 1917 |
-
{
|
| 1918 |
-
"epoch": 0.06,
|
| 1919 |
-
"learning_rate": 1.987088408146473e-05,
|
| 1920 |
-
"loss": 0.9887,
|
| 1921 |
-
"step": 1495
|
| 1922 |
-
},
|
| 1923 |
-
{
|
| 1924 |
-
"epoch": 0.06,
|
| 1925 |
-
"learning_rate": 1.9869849680837932e-05,
|
| 1926 |
-
"loss": 0.9645,
|
| 1927 |
-
"step": 1500
|
| 1928 |
-
},
|
| 1929 |
-
{
|
| 1930 |
-
"epoch": 0.06,
|
| 1931 |
-
"eval_loss": 0.9189525246620178,
|
| 1932 |
-
"eval_runtime": 138.1629,
|
| 1933 |
-
"eval_samples_per_second": 17.125,
|
| 1934 |
-
"eval_steps_per_second": 2.859,
|
| 1935 |
-
"step": 1500
|
| 1936 |
-
},
|
| 1937 |
-
{
|
| 1938 |
-
"epoch": 0.06,
|
| 1939 |
-
"learning_rate": 1.9868811180361404e-05,
|
| 1940 |
-
"loss": 0.9817,
|
| 1941 |
-
"step": 1505
|
| 1942 |
-
},
|
| 1943 |
-
{
|
| 1944 |
-
"epoch": 0.06,
|
| 1945 |
-
"learning_rate": 1.9867768580466538e-05,
|
| 1946 |
-
"loss": 0.9345,
|
| 1947 |
-
"step": 1510
|
| 1948 |
-
},
|
| 1949 |
-
{
|
| 1950 |
-
"epoch": 0.06,
|
| 1951 |
-
"learning_rate": 1.9866721881586414e-05,
|
| 1952 |
-
"loss": 0.969,
|
| 1953 |
-
"step": 1515
|
| 1954 |
-
},
|
| 1955 |
-
{
|
| 1956 |
-
"epoch": 0.06,
|
| 1957 |
-
"learning_rate": 1.986567108415583e-05,
|
| 1958 |
-
"loss": 0.9457,
|
| 1959 |
-
"step": 1520
|
| 1960 |
-
},
|
| 1961 |
-
{
|
| 1962 |
-
"epoch": 0.06,
|
| 1963 |
-
"learning_rate": 1.9864616188611273e-05,
|
| 1964 |
-
"loss": 0.9405,
|
| 1965 |
-
"step": 1525
|
| 1966 |
-
},
|
| 1967 |
-
{
|
| 1968 |
-
"epoch": 0.06,
|
| 1969 |
-
"learning_rate": 1.9863557195390934e-05,
|
| 1970 |
-
"loss": 0.9762,
|
| 1971 |
-
"step": 1530
|
| 1972 |
-
},
|
| 1973 |
-
{
|
| 1974 |
-
"epoch": 0.06,
|
| 1975 |
-
"learning_rate": 1.9862494104934716e-05,
|
| 1976 |
-
"loss": 0.9651,
|
| 1977 |
-
"step": 1535
|
| 1978 |
-
},
|
| 1979 |
-
{
|
| 1980 |
-
"epoch": 0.06,
|
| 1981 |
-
"learning_rate": 1.9861426917684217e-05,
|
| 1982 |
-
"loss": 0.9591,
|
| 1983 |
-
"step": 1540
|
| 1984 |
-
},
|
| 1985 |
-
{
|
| 1986 |
-
"epoch": 0.06,
|
| 1987 |
-
"learning_rate": 1.9860355634082733e-05,
|
| 1988 |
-
"loss": 0.9416,
|
| 1989 |
-
"step": 1545
|
| 1990 |
-
},
|
| 1991 |
-
{
|
| 1992 |
-
"epoch": 0.06,
|
| 1993 |
-
"learning_rate": 1.985928025457527e-05,
|
| 1994 |
-
"loss": 1.0004,
|
| 1995 |
-
"step": 1550
|
| 1996 |
-
},
|
| 1997 |
-
{
|
| 1998 |
-
"epoch": 0.06,
|
| 1999 |
-
"learning_rate": 1.985820077960853e-05,
|
| 2000 |
-
"loss": 0.97,
|
| 2001 |
-
"step": 1555
|
| 2002 |
-
},
|
| 2003 |
-
{
|
| 2004 |
-
"epoch": 0.06,
|
| 2005 |
-
"learning_rate": 1.9857117209630917e-05,
|
| 2006 |
-
"loss": 0.9593,
|
| 2007 |
-
"step": 1560
|
| 2008 |
-
},
|
| 2009 |
-
{
|
| 2010 |
-
"epoch": 0.06,
|
| 2011 |
-
"learning_rate": 1.9856029545092536e-05,
|
| 2012 |
-
"loss": 0.9151,
|
| 2013 |
-
"step": 1565
|
| 2014 |
-
},
|
| 2015 |
-
{
|
| 2016 |
-
"epoch": 0.06,
|
| 2017 |
-
"learning_rate": 1.9854937786445195e-05,
|
| 2018 |
-
"loss": 0.9333,
|
| 2019 |
-
"step": 1570
|
| 2020 |
-
},
|
| 2021 |
-
{
|
| 2022 |
-
"epoch": 0.06,
|
| 2023 |
-
"learning_rate": 1.98538419341424e-05,
|
| 2024 |
-
"loss": 0.9953,
|
| 2025 |
-
"step": 1575
|
| 2026 |
-
},
|
| 2027 |
-
{
|
| 2028 |
-
"epoch": 0.06,
|
| 2029 |
-
"learning_rate": 1.9852741988639356e-05,
|
| 2030 |
-
"loss": 0.9662,
|
| 2031 |
-
"step": 1580
|
| 2032 |
-
},
|
| 2033 |
-
{
|
| 2034 |
-
"epoch": 0.06,
|
| 2035 |
-
"learning_rate": 1.9851637950392975e-05,
|
| 2036 |
-
"loss": 0.973,
|
| 2037 |
-
"step": 1585
|
| 2038 |
-
},
|
| 2039 |
-
{
|
| 2040 |
-
"epoch": 0.06,
|
| 2041 |
-
"learning_rate": 1.9850529819861867e-05,
|
| 2042 |
-
"loss": 0.9741,
|
| 2043 |
-
"step": 1590
|
| 2044 |
-
},
|
| 2045 |
-
{
|
| 2046 |
-
"epoch": 0.06,
|
| 2047 |
-
"learning_rate": 1.9849417597506335e-05,
|
| 2048 |
-
"loss": 0.9171,
|
| 2049 |
-
"step": 1595
|
| 2050 |
-
},
|
| 2051 |
-
{
|
| 2052 |
-
"epoch": 0.06,
|
| 2053 |
-
"learning_rate": 1.9848301283788384e-05,
|
| 2054 |
-
"loss": 0.9745,
|
| 2055 |
-
"step": 1600
|
| 2056 |
-
},
|
| 2057 |
-
{
|
| 2058 |
-
"epoch": 0.06,
|
| 2059 |
-
"eval_loss": 0.9158608317375183,
|
| 2060 |
-
"eval_runtime": 138.1513,
|
| 2061 |
-
"eval_samples_per_second": 17.126,
|
| 2062 |
-
"eval_steps_per_second": 2.859,
|
| 2063 |
-
"step": 1600
|
| 2064 |
}
|
| 2065 |
],
|
| 2066 |
"logging_steps": 5,
|
|
@@ -2068,7 +260,7 @@
|
|
| 2068 |
"num_input_tokens_seen": 0,
|
| 2069 |
"num_train_epochs": 1,
|
| 2070 |
"save_steps": 200,
|
| 2071 |
-
"total_flos":
|
| 2072 |
"trial_name": null,
|
| 2073 |
"trial_params": null
|
| 2074 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.008123683074814043,
|
| 5 |
+
"eval_steps": 400,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 8.097165991902834e-09,
|
| 14 |
"loss": 3.5463,
|
| 15 |
"step": 1
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 0.0,
|
| 19 |
+
"learning_rate": 4.048582995951417e-08,
|
| 20 |
+
"loss": 3.6591,
|
| 21 |
"step": 5
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
+
"learning_rate": 8.097165991902834e-08,
|
| 26 |
+
"loss": 3.6377,
|
| 27 |
"step": 10
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
+
"learning_rate": 1.214574898785425e-07,
|
| 32 |
+
"loss": 3.3634,
|
| 33 |
"step": 15
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
+
"learning_rate": 1.6194331983805668e-07,
|
| 38 |
+
"loss": 3.069,
|
| 39 |
"step": 20
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
+
"learning_rate": 2.0242914979757083e-07,
|
| 44 |
+
"loss": 2.6403,
|
| 45 |
"step": 25
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
+
"learning_rate": 2.42914979757085e-07,
|
| 50 |
+
"loss": 2.1059,
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
+
"learning_rate": 2.8340080971659917e-07,
|
| 56 |
+
"loss": 1.7116,
|
| 57 |
"step": 35
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.0,
|
| 61 |
+
"learning_rate": 3.2388663967611335e-07,
|
| 62 |
+
"loss": 1.2154,
|
| 63 |
"step": 40
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
+
"learning_rate": 3.6437246963562754e-07,
|
| 68 |
+
"loss": 1.1809,
|
| 69 |
"step": 45
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.0,
|
| 73 |
+
"learning_rate": 4.0485829959514166e-07,
|
| 74 |
+
"loss": 1.1222,
|
| 75 |
"step": 50
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.0,
|
| 79 |
+
"learning_rate": 4.4534412955465585e-07,
|
| 80 |
+
"loss": 1.0215,
|
| 81 |
"step": 55
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.0,
|
| 85 |
+
"learning_rate": 4.8582995951417e-07,
|
| 86 |
+
"loss": 0.9725,
|
| 87 |
"step": 60
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.0,
|
| 91 |
+
"learning_rate": 5.263157894736842e-07,
|
| 92 |
+
"loss": 0.9506,
|
| 93 |
"step": 65
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.0,
|
| 97 |
+
"learning_rate": 5.668016194331983e-07,
|
| 98 |
+
"loss": 0.8981,
|
| 99 |
"step": 70
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"epoch": 0.0,
|
| 103 |
+
"learning_rate": 6.072874493927125e-07,
|
| 104 |
+
"loss": 0.8955,
|
| 105 |
"step": 75
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.0,
|
| 109 |
+
"learning_rate": 6.477732793522267e-07,
|
| 110 |
+
"loss": 0.8819,
|
| 111 |
"step": 80
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 0.0,
|
| 115 |
+
"learning_rate": 6.882591093117408e-07,
|
| 116 |
+
"loss": 0.9384,
|
| 117 |
"step": 85
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"epoch": 0.0,
|
| 121 |
+
"learning_rate": 7.287449392712551e-07,
|
| 122 |
+
"loss": 0.9053,
|
| 123 |
"step": 90
|
| 124 |
},
|
| 125 |
{
|
| 126 |
"epoch": 0.0,
|
| 127 |
+
"learning_rate": 7.692307692307693e-07,
|
| 128 |
+
"loss": 0.8439,
|
| 129 |
"step": 95
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 0.0,
|
| 133 |
+
"learning_rate": 8.097165991902833e-07,
|
| 134 |
+
"loss": 0.8176,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 135 |
"step": 100
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.0,
|
| 139 |
+
"learning_rate": 8.502024291497975e-07,
|
| 140 |
+
"loss": 0.8403,
|
| 141 |
"step": 105
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 0.0,
|
| 145 |
+
"learning_rate": 8.906882591093117e-07,
|
| 146 |
+
"loss": 0.8365,
|
| 147 |
"step": 110
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 0.0,
|
| 151 |
+
"learning_rate": 9.311740890688259e-07,
|
| 152 |
+
"loss": 0.8354,
|
| 153 |
"step": 115
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.0,
|
| 157 |
+
"learning_rate": 9.7165991902834e-07,
|
| 158 |
+
"loss": 0.7901,
|
| 159 |
"step": 120
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 0.01,
|
| 163 |
+
"learning_rate": 1.0121457489878542e-06,
|
| 164 |
+
"loss": 0.8182,
|
| 165 |
"step": 125
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 0.01,
|
| 169 |
+
"learning_rate": 1.0526315789473683e-06,
|
| 170 |
+
"loss": 0.8021,
|
| 171 |
"step": 130
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 0.01,
|
| 175 |
+
"learning_rate": 1.0931174089068826e-06,
|
| 176 |
+
"loss": 0.7999,
|
| 177 |
"step": 135
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.01,
|
| 181 |
+
"learning_rate": 1.1336032388663967e-06,
|
| 182 |
+
"loss": 0.7896,
|
| 183 |
"step": 140
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 0.01,
|
| 187 |
+
"learning_rate": 1.1740890688259108e-06,
|
| 188 |
+
"loss": 0.8225,
|
| 189 |
"step": 145
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 0.01,
|
| 193 |
+
"learning_rate": 1.214574898785425e-06,
|
| 194 |
+
"loss": 0.8044,
|
| 195 |
"step": 150
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 0.01,
|
| 199 |
+
"learning_rate": 1.2550607287449393e-06,
|
| 200 |
+
"loss": 0.7872,
|
| 201 |
"step": 155
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.01,
|
| 205 |
+
"learning_rate": 1.2955465587044534e-06,
|
| 206 |
+
"loss": 0.826,
|
| 207 |
"step": 160
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 0.01,
|
| 211 |
+
"learning_rate": 1.3360323886639675e-06,
|
| 212 |
+
"loss": 0.772,
|
| 213 |
"step": 165
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 0.01,
|
| 217 |
+
"learning_rate": 1.3765182186234816e-06,
|
| 218 |
+
"loss": 0.7842,
|
| 219 |
"step": 170
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.01,
|
| 223 |
+
"learning_rate": 1.4170040485829959e-06,
|
| 224 |
+
"loss": 0.7833,
|
| 225 |
"step": 175
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 0.01,
|
| 229 |
+
"learning_rate": 1.4574898785425101e-06,
|
| 230 |
+
"loss": 0.7697,
|
| 231 |
"step": 180
|
| 232 |
},
|
| 233 |
{
|
| 234 |
"epoch": 0.01,
|
| 235 |
+
"learning_rate": 1.4979757085020242e-06,
|
| 236 |
+
"loss": 0.7515,
|
| 237 |
"step": 185
|
| 238 |
},
|
| 239 |
{
|
| 240 |
"epoch": 0.01,
|
| 241 |
+
"learning_rate": 1.5384615384615385e-06,
|
| 242 |
+
"loss": 0.7914,
|
| 243 |
"step": 190
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 0.01,
|
| 247 |
+
"learning_rate": 1.5789473684210526e-06,
|
| 248 |
+
"loss": 0.7729,
|
| 249 |
"step": 195
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.01,
|
| 253 |
+
"learning_rate": 1.6194331983805667e-06,
|
| 254 |
+
"loss": 0.7222,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 255 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 256 |
}
|
| 257 |
],
|
| 258 |
"logging_steps": 5,
|
|
|
|
| 260 |
"num_input_tokens_seen": 0,
|
| 261 |
"num_train_epochs": 1,
|
| 262 |
"save_steps": 200,
|
| 263 |
+
"total_flos": 27833727655936.0,
|
| 264 |
"trial_name": null,
|
| 265 |
"trial_params": null
|
| 266 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5752
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d21c94b3b6f8d492e13a090e98295d293df38fd5c7e7fd876836cbc0e9099e77
|
| 3 |
size 5752
|