Upload folder using huggingface_hub
Browse files- adapter_model.bin +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +188 -176
- training_args.bin +1 -1
adapter_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 100299853
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d16c842d2d6875baaa0ed21d3120cda04cd53314c22ca8fcfba7479ab413f44
|
| 3 |
size 100299853
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 200654493
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:df17622e5738ff9a5ea25296f390902c1b6aea1dad064e373a1b7628941d974a
|
| 3 |
size 200654493
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14575
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cad4573d0b2583661233eff441f718a499a533f60da927777c826a562a2040ab
|
| 3 |
size 14575
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ab2901562c0d43982c3be6d04d508fbad06acb348a5977583b18c95b785657d
|
| 3 |
size 627
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 2.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -11,375 +11,387 @@
|
|
| 11 |
{
|
| 12 |
"epoch": 0.05,
|
| 13 |
"learning_rate": 3.0769230769230774e-05,
|
| 14 |
-
"loss": 2.
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
"learning_rate": 6.153846153846155e-05,
|
| 20 |
-
"loss": 2.
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.14,
|
| 25 |
"learning_rate": 9.230769230769232e-05,
|
| 26 |
-
"loss": 2.
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.19,
|
| 31 |
"learning_rate": 0.0001230769230769231,
|
| 32 |
-
"loss":
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.24,
|
| 37 |
"learning_rate": 0.00015384615384615385,
|
| 38 |
-
"loss":
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
"learning_rate": 0.00018461538461538463,
|
| 44 |
-
"loss":
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate": 0.
|
| 56 |
-
"loss": 2.
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.43,
|
| 61 |
-
"learning_rate": 0.
|
| 62 |
-
"loss":
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch": 0.
|
| 67 |
-
"learning_rate": 0.
|
| 68 |
-
"loss": 1.
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch": 0.
|
| 73 |
-
"learning_rate": 0.
|
| 74 |
-
"loss":
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch": 0.
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss":
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch": 0.
|
| 85 |
-
"learning_rate": 0.
|
| 86 |
-
"loss": 1.
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch": 0.
|
| 91 |
-
"learning_rate": 0.
|
| 92 |
-
"loss":
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate": 0.
|
| 98 |
-
"loss": 1.
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch": 0.
|
| 103 |
-
"learning_rate": 0.
|
| 104 |
-
"loss": 1.
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch": 0.
|
| 109 |
-
"learning_rate": 0.
|
| 110 |
-
"loss": 2.
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch": 0.
|
| 115 |
-
"learning_rate": 0.
|
| 116 |
-
"loss": 1.
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch": 0.
|
| 121 |
-
"learning_rate": 0.
|
| 122 |
-
"loss": 2.
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch": 0.
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss": 2.
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch": 1.
|
| 133 |
-
"learning_rate": 0.
|
| 134 |
-
"loss": 2.
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 1.
|
| 139 |
-
"learning_rate": 0.
|
| 140 |
-
"loss": 1.
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
-
"epoch": 1.
|
| 145 |
-
"learning_rate": 0.
|
| 146 |
-
"loss":
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
-
"epoch": 1.
|
| 151 |
-
"learning_rate": 0.
|
| 152 |
-
"loss": 1.
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
-
"epoch": 1.
|
| 157 |
-
"learning_rate": 0.
|
| 158 |
-
"loss": 1.
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
-
"epoch": 1.
|
| 163 |
-
"learning_rate": 0.
|
| 164 |
-
"loss": 1.
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
-
"epoch": 1.
|
| 169 |
-
"learning_rate": 0.
|
| 170 |
-
"loss": 1.
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
-
"epoch": 1.
|
| 175 |
-
"learning_rate": 0.
|
| 176 |
-
"loss": 1.
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 1.
|
| 181 |
-
"learning_rate": 0.
|
| 182 |
-
"loss": 1.
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
-
"epoch": 1.
|
| 187 |
-
"learning_rate": 0.
|
| 188 |
-
"loss": 1.
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
-
"epoch": 1.
|
| 193 |
-
"learning_rate": 0.
|
| 194 |
-
"loss": 1.
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
-
"epoch": 1.
|
| 199 |
-
"learning_rate": 0.
|
| 200 |
-
"loss": 1.
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
-
"epoch": 1.
|
| 205 |
-
"learning_rate": 0.
|
| 206 |
-
"loss": 1.
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
-
"epoch": 1.
|
| 211 |
-
"learning_rate": 0.
|
| 212 |
-
"loss": 1.
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
-
"epoch": 1.
|
| 217 |
-
"learning_rate": 9.
|
| 218 |
-
"loss": 1.
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 1.
|
| 223 |
-
"learning_rate": 9.
|
| 224 |
-
"loss": 1.
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
-
"epoch": 1.
|
| 229 |
-
"learning_rate":
|
| 230 |
-
"loss": 1.
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
-
"epoch": 1.
|
| 235 |
-
"learning_rate": 8.
|
| 236 |
-
"loss": 1.
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
-
"epoch": 1.
|
| 241 |
-
"learning_rate": 8.
|
| 242 |
-
"loss": 1.
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
-
"epoch": 1.
|
| 247 |
-
"learning_rate":
|
| 248 |
-
"loss": 1.
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
-
"epoch": 1.
|
| 253 |
-
"learning_rate": 7.
|
| 254 |
-
"loss": 1.
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
-
"epoch":
|
| 259 |
-
"learning_rate": 7.
|
| 260 |
-
"loss": 1.
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 2.
|
| 265 |
-
"learning_rate":
|
| 266 |
-
"loss": 1.
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
-
"epoch": 2.
|
| 271 |
-
"learning_rate": 6.
|
| 272 |
-
"loss": 1.
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
-
"epoch": 2.
|
| 277 |
-
"learning_rate":
|
| 278 |
-
"loss": 1.
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
-
"epoch": 2.
|
| 283 |
-
"learning_rate":
|
| 284 |
-
"loss": 1.
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
-
"epoch": 2.
|
| 289 |
-
"learning_rate": 5.
|
| 290 |
-
"loss": 1.
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
-
"epoch": 2.
|
| 295 |
-
"learning_rate":
|
| 296 |
-
"loss": 1.
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
-
"epoch": 2.
|
| 301 |
-
"learning_rate": 4.
|
| 302 |
-
"loss": 1.
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 2.
|
| 307 |
-
"learning_rate": 4.
|
| 308 |
-
"loss": 1.
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
-
"epoch": 2.
|
| 313 |
-
"learning_rate":
|
| 314 |
-
"loss": 1.
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
-
"epoch": 2.
|
| 319 |
-
"learning_rate": 3.
|
| 320 |
-
"loss": 1.
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
-
"epoch": 2.
|
| 325 |
-
"learning_rate": 3.
|
| 326 |
-
"loss": 1.
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
-
"epoch": 2.
|
| 331 |
-
"learning_rate":
|
| 332 |
-
"loss": 1.
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
-
"epoch": 2.
|
| 337 |
-
"learning_rate": 2.
|
| 338 |
-
"loss": 1.
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
-
"epoch": 2.
|
| 343 |
-
"learning_rate":
|
| 344 |
-
"loss": 1.
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 2.
|
| 349 |
-
"learning_rate":
|
| 350 |
-
"loss": 1.
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
-
"epoch": 2.
|
| 355 |
-
"learning_rate": 1.
|
| 356 |
-
"loss": 1.
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
-
"epoch": 2.
|
| 361 |
-
"learning_rate":
|
| 362 |
-
"loss": 1.
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
-
"epoch": 2.
|
| 367 |
-
"learning_rate":
|
| 368 |
-
"loss": 1.
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
-
"epoch": 2.
|
| 373 |
-
"learning_rate":
|
| 374 |
-
"loss": 1.
|
| 375 |
"step": 122
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 376 |
}
|
| 377 |
],
|
| 378 |
"logging_steps": 2,
|
| 379 |
-
"max_steps":
|
| 380 |
"num_train_epochs": 3,
|
| 381 |
"save_steps": 500,
|
| 382 |
-
"total_flos": 3.
|
| 383 |
"trial_name": null,
|
| 384 |
"trial_params": null
|
| 385 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.991097922848665,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 126,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 0.05,
|
| 13 |
"learning_rate": 3.0769230769230774e-05,
|
| 14 |
+
"loss": 2.1273,
|
| 15 |
"step": 2
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.09,
|
| 19 |
"learning_rate": 6.153846153846155e-05,
|
| 20 |
+
"loss": 2.0994,
|
| 21 |
"step": 4
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 0.14,
|
| 25 |
"learning_rate": 9.230769230769232e-05,
|
| 26 |
+
"loss": 2.0854,
|
| 27 |
"step": 6
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.19,
|
| 31 |
"learning_rate": 0.0001230769230769231,
|
| 32 |
+
"loss": 1.9979,
|
| 33 |
"step": 8
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.24,
|
| 37 |
"learning_rate": 0.00015384615384615385,
|
| 38 |
+
"loss": 1.9966,
|
| 39 |
"step": 10
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.28,
|
| 43 |
"learning_rate": 0.00018461538461538463,
|
| 44 |
+
"loss": 1.9579,
|
| 45 |
"step": 12
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.33,
|
| 49 |
+
"learning_rate": 0.00019823008849557524,
|
| 50 |
+
"loss": 2.0361,
|
| 51 |
"step": 14
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.38,
|
| 55 |
+
"learning_rate": 0.00019469026548672567,
|
| 56 |
+
"loss": 2.073,
|
| 57 |
"step": 16
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.43,
|
| 61 |
+
"learning_rate": 0.00019115044247787613,
|
| 62 |
+
"loss": 2.0252,
|
| 63 |
"step": 18
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.47,
|
| 67 |
+
"learning_rate": 0.00018761061946902656,
|
| 68 |
+
"loss": 1.9864,
|
| 69 |
"step": 20
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.52,
|
| 73 |
+
"learning_rate": 0.000184070796460177,
|
| 74 |
+
"loss": 1.9977,
|
| 75 |
"step": 22
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.57,
|
| 79 |
+
"learning_rate": 0.00018053097345132742,
|
| 80 |
+
"loss": 1.988,
|
| 81 |
"step": 24
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.62,
|
| 85 |
+
"learning_rate": 0.0001769911504424779,
|
| 86 |
+
"loss": 1.998,
|
| 87 |
"step": 26
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.66,
|
| 91 |
+
"learning_rate": 0.00017345132743362834,
|
| 92 |
+
"loss": 2.0194,
|
| 93 |
"step": 28
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.71,
|
| 97 |
+
"learning_rate": 0.00016991150442477877,
|
| 98 |
+
"loss": 1.9534,
|
| 99 |
"step": 30
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.76,
|
| 103 |
+
"learning_rate": 0.0001663716814159292,
|
| 104 |
+
"loss": 1.9289,
|
| 105 |
"step": 32
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 0.81,
|
| 109 |
+
"learning_rate": 0.00016283185840707966,
|
| 110 |
+
"loss": 2.0024,
|
| 111 |
"step": 34
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.85,
|
| 115 |
+
"learning_rate": 0.0001592920353982301,
|
| 116 |
+
"loss": 1.9104,
|
| 117 |
"step": 36
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.9,
|
| 121 |
+
"learning_rate": 0.00015575221238938055,
|
| 122 |
+
"loss": 2.0109,
|
| 123 |
"step": 38
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 0.95,
|
| 127 |
+
"learning_rate": 0.00015221238938053098,
|
| 128 |
+
"loss": 2.0065,
|
| 129 |
"step": 40
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 1.0,
|
| 133 |
+
"learning_rate": 0.00014867256637168144,
|
| 134 |
+
"loss": 2.0736,
|
| 135 |
"step": 42
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 1.04,
|
| 139 |
+
"learning_rate": 0.00014513274336283187,
|
| 140 |
+
"loss": 1.8957,
|
| 141 |
"step": 44
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 1.09,
|
| 145 |
+
"learning_rate": 0.0001415929203539823,
|
| 146 |
+
"loss": 2.0225,
|
| 147 |
"step": 46
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 1.14,
|
| 151 |
+
"learning_rate": 0.00013805309734513276,
|
| 152 |
+
"loss": 1.8477,
|
| 153 |
"step": 48
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 1.19,
|
| 157 |
+
"learning_rate": 0.00013451327433628321,
|
| 158 |
+
"loss": 1.8566,
|
| 159 |
"step": 50
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 1.23,
|
| 163 |
+
"learning_rate": 0.00013097345132743365,
|
| 164 |
+
"loss": 1.8938,
|
| 165 |
"step": 52
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 1.28,
|
| 169 |
+
"learning_rate": 0.00012743362831858408,
|
| 170 |
+
"loss": 1.7651,
|
| 171 |
"step": 54
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 1.33,
|
| 175 |
+
"learning_rate": 0.0001238938053097345,
|
| 176 |
+
"loss": 1.7856,
|
| 177 |
"step": 56
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 1.38,
|
| 181 |
+
"learning_rate": 0.00012035398230088497,
|
| 182 |
+
"loss": 1.8789,
|
| 183 |
"step": 58
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 1.42,
|
| 187 |
+
"learning_rate": 0.00011681415929203541,
|
| 188 |
+
"loss": 1.8553,
|
| 189 |
"step": 60
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 1.47,
|
| 193 |
+
"learning_rate": 0.00011327433628318584,
|
| 194 |
+
"loss": 1.8999,
|
| 195 |
"step": 62
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 1.52,
|
| 199 |
+
"learning_rate": 0.00010973451327433629,
|
| 200 |
+
"loss": 1.8786,
|
| 201 |
"step": 64
|
| 202 |
},
|
| 203 |
{
|
| 204 |
+
"epoch": 1.57,
|
| 205 |
+
"learning_rate": 0.00010619469026548674,
|
| 206 |
+
"loss": 1.9334,
|
| 207 |
"step": 66
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 1.61,
|
| 211 |
+
"learning_rate": 0.00010265486725663717,
|
| 212 |
+
"loss": 1.8059,
|
| 213 |
"step": 68
|
| 214 |
},
|
| 215 |
{
|
| 216 |
+
"epoch": 1.66,
|
| 217 |
+
"learning_rate": 9.911504424778762e-05,
|
| 218 |
+
"loss": 1.9664,
|
| 219 |
"step": 70
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 1.71,
|
| 223 |
+
"learning_rate": 9.557522123893806e-05,
|
| 224 |
+
"loss": 1.7888,
|
| 225 |
"step": 72
|
| 226 |
},
|
| 227 |
{
|
| 228 |
+
"epoch": 1.76,
|
| 229 |
+
"learning_rate": 9.20353982300885e-05,
|
| 230 |
+
"loss": 1.83,
|
| 231 |
"step": 74
|
| 232 |
},
|
| 233 |
{
|
| 234 |
+
"epoch": 1.8,
|
| 235 |
+
"learning_rate": 8.849557522123895e-05,
|
| 236 |
+
"loss": 1.902,
|
| 237 |
"step": 76
|
| 238 |
},
|
| 239 |
{
|
| 240 |
+
"epoch": 1.85,
|
| 241 |
+
"learning_rate": 8.495575221238938e-05,
|
| 242 |
+
"loss": 1.7865,
|
| 243 |
"step": 78
|
| 244 |
},
|
| 245 |
{
|
| 246 |
+
"epoch": 1.9,
|
| 247 |
+
"learning_rate": 8.141592920353983e-05,
|
| 248 |
+
"loss": 1.8831,
|
| 249 |
"step": 80
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 1.95,
|
| 253 |
+
"learning_rate": 7.787610619469027e-05,
|
| 254 |
+
"loss": 1.8708,
|
| 255 |
"step": 82
|
| 256 |
},
|
| 257 |
{
|
| 258 |
+
"epoch": 1.99,
|
| 259 |
+
"learning_rate": 7.433628318584072e-05,
|
| 260 |
+
"loss": 1.9322,
|
| 261 |
"step": 84
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 2.04,
|
| 265 |
+
"learning_rate": 7.079646017699115e-05,
|
| 266 |
+
"loss": 1.7906,
|
| 267 |
"step": 86
|
| 268 |
},
|
| 269 |
{
|
| 270 |
+
"epoch": 2.09,
|
| 271 |
+
"learning_rate": 6.725663716814161e-05,
|
| 272 |
+
"loss": 1.8455,
|
| 273 |
"step": 88
|
| 274 |
},
|
| 275 |
{
|
| 276 |
+
"epoch": 2.14,
|
| 277 |
+
"learning_rate": 6.371681415929204e-05,
|
| 278 |
+
"loss": 1.7047,
|
| 279 |
"step": 90
|
| 280 |
},
|
| 281 |
{
|
| 282 |
+
"epoch": 2.18,
|
| 283 |
+
"learning_rate": 6.017699115044248e-05,
|
| 284 |
+
"loss": 1.6546,
|
| 285 |
"step": 92
|
| 286 |
},
|
| 287 |
{
|
| 288 |
+
"epoch": 2.23,
|
| 289 |
+
"learning_rate": 5.663716814159292e-05,
|
| 290 |
+
"loss": 1.7517,
|
| 291 |
"step": 94
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 2.28,
|
| 295 |
+
"learning_rate": 5.309734513274337e-05,
|
| 296 |
+
"loss": 1.7851,
|
| 297 |
"step": 96
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 2.33,
|
| 301 |
+
"learning_rate": 4.955752212389381e-05,
|
| 302 |
+
"loss": 1.6925,
|
| 303 |
"step": 98
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 2.37,
|
| 307 |
+
"learning_rate": 4.601769911504425e-05,
|
| 308 |
+
"loss": 1.8215,
|
| 309 |
"step": 100
|
| 310 |
},
|
| 311 |
{
|
| 312 |
+
"epoch": 2.42,
|
| 313 |
+
"learning_rate": 4.247787610619469e-05,
|
| 314 |
+
"loss": 1.7755,
|
| 315 |
"step": 102
|
| 316 |
},
|
| 317 |
{
|
| 318 |
+
"epoch": 2.47,
|
| 319 |
+
"learning_rate": 3.893805309734514e-05,
|
| 320 |
+
"loss": 1.7762,
|
| 321 |
"step": 104
|
| 322 |
},
|
| 323 |
{
|
| 324 |
+
"epoch": 2.52,
|
| 325 |
+
"learning_rate": 3.5398230088495574e-05,
|
| 326 |
+
"loss": 1.7803,
|
| 327 |
"step": 106
|
| 328 |
},
|
| 329 |
{
|
| 330 |
+
"epoch": 2.56,
|
| 331 |
+
"learning_rate": 3.185840707964602e-05,
|
| 332 |
+
"loss": 1.8136,
|
| 333 |
"step": 108
|
| 334 |
},
|
| 335 |
{
|
| 336 |
+
"epoch": 2.61,
|
| 337 |
+
"learning_rate": 2.831858407079646e-05,
|
| 338 |
+
"loss": 1.6939,
|
| 339 |
"step": 110
|
| 340 |
},
|
| 341 |
{
|
| 342 |
+
"epoch": 2.66,
|
| 343 |
+
"learning_rate": 2.4778761061946905e-05,
|
| 344 |
+
"loss": 1.6688,
|
| 345 |
"step": 112
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 2.71,
|
| 349 |
+
"learning_rate": 2.1238938053097346e-05,
|
| 350 |
+
"loss": 1.7594,
|
| 351 |
"step": 114
|
| 352 |
},
|
| 353 |
{
|
| 354 |
+
"epoch": 2.75,
|
| 355 |
+
"learning_rate": 1.7699115044247787e-05,
|
| 356 |
+
"loss": 1.8506,
|
| 357 |
"step": 116
|
| 358 |
},
|
| 359 |
{
|
| 360 |
+
"epoch": 2.8,
|
| 361 |
+
"learning_rate": 1.415929203539823e-05,
|
| 362 |
+
"loss": 1.7872,
|
| 363 |
"step": 118
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 2.85,
|
| 367 |
+
"learning_rate": 1.0619469026548673e-05,
|
| 368 |
+
"loss": 1.8039,
|
| 369 |
"step": 120
|
| 370 |
},
|
| 371 |
{
|
| 372 |
+
"epoch": 2.9,
|
| 373 |
+
"learning_rate": 7.079646017699115e-06,
|
| 374 |
+
"loss": 1.7888,
|
| 375 |
"step": 122
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.94,
|
| 379 |
+
"learning_rate": 3.5398230088495575e-06,
|
| 380 |
+
"loss": 1.8769,
|
| 381 |
+
"step": 124
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 2.99,
|
| 385 |
+
"learning_rate": 0.0,
|
| 386 |
+
"loss": 1.7694,
|
| 387 |
+
"step": 126
|
| 388 |
}
|
| 389 |
],
|
| 390 |
"logging_steps": 2,
|
| 391 |
+
"max_steps": 126,
|
| 392 |
"num_train_epochs": 3,
|
| 393 |
"save_steps": 500,
|
| 394 |
+
"total_flos": 3.333935476821197e+16,
|
| 395 |
"trial_name": null,
|
| 396 |
"trial_params": null
|
| 397 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4027
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2b4c8be2afa92d1340e64da2cf4f2dc5ca13d2c797ac60a3ebdc68340983a85
|
| 3 |
size 4027
|