Commit
•
20716ad
1
Parent(s):
1e592c3
Training in progress, step 15, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -19,13 +19,13 @@
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
|
|
|
| 22 |
"down_proj",
|
| 23 |
-
"
|
| 24 |
"k_proj",
|
|
|
|
| 25 |
"o_proj",
|
| 26 |
-
"q_proj"
|
| 27 |
-
"gate_proj",
|
| 28 |
-
"up_proj"
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM"
|
| 31 |
}
|
|
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
| 22 |
+
"gate_proj",
|
| 23 |
"down_proj",
|
| 24 |
+
"up_proj",
|
| 25 |
"k_proj",
|
| 26 |
+
"v_proj",
|
| 27 |
"o_proj",
|
| 28 |
+
"q_proj"
|
|
|
|
|
|
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM"
|
| 31 |
}
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 125248064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae50d39e33235bdaa29bf052355d3b176b9c6e7495b18478180b9c9b837d9b0f
|
| 3 |
size 125248064
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 63350356
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf53fc4dcd28a566f9ea697f2e4e6a0f10991ca8c1cbeaac79568ecdc1145b3f
|
| 3 |
size 63350356
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5d185880e7fabcea80425635bb46b445f79d8628025d1e32bd0c3c1a767c69f
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e82bfeff0cfcd6e162a107e74886e0b3b1d95bebbcebce75e628978fbbb7c65d
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -23,897 +23,87 @@
|
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
"learning_rate": 0.0002,
|
| 26 |
-
"loss": 2.
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
-
"learning_rate": 0.
|
| 32 |
-
"loss": 2.
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
-
"learning_rate": 0.
|
| 38 |
-
"loss": 1.
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
-
"learning_rate": 0.
|
| 44 |
-
"loss": 2.
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
-
"learning_rate": 0.
|
| 56 |
-
"loss": 1.
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.0,
|
| 61 |
-
"learning_rate": 0.
|
| 62 |
-
"loss": 2.
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
-
"learning_rate": 0.
|
| 68 |
-
"loss": 2.
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.0,
|
| 73 |
-
"learning_rate": 0.
|
| 74 |
-
"loss": 1.
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.0,
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss": 2.
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.01,
|
| 85 |
-
"learning_rate": 0.
|
| 86 |
-
"loss": 1.
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.01,
|
| 91 |
-
"learning_rate": 0.
|
| 92 |
-
"loss": 1.
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.01,
|
| 97 |
-
"learning_rate": 0.
|
| 98 |
-
"loss": 1.
|
| 99 |
"step": 15
|
| 100 |
-
},
|
| 101 |
-
{
|
| 102 |
-
"epoch": 0.01,
|
| 103 |
-
"learning_rate": 0.00018231292517006804,
|
| 104 |
-
"loss": 1.5366,
|
| 105 |
-
"step": 16
|
| 106 |
-
},
|
| 107 |
-
{
|
| 108 |
-
"epoch": 0.01,
|
| 109 |
-
"learning_rate": 0.00018095238095238095,
|
| 110 |
-
"loss": 1.7421,
|
| 111 |
-
"step": 17
|
| 112 |
-
},
|
| 113 |
-
{
|
| 114 |
-
"epoch": 0.01,
|
| 115 |
-
"learning_rate": 0.0001795918367346939,
|
| 116 |
-
"loss": 2.5308,
|
| 117 |
-
"step": 18
|
| 118 |
-
},
|
| 119 |
-
{
|
| 120 |
-
"epoch": 0.01,
|
| 121 |
-
"learning_rate": 0.00017823129251700681,
|
| 122 |
-
"loss": 1.5277,
|
| 123 |
-
"step": 19
|
| 124 |
-
},
|
| 125 |
-
{
|
| 126 |
-
"epoch": 0.01,
|
| 127 |
-
"learning_rate": 0.00017687074829931973,
|
| 128 |
-
"loss": 1.9084,
|
| 129 |
-
"step": 20
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.01,
|
| 133 |
-
"learning_rate": 0.00017551020408163265,
|
| 134 |
-
"loss": 1.9708,
|
| 135 |
-
"step": 21
|
| 136 |
-
},
|
| 137 |
-
{
|
| 138 |
-
"epoch": 0.01,
|
| 139 |
-
"learning_rate": 0.0001741496598639456,
|
| 140 |
-
"loss": 1.8567,
|
| 141 |
-
"step": 22
|
| 142 |
-
},
|
| 143 |
-
{
|
| 144 |
-
"epoch": 0.01,
|
| 145 |
-
"learning_rate": 0.0001727891156462585,
|
| 146 |
-
"loss": 1.5694,
|
| 147 |
-
"step": 23
|
| 148 |
-
},
|
| 149 |
-
{
|
| 150 |
-
"epoch": 0.01,
|
| 151 |
-
"learning_rate": 0.00017142857142857143,
|
| 152 |
-
"loss": 1.8519,
|
| 153 |
-
"step": 24
|
| 154 |
-
},
|
| 155 |
-
{
|
| 156 |
-
"epoch": 0.01,
|
| 157 |
-
"learning_rate": 0.00017006802721088434,
|
| 158 |
-
"loss": 1.663,
|
| 159 |
-
"step": 25
|
| 160 |
-
},
|
| 161 |
-
{
|
| 162 |
-
"epoch": 0.01,
|
| 163 |
-
"learning_rate": 0.00016870748299319729,
|
| 164 |
-
"loss": 1.9309,
|
| 165 |
-
"step": 26
|
| 166 |
-
},
|
| 167 |
-
{
|
| 168 |
-
"epoch": 0.01,
|
| 169 |
-
"learning_rate": 0.00016734693877551023,
|
| 170 |
-
"loss": 1.551,
|
| 171 |
-
"step": 27
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.01,
|
| 175 |
-
"learning_rate": 0.00016598639455782315,
|
| 176 |
-
"loss": 1.4782,
|
| 177 |
-
"step": 28
|
| 178 |
-
},
|
| 179 |
-
{
|
| 180 |
-
"epoch": 0.01,
|
| 181 |
-
"learning_rate": 0.00016462585034013606,
|
| 182 |
-
"loss": 1.6643,
|
| 183 |
-
"step": 29
|
| 184 |
-
},
|
| 185 |
-
{
|
| 186 |
-
"epoch": 0.01,
|
| 187 |
-
"learning_rate": 0.00016326530612244898,
|
| 188 |
-
"loss": 1.5571,
|
| 189 |
-
"step": 30
|
| 190 |
-
},
|
| 191 |
-
{
|
| 192 |
-
"epoch": 0.01,
|
| 193 |
-
"learning_rate": 0.00016190476190476192,
|
| 194 |
-
"loss": 1.6747,
|
| 195 |
-
"step": 31
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 0.01,
|
| 199 |
-
"learning_rate": 0.00016054421768707484,
|
| 200 |
-
"loss": 1.3539,
|
| 201 |
-
"step": 32
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 0.01,
|
| 205 |
-
"learning_rate": 0.00015918367346938776,
|
| 206 |
-
"loss": 2.0579,
|
| 207 |
-
"step": 33
|
| 208 |
-
},
|
| 209 |
-
{
|
| 210 |
-
"epoch": 0.01,
|
| 211 |
-
"learning_rate": 0.00015782312925170067,
|
| 212 |
-
"loss": 1.5418,
|
| 213 |
-
"step": 34
|
| 214 |
-
},
|
| 215 |
-
{
|
| 216 |
-
"epoch": 0.01,
|
| 217 |
-
"learning_rate": 0.00015646258503401362,
|
| 218 |
-
"loss": 1.8159,
|
| 219 |
-
"step": 35
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 0.01,
|
| 223 |
-
"learning_rate": 0.00015510204081632654,
|
| 224 |
-
"loss": 1.6081,
|
| 225 |
-
"step": 36
|
| 226 |
-
},
|
| 227 |
-
{
|
| 228 |
-
"epoch": 0.01,
|
| 229 |
-
"learning_rate": 0.00015374149659863945,
|
| 230 |
-
"loss": 1.5529,
|
| 231 |
-
"step": 37
|
| 232 |
-
},
|
| 233 |
-
{
|
| 234 |
-
"epoch": 0.02,
|
| 235 |
-
"learning_rate": 0.00015238095238095237,
|
| 236 |
-
"loss": 1.5456,
|
| 237 |
-
"step": 38
|
| 238 |
-
},
|
| 239 |
-
{
|
| 240 |
-
"epoch": 0.02,
|
| 241 |
-
"learning_rate": 0.0001510204081632653,
|
| 242 |
-
"loss": 2.1016,
|
| 243 |
-
"step": 39
|
| 244 |
-
},
|
| 245 |
-
{
|
| 246 |
-
"epoch": 0.02,
|
| 247 |
-
"learning_rate": 0.00014965986394557826,
|
| 248 |
-
"loss": 1.4394,
|
| 249 |
-
"step": 40
|
| 250 |
-
},
|
| 251 |
-
{
|
| 252 |
-
"epoch": 0.02,
|
| 253 |
-
"learning_rate": 0.00014829931972789117,
|
| 254 |
-
"loss": 1.4033,
|
| 255 |
-
"step": 41
|
| 256 |
-
},
|
| 257 |
-
{
|
| 258 |
-
"epoch": 0.02,
|
| 259 |
-
"learning_rate": 0.0001469387755102041,
|
| 260 |
-
"loss": 1.5887,
|
| 261 |
-
"step": 42
|
| 262 |
-
},
|
| 263 |
-
{
|
| 264 |
-
"epoch": 0.02,
|
| 265 |
-
"learning_rate": 0.000145578231292517,
|
| 266 |
-
"loss": 1.8761,
|
| 267 |
-
"step": 43
|
| 268 |
-
},
|
| 269 |
-
{
|
| 270 |
-
"epoch": 0.02,
|
| 271 |
-
"learning_rate": 0.00014421768707482995,
|
| 272 |
-
"loss": 1.67,
|
| 273 |
-
"step": 44
|
| 274 |
-
},
|
| 275 |
-
{
|
| 276 |
-
"epoch": 0.02,
|
| 277 |
-
"learning_rate": 0.00014285714285714287,
|
| 278 |
-
"loss": 1.8064,
|
| 279 |
-
"step": 45
|
| 280 |
-
},
|
| 281 |
-
{
|
| 282 |
-
"epoch": 0.02,
|
| 283 |
-
"learning_rate": 0.00014149659863945578,
|
| 284 |
-
"loss": 1.4283,
|
| 285 |
-
"step": 46
|
| 286 |
-
},
|
| 287 |
-
{
|
| 288 |
-
"epoch": 0.02,
|
| 289 |
-
"learning_rate": 0.0001401360544217687,
|
| 290 |
-
"loss": 1.8028,
|
| 291 |
-
"step": 47
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 0.02,
|
| 295 |
-
"learning_rate": 0.00013877551020408165,
|
| 296 |
-
"loss": 1.5101,
|
| 297 |
-
"step": 48
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 0.02,
|
| 301 |
-
"learning_rate": 0.00013741496598639456,
|
| 302 |
-
"loss": 1.7825,
|
| 303 |
-
"step": 49
|
| 304 |
-
},
|
| 305 |
-
{
|
| 306 |
-
"epoch": 0.02,
|
| 307 |
-
"learning_rate": 0.00013605442176870748,
|
| 308 |
-
"loss": 1.3797,
|
| 309 |
-
"step": 50
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 0.02,
|
| 313 |
-
"learning_rate": 0.0001346938775510204,
|
| 314 |
-
"loss": 1.7408,
|
| 315 |
-
"step": 51
|
| 316 |
-
},
|
| 317 |
-
{
|
| 318 |
-
"epoch": 0.02,
|
| 319 |
-
"learning_rate": 0.00013333333333333334,
|
| 320 |
-
"loss": 1.6008,
|
| 321 |
-
"step": 52
|
| 322 |
-
},
|
| 323 |
-
{
|
| 324 |
-
"epoch": 0.02,
|
| 325 |
-
"learning_rate": 0.00013197278911564626,
|
| 326 |
-
"loss": 1.5121,
|
| 327 |
-
"step": 53
|
| 328 |
-
},
|
| 329 |
-
{
|
| 330 |
-
"epoch": 0.02,
|
| 331 |
-
"learning_rate": 0.00013061224489795917,
|
| 332 |
-
"loss": 1.3599,
|
| 333 |
-
"step": 54
|
| 334 |
-
},
|
| 335 |
-
{
|
| 336 |
-
"epoch": 0.02,
|
| 337 |
-
"learning_rate": 0.00012925170068027212,
|
| 338 |
-
"loss": 1.4465,
|
| 339 |
-
"step": 55
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 0.02,
|
| 343 |
-
"learning_rate": 0.00012789115646258506,
|
| 344 |
-
"loss": 1.6332,
|
| 345 |
-
"step": 56
|
| 346 |
-
},
|
| 347 |
-
{
|
| 348 |
-
"epoch": 0.02,
|
| 349 |
-
"learning_rate": 0.00012653061224489798,
|
| 350 |
-
"loss": 1.6305,
|
| 351 |
-
"step": 57
|
| 352 |
-
},
|
| 353 |
-
{
|
| 354 |
-
"epoch": 0.02,
|
| 355 |
-
"learning_rate": 0.0001251700680272109,
|
| 356 |
-
"loss": 1.5599,
|
| 357 |
-
"step": 58
|
| 358 |
-
},
|
| 359 |
-
{
|
| 360 |
-
"epoch": 0.02,
|
| 361 |
-
"learning_rate": 0.0001238095238095238,
|
| 362 |
-
"loss": 1.8265,
|
| 363 |
-
"step": 59
|
| 364 |
-
},
|
| 365 |
-
{
|
| 366 |
-
"epoch": 0.02,
|
| 367 |
-
"learning_rate": 0.00012244897959183676,
|
| 368 |
-
"loss": 1.7149,
|
| 369 |
-
"step": 60
|
| 370 |
-
},
|
| 371 |
-
{
|
| 372 |
-
"epoch": 0.02,
|
| 373 |
-
"learning_rate": 0.00012108843537414967,
|
| 374 |
-
"loss": 1.7577,
|
| 375 |
-
"step": 61
|
| 376 |
-
},
|
| 377 |
-
{
|
| 378 |
-
"epoch": 0.02,
|
| 379 |
-
"learning_rate": 0.00011972789115646259,
|
| 380 |
-
"loss": 1.5519,
|
| 381 |
-
"step": 62
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 0.03,
|
| 385 |
-
"learning_rate": 0.00011836734693877552,
|
| 386 |
-
"loss": 1.4681,
|
| 387 |
-
"step": 63
|
| 388 |
-
},
|
| 389 |
-
{
|
| 390 |
-
"epoch": 0.03,
|
| 391 |
-
"learning_rate": 0.00011700680272108844,
|
| 392 |
-
"loss": 1.3328,
|
| 393 |
-
"step": 64
|
| 394 |
-
},
|
| 395 |
-
{
|
| 396 |
-
"epoch": 0.03,
|
| 397 |
-
"learning_rate": 0.00011564625850340137,
|
| 398 |
-
"loss": 1.4249,
|
| 399 |
-
"step": 65
|
| 400 |
-
},
|
| 401 |
-
{
|
| 402 |
-
"epoch": 0.03,
|
| 403 |
-
"learning_rate": 0.00011428571428571428,
|
| 404 |
-
"loss": 1.8161,
|
| 405 |
-
"step": 66
|
| 406 |
-
},
|
| 407 |
-
{
|
| 408 |
-
"epoch": 0.03,
|
| 409 |
-
"learning_rate": 0.00011292517006802721,
|
| 410 |
-
"loss": 1.4775,
|
| 411 |
-
"step": 67
|
| 412 |
-
},
|
| 413 |
-
{
|
| 414 |
-
"epoch": 0.03,
|
| 415 |
-
"learning_rate": 0.00011156462585034013,
|
| 416 |
-
"loss": 1.5223,
|
| 417 |
-
"step": 68
|
| 418 |
-
},
|
| 419 |
-
{
|
| 420 |
-
"epoch": 0.03,
|
| 421 |
-
"learning_rate": 0.00011020408163265306,
|
| 422 |
-
"loss": 1.5734,
|
| 423 |
-
"step": 69
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.03,
|
| 427 |
-
"learning_rate": 0.000108843537414966,
|
| 428 |
-
"loss": 1.5901,
|
| 429 |
-
"step": 70
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 0.03,
|
| 433 |
-
"learning_rate": 0.00010748299319727892,
|
| 434 |
-
"loss": 2.1238,
|
| 435 |
-
"step": 71
|
| 436 |
-
},
|
| 437 |
-
{
|
| 438 |
-
"epoch": 0.03,
|
| 439 |
-
"learning_rate": 0.00010612244897959185,
|
| 440 |
-
"loss": 1.8808,
|
| 441 |
-
"step": 72
|
| 442 |
-
},
|
| 443 |
-
{
|
| 444 |
-
"epoch": 0.03,
|
| 445 |
-
"learning_rate": 0.00010476190476190477,
|
| 446 |
-
"loss": 1.3648,
|
| 447 |
-
"step": 73
|
| 448 |
-
},
|
| 449 |
-
{
|
| 450 |
-
"epoch": 0.03,
|
| 451 |
-
"learning_rate": 0.0001034013605442177,
|
| 452 |
-
"loss": 1.3667,
|
| 453 |
-
"step": 74
|
| 454 |
-
},
|
| 455 |
-
{
|
| 456 |
-
"epoch": 0.03,
|
| 457 |
-
"learning_rate": 0.00010204081632653062,
|
| 458 |
-
"loss": 1.3925,
|
| 459 |
-
"step": 75
|
| 460 |
-
},
|
| 461 |
-
{
|
| 462 |
-
"epoch": 0.03,
|
| 463 |
-
"learning_rate": 0.00010068027210884355,
|
| 464 |
-
"loss": 1.8534,
|
| 465 |
-
"step": 76
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.03,
|
| 469 |
-
"learning_rate": 9.931972789115646e-05,
|
| 470 |
-
"loss": 1.6686,
|
| 471 |
-
"step": 77
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 0.03,
|
| 475 |
-
"learning_rate": 9.79591836734694e-05,
|
| 476 |
-
"loss": 1.5057,
|
| 477 |
-
"step": 78
|
| 478 |
-
},
|
| 479 |
-
{
|
| 480 |
-
"epoch": 0.03,
|
| 481 |
-
"learning_rate": 9.659863945578231e-05,
|
| 482 |
-
"loss": 1.4539,
|
| 483 |
-
"step": 79
|
| 484 |
-
},
|
| 485 |
-
{
|
| 486 |
-
"epoch": 0.03,
|
| 487 |
-
"learning_rate": 9.523809523809524e-05,
|
| 488 |
-
"loss": 1.6102,
|
| 489 |
-
"step": 80
|
| 490 |
-
},
|
| 491 |
-
{
|
| 492 |
-
"epoch": 0.03,
|
| 493 |
-
"learning_rate": 9.387755102040817e-05,
|
| 494 |
-
"loss": 1.3404,
|
| 495 |
-
"step": 81
|
| 496 |
-
},
|
| 497 |
-
{
|
| 498 |
-
"epoch": 0.03,
|
| 499 |
-
"learning_rate": 9.25170068027211e-05,
|
| 500 |
-
"loss": 1.5923,
|
| 501 |
-
"step": 82
|
| 502 |
-
},
|
| 503 |
-
{
|
| 504 |
-
"epoch": 0.03,
|
| 505 |
-
"learning_rate": 9.115646258503402e-05,
|
| 506 |
-
"loss": 1.34,
|
| 507 |
-
"step": 83
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.03,
|
| 511 |
-
"learning_rate": 8.979591836734695e-05,
|
| 512 |
-
"loss": 1.9159,
|
| 513 |
-
"step": 84
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 0.03,
|
| 517 |
-
"learning_rate": 8.843537414965987e-05,
|
| 518 |
-
"loss": 1.5353,
|
| 519 |
-
"step": 85
|
| 520 |
-
},
|
| 521 |
-
{
|
| 522 |
-
"epoch": 0.03,
|
| 523 |
-
"learning_rate": 8.70748299319728e-05,
|
| 524 |
-
"loss": 1.5158,
|
| 525 |
-
"step": 86
|
| 526 |
-
},
|
| 527 |
-
{
|
| 528 |
-
"epoch": 0.03,
|
| 529 |
-
"learning_rate": 8.571428571428571e-05,
|
| 530 |
-
"loss": 1.3898,
|
| 531 |
-
"step": 87
|
| 532 |
-
},
|
| 533 |
-
{
|
| 534 |
-
"epoch": 0.04,
|
| 535 |
-
"learning_rate": 8.435374149659864e-05,
|
| 536 |
-
"loss": 1.6535,
|
| 537 |
-
"step": 88
|
| 538 |
-
},
|
| 539 |
-
{
|
| 540 |
-
"epoch": 0.04,
|
| 541 |
-
"learning_rate": 8.299319727891157e-05,
|
| 542 |
-
"loss": 1.6544,
|
| 543 |
-
"step": 89
|
| 544 |
-
},
|
| 545 |
-
{
|
| 546 |
-
"epoch": 0.04,
|
| 547 |
-
"learning_rate": 8.163265306122449e-05,
|
| 548 |
-
"loss": 1.8656,
|
| 549 |
-
"step": 90
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 0.04,
|
| 553 |
-
"learning_rate": 8.027210884353742e-05,
|
| 554 |
-
"loss": 1.465,
|
| 555 |
-
"step": 91
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 0.04,
|
| 559 |
-
"learning_rate": 7.891156462585034e-05,
|
| 560 |
-
"loss": 1.0163,
|
| 561 |
-
"step": 92
|
| 562 |
-
},
|
| 563 |
-
{
|
| 564 |
-
"epoch": 0.04,
|
| 565 |
-
"learning_rate": 7.755102040816327e-05,
|
| 566 |
-
"loss": 1.61,
|
| 567 |
-
"step": 93
|
| 568 |
-
},
|
| 569 |
-
{
|
| 570 |
-
"epoch": 0.04,
|
| 571 |
-
"learning_rate": 7.619047619047618e-05,
|
| 572 |
-
"loss": 2.0532,
|
| 573 |
-
"step": 94
|
| 574 |
-
},
|
| 575 |
-
{
|
| 576 |
-
"epoch": 0.04,
|
| 577 |
-
"learning_rate": 7.482993197278913e-05,
|
| 578 |
-
"loss": 1.4248,
|
| 579 |
-
"step": 95
|
| 580 |
-
},
|
| 581 |
-
{
|
| 582 |
-
"epoch": 0.04,
|
| 583 |
-
"learning_rate": 7.346938775510205e-05,
|
| 584 |
-
"loss": 1.4999,
|
| 585 |
-
"step": 96
|
| 586 |
-
},
|
| 587 |
-
{
|
| 588 |
-
"epoch": 0.04,
|
| 589 |
-
"learning_rate": 7.210884353741498e-05,
|
| 590 |
-
"loss": 1.6352,
|
| 591 |
-
"step": 97
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 0.04,
|
| 595 |
-
"learning_rate": 7.074829931972789e-05,
|
| 596 |
-
"loss": 1.6345,
|
| 597 |
-
"step": 98
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 0.04,
|
| 601 |
-
"learning_rate": 6.938775510204082e-05,
|
| 602 |
-
"loss": 1.6019,
|
| 603 |
-
"step": 99
|
| 604 |
-
},
|
| 605 |
-
{
|
| 606 |
-
"epoch": 0.04,
|
| 607 |
-
"learning_rate": 6.802721088435374e-05,
|
| 608 |
-
"loss": 1.4834,
|
| 609 |
-
"step": 100
|
| 610 |
-
},
|
| 611 |
-
{
|
| 612 |
-
"epoch": 0.04,
|
| 613 |
-
"learning_rate": 6.666666666666667e-05,
|
| 614 |
-
"loss": 1.4514,
|
| 615 |
-
"step": 101
|
| 616 |
-
},
|
| 617 |
-
{
|
| 618 |
-
"epoch": 0.04,
|
| 619 |
-
"learning_rate": 6.530612244897959e-05,
|
| 620 |
-
"loss": 1.5484,
|
| 621 |
-
"step": 102
|
| 622 |
-
},
|
| 623 |
-
{
|
| 624 |
-
"epoch": 0.04,
|
| 625 |
-
"learning_rate": 6.394557823129253e-05,
|
| 626 |
-
"loss": 1.6164,
|
| 627 |
-
"step": 103
|
| 628 |
-
},
|
| 629 |
-
{
|
| 630 |
-
"epoch": 0.04,
|
| 631 |
-
"learning_rate": 6.258503401360545e-05,
|
| 632 |
-
"loss": 1.6819,
|
| 633 |
-
"step": 104
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 0.04,
|
| 637 |
-
"learning_rate": 6.122448979591838e-05,
|
| 638 |
-
"loss": 1.8219,
|
| 639 |
-
"step": 105
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 0.04,
|
| 643 |
-
"learning_rate": 5.9863945578231295e-05,
|
| 644 |
-
"loss": 1.385,
|
| 645 |
-
"step": 106
|
| 646 |
-
},
|
| 647 |
-
{
|
| 648 |
-
"epoch": 0.04,
|
| 649 |
-
"learning_rate": 5.850340136054422e-05,
|
| 650 |
-
"loss": 1.6358,
|
| 651 |
-
"step": 107
|
| 652 |
-
},
|
| 653 |
-
{
|
| 654 |
-
"epoch": 0.04,
|
| 655 |
-
"learning_rate": 5.714285714285714e-05,
|
| 656 |
-
"loss": 1.5059,
|
| 657 |
-
"step": 108
|
| 658 |
-
},
|
| 659 |
-
{
|
| 660 |
-
"epoch": 0.04,
|
| 661 |
-
"learning_rate": 5.5782312925170065e-05,
|
| 662 |
-
"loss": 1.6794,
|
| 663 |
-
"step": 109
|
| 664 |
-
},
|
| 665 |
-
{
|
| 666 |
-
"epoch": 0.04,
|
| 667 |
-
"learning_rate": 5.4421768707483e-05,
|
| 668 |
-
"loss": 1.4888,
|
| 669 |
-
"step": 110
|
| 670 |
-
},
|
| 671 |
-
{
|
| 672 |
-
"epoch": 0.04,
|
| 673 |
-
"learning_rate": 5.3061224489795926e-05,
|
| 674 |
-
"loss": 1.3177,
|
| 675 |
-
"step": 111
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 0.04,
|
| 679 |
-
"learning_rate": 5.170068027210885e-05,
|
| 680 |
-
"loss": 1.5026,
|
| 681 |
-
"step": 112
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 0.05,
|
| 685 |
-
"learning_rate": 5.034013605442177e-05,
|
| 686 |
-
"loss": 1.8063,
|
| 687 |
-
"step": 113
|
| 688 |
-
},
|
| 689 |
-
{
|
| 690 |
-
"epoch": 0.05,
|
| 691 |
-
"learning_rate": 4.89795918367347e-05,
|
| 692 |
-
"loss": 1.8037,
|
| 693 |
-
"step": 114
|
| 694 |
-
},
|
| 695 |
-
{
|
| 696 |
-
"epoch": 0.05,
|
| 697 |
-
"learning_rate": 4.761904761904762e-05,
|
| 698 |
-
"loss": 1.6401,
|
| 699 |
-
"step": 115
|
| 700 |
-
},
|
| 701 |
-
{
|
| 702 |
-
"epoch": 0.05,
|
| 703 |
-
"learning_rate": 4.625850340136055e-05,
|
| 704 |
-
"loss": 1.7781,
|
| 705 |
-
"step": 116
|
| 706 |
-
},
|
| 707 |
-
{
|
| 708 |
-
"epoch": 0.05,
|
| 709 |
-
"learning_rate": 4.4897959183673474e-05,
|
| 710 |
-
"loss": 1.7268,
|
| 711 |
-
"step": 117
|
| 712 |
-
},
|
| 713 |
-
{
|
| 714 |
-
"epoch": 0.05,
|
| 715 |
-
"learning_rate": 4.35374149659864e-05,
|
| 716 |
-
"loss": 1.6482,
|
| 717 |
-
"step": 118
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 0.05,
|
| 721 |
-
"learning_rate": 4.217687074829932e-05,
|
| 722 |
-
"loss": 1.5628,
|
| 723 |
-
"step": 119
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 0.05,
|
| 727 |
-
"learning_rate": 4.0816326530612245e-05,
|
| 728 |
-
"loss": 1.3526,
|
| 729 |
-
"step": 120
|
| 730 |
-
},
|
| 731 |
-
{
|
| 732 |
-
"epoch": 0.05,
|
| 733 |
-
"learning_rate": 3.945578231292517e-05,
|
| 734 |
-
"loss": 1.3096,
|
| 735 |
-
"step": 121
|
| 736 |
-
},
|
| 737 |
-
{
|
| 738 |
-
"epoch": 0.05,
|
| 739 |
-
"learning_rate": 3.809523809523809e-05,
|
| 740 |
-
"loss": 1.6429,
|
| 741 |
-
"step": 122
|
| 742 |
-
},
|
| 743 |
-
{
|
| 744 |
-
"epoch": 0.05,
|
| 745 |
-
"learning_rate": 3.673469387755102e-05,
|
| 746 |
-
"loss": 1.4119,
|
| 747 |
-
"step": 123
|
| 748 |
-
},
|
| 749 |
-
{
|
| 750 |
-
"epoch": 0.05,
|
| 751 |
-
"learning_rate": 3.5374149659863946e-05,
|
| 752 |
-
"loss": 1.4932,
|
| 753 |
-
"step": 124
|
| 754 |
-
},
|
| 755 |
-
{
|
| 756 |
-
"epoch": 0.05,
|
| 757 |
-
"learning_rate": 3.401360544217687e-05,
|
| 758 |
-
"loss": 1.8325,
|
| 759 |
-
"step": 125
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 0.05,
|
| 763 |
-
"learning_rate": 3.265306122448979e-05,
|
| 764 |
-
"loss": 1.6309,
|
| 765 |
-
"step": 126
|
| 766 |
-
},
|
| 767 |
-
{
|
| 768 |
-
"epoch": 0.05,
|
| 769 |
-
"learning_rate": 3.1292517006802724e-05,
|
| 770 |
-
"loss": 1.6754,
|
| 771 |
-
"step": 127
|
| 772 |
-
},
|
| 773 |
-
{
|
| 774 |
-
"epoch": 0.05,
|
| 775 |
-
"learning_rate": 2.9931972789115647e-05,
|
| 776 |
-
"loss": 1.1972,
|
| 777 |
-
"step": 128
|
| 778 |
-
},
|
| 779 |
-
{
|
| 780 |
-
"epoch": 0.05,
|
| 781 |
-
"learning_rate": 2.857142857142857e-05,
|
| 782 |
-
"loss": 1.6411,
|
| 783 |
-
"step": 129
|
| 784 |
-
},
|
| 785 |
-
{
|
| 786 |
-
"epoch": 0.05,
|
| 787 |
-
"learning_rate": 2.72108843537415e-05,
|
| 788 |
-
"loss": 2.031,
|
| 789 |
-
"step": 130
|
| 790 |
-
},
|
| 791 |
-
{
|
| 792 |
-
"epoch": 0.05,
|
| 793 |
-
"learning_rate": 2.5850340136054425e-05,
|
| 794 |
-
"loss": 1.726,
|
| 795 |
-
"step": 131
|
| 796 |
-
},
|
| 797 |
-
{
|
| 798 |
-
"epoch": 0.05,
|
| 799 |
-
"learning_rate": 2.448979591836735e-05,
|
| 800 |
-
"loss": 1.0537,
|
| 801 |
-
"step": 132
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 0.05,
|
| 805 |
-
"learning_rate": 2.3129251700680275e-05,
|
| 806 |
-
"loss": 1.2133,
|
| 807 |
-
"step": 133
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 0.05,
|
| 811 |
-
"learning_rate": 2.17687074829932e-05,
|
| 812 |
-
"loss": 1.4264,
|
| 813 |
-
"step": 134
|
| 814 |
-
},
|
| 815 |
-
{
|
| 816 |
-
"epoch": 0.05,
|
| 817 |
-
"learning_rate": 2.0408163265306123e-05,
|
| 818 |
-
"loss": 1.8015,
|
| 819 |
-
"step": 135
|
| 820 |
-
},
|
| 821 |
-
{
|
| 822 |
-
"epoch": 0.05,
|
| 823 |
-
"learning_rate": 1.9047619047619046e-05,
|
| 824 |
-
"loss": 1.6847,
|
| 825 |
-
"step": 136
|
| 826 |
-
},
|
| 827 |
-
{
|
| 828 |
-
"epoch": 0.05,
|
| 829 |
-
"learning_rate": 1.7687074829931973e-05,
|
| 830 |
-
"loss": 1.6395,
|
| 831 |
-
"step": 137
|
| 832 |
-
},
|
| 833 |
-
{
|
| 834 |
-
"epoch": 0.06,
|
| 835 |
-
"learning_rate": 1.6326530612244897e-05,
|
| 836 |
-
"loss": 1.3069,
|
| 837 |
-
"step": 138
|
| 838 |
-
},
|
| 839 |
-
{
|
| 840 |
-
"epoch": 0.06,
|
| 841 |
-
"learning_rate": 1.4965986394557824e-05,
|
| 842 |
-
"loss": 1.7292,
|
| 843 |
-
"step": 139
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 0.06,
|
| 847 |
-
"learning_rate": 1.360544217687075e-05,
|
| 848 |
-
"loss": 1.5996,
|
| 849 |
-
"step": 140
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 0.06,
|
| 853 |
-
"learning_rate": 1.2244897959183674e-05,
|
| 854 |
-
"loss": 1.5048,
|
| 855 |
-
"step": 141
|
| 856 |
-
},
|
| 857 |
-
{
|
| 858 |
-
"epoch": 0.06,
|
| 859 |
-
"learning_rate": 1.08843537414966e-05,
|
| 860 |
-
"loss": 1.5075,
|
| 861 |
-
"step": 142
|
| 862 |
-
},
|
| 863 |
-
{
|
| 864 |
-
"epoch": 0.06,
|
| 865 |
-
"learning_rate": 9.523809523809523e-06,
|
| 866 |
-
"loss": 1.4126,
|
| 867 |
-
"step": 143
|
| 868 |
-
},
|
| 869 |
-
{
|
| 870 |
-
"epoch": 0.06,
|
| 871 |
-
"learning_rate": 8.163265306122448e-06,
|
| 872 |
-
"loss": 1.4858,
|
| 873 |
-
"step": 144
|
| 874 |
-
},
|
| 875 |
-
{
|
| 876 |
-
"epoch": 0.06,
|
| 877 |
-
"learning_rate": 6.802721088435375e-06,
|
| 878 |
-
"loss": 1.4445,
|
| 879 |
-
"step": 145
|
| 880 |
-
},
|
| 881 |
-
{
|
| 882 |
-
"epoch": 0.06,
|
| 883 |
-
"learning_rate": 5.4421768707483e-06,
|
| 884 |
-
"loss": 1.2193,
|
| 885 |
-
"step": 146
|
| 886 |
-
},
|
| 887 |
-
{
|
| 888 |
-
"epoch": 0.06,
|
| 889 |
-
"learning_rate": 4.081632653061224e-06,
|
| 890 |
-
"loss": 1.5606,
|
| 891 |
-
"step": 147
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 0.06,
|
| 895 |
-
"learning_rate": 2.72108843537415e-06,
|
| 896 |
-
"loss": 1.5721,
|
| 897 |
-
"step": 148
|
| 898 |
-
},
|
| 899 |
-
{
|
| 900 |
-
"epoch": 0.06,
|
| 901 |
-
"learning_rate": 1.360544217687075e-06,
|
| 902 |
-
"loss": 1.8318,
|
| 903 |
-
"step": 149
|
| 904 |
-
},
|
| 905 |
-
{
|
| 906 |
-
"epoch": 0.06,
|
| 907 |
-
"learning_rate": 0.0,
|
| 908 |
-
"loss": 1.4401,
|
| 909 |
-
"step": 150
|
| 910 |
}
|
| 911 |
],
|
| 912 |
"logging_steps": 1,
|
| 913 |
-
"max_steps":
|
| 914 |
"num_train_epochs": 1,
|
| 915 |
"save_steps": 5,
|
| 916 |
-
"total_flos":
|
| 917 |
"trial_name": null,
|
| 918 |
"trial_params": null
|
| 919 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.006,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 15,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
"learning_rate": 0.0002,
|
| 26 |
+
"loss": 2.0719,
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
+
"learning_rate": 0.00019793814432989693,
|
| 32 |
+
"loss": 2.1789,
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
+
"learning_rate": 0.00019587628865979381,
|
| 38 |
+
"loss": 1.9318,
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
+
"learning_rate": 0.00019381443298969073,
|
| 44 |
+
"loss": 2.4083,
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
+
"learning_rate": 0.00019175257731958765,
|
| 50 |
+
"loss": 2.6823,
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
+
"learning_rate": 0.00018969072164948454,
|
| 56 |
+
"loss": 1.5946,
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.0,
|
| 61 |
+
"learning_rate": 0.00018762886597938145,
|
| 62 |
+
"loss": 2.0224,
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
+
"learning_rate": 0.00018556701030927837,
|
| 68 |
+
"loss": 2.0527,
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.0,
|
| 73 |
+
"learning_rate": 0.00018350515463917526,
|
| 74 |
+
"loss": 1.8021,
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.0,
|
| 79 |
+
"learning_rate": 0.00018144329896907217,
|
| 80 |
+
"loss": 2.0241,
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.01,
|
| 85 |
+
"learning_rate": 0.0001793814432989691,
|
| 86 |
+
"loss": 1.8196,
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.01,
|
| 91 |
+
"learning_rate": 0.00017731958762886598,
|
| 92 |
+
"loss": 1.8936,
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.01,
|
| 97 |
+
"learning_rate": 0.0001752577319587629,
|
| 98 |
+
"loss": 1.797,
|
| 99 |
"step": 15
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
}
|
| 101 |
],
|
| 102 |
"logging_steps": 1,
|
| 103 |
+
"max_steps": 100,
|
| 104 |
"num_train_epochs": 1,
|
| 105 |
"save_steps": 5,
|
| 106 |
+
"total_flos": 1808974570045440.0,
|
| 107 |
"trial_name": null,
|
| 108 |
"trial_params": null
|
| 109 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4600
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b9fc228c3c47641175a267c20a8bed63fa9082c03bc18b5aad785a75b8728f1
|
| 3 |
size 4600
|