Commit
·
042a5a5
1
Parent(s):
0c36757
Training in progress, step 15, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -19,12 +19,12 @@
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
| 22 |
-
"gate_proj",
|
| 23 |
-
"o_proj",
|
| 24 |
"k_proj",
|
| 25 |
-
"v_proj",
|
| 26 |
"down_proj",
|
|
|
|
|
|
|
| 27 |
"up_proj",
|
|
|
|
| 28 |
"q_proj"
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM"
|
|
|
|
| 19 |
"rank_pattern": {},
|
| 20 |
"revision": null,
|
| 21 |
"target_modules": [
|
|
|
|
|
|
|
| 22 |
"k_proj",
|
|
|
|
| 23 |
"down_proj",
|
| 24 |
+
"o_proj",
|
| 25 |
+
"v_proj",
|
| 26 |
"up_proj",
|
| 27 |
+
"gate_proj",
|
| 28 |
"q_proj"
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM"
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 83945296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e111b7672c5d2a8f54e7966b44c3966154845bb02803264d2c43c90f3a082e65
|
| 3 |
size 83945296
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 42545748
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9dc26a38ae8770b17827642290912a8d4383f5eea19af27d7c96b34013b80420
|
| 3 |
size 42545748
|
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:a81816f8c172759b66460138682effef4b6abce5a23baf230c81f044a0c8ede6
|
| 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 sha256:9d495c26c3e057e915a49e6c3755bab17f11955ce2bc01fcf07818c9d8f4dad2
|
| 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,
|
|
@@ -23,597 +23,87 @@
|
|
| 23 |
{
|
| 24 |
"epoch": 0.0,
|
| 25 |
"learning_rate": 0.0002,
|
| 26 |
-
"loss": 1.
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
-
"learning_rate": 0.
|
| 32 |
-
"loss": 1.
|
| 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": 1.
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
-
"learning_rate": 0.
|
| 68 |
-
"loss": 1.
|
| 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": 1.
|
| 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": 0.0001731958762886598,
|
| 104 |
-
"loss": 1.5251,
|
| 105 |
-
"step": 16
|
| 106 |
-
},
|
| 107 |
-
{
|
| 108 |
-
"epoch": 0.01,
|
| 109 |
-
"learning_rate": 0.0001711340206185567,
|
| 110 |
-
"loss": 1.7227,
|
| 111 |
-
"step": 17
|
| 112 |
-
},
|
| 113 |
-
{
|
| 114 |
-
"epoch": 0.01,
|
| 115 |
-
"learning_rate": 0.00016907216494845361,
|
| 116 |
-
"loss": 2.5511,
|
| 117 |
-
"step": 18
|
| 118 |
-
},
|
| 119 |
-
{
|
| 120 |
-
"epoch": 0.01,
|
| 121 |
-
"learning_rate": 0.00016701030927835053,
|
| 122 |
-
"loss": 1.4848,
|
| 123 |
-
"step": 19
|
| 124 |
-
},
|
| 125 |
-
{
|
| 126 |
-
"epoch": 0.01,
|
| 127 |
-
"learning_rate": 0.00016494845360824742,
|
| 128 |
-
"loss": 1.9532,
|
| 129 |
-
"step": 20
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.01,
|
| 133 |
-
"learning_rate": 0.00016288659793814434,
|
| 134 |
-
"loss": 1.879,
|
| 135 |
-
"step": 21
|
| 136 |
-
},
|
| 137 |
-
{
|
| 138 |
-
"epoch": 0.01,
|
| 139 |
-
"learning_rate": 0.00016082474226804125,
|
| 140 |
-
"loss": 1.7689,
|
| 141 |
-
"step": 22
|
| 142 |
-
},
|
| 143 |
-
{
|
| 144 |
-
"epoch": 0.01,
|
| 145 |
-
"learning_rate": 0.00015876288659793814,
|
| 146 |
-
"loss": 1.5917,
|
| 147 |
-
"step": 23
|
| 148 |
-
},
|
| 149 |
-
{
|
| 150 |
-
"epoch": 0.01,
|
| 151 |
-
"learning_rate": 0.00015670103092783506,
|
| 152 |
-
"loss": 1.9113,
|
| 153 |
-
"step": 24
|
| 154 |
-
},
|
| 155 |
-
{
|
| 156 |
-
"epoch": 0.01,
|
| 157 |
-
"learning_rate": 0.00015463917525773197,
|
| 158 |
-
"loss": 1.5997,
|
| 159 |
-
"step": 25
|
| 160 |
-
},
|
| 161 |
-
{
|
| 162 |
-
"epoch": 0.01,
|
| 163 |
-
"learning_rate": 0.00015257731958762886,
|
| 164 |
-
"loss": 1.7558,
|
| 165 |
-
"step": 26
|
| 166 |
-
},
|
| 167 |
-
{
|
| 168 |
-
"epoch": 0.01,
|
| 169 |
-
"learning_rate": 0.00015051546391752578,
|
| 170 |
-
"loss": 1.4693,
|
| 171 |
-
"step": 27
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.01,
|
| 175 |
-
"learning_rate": 0.0001484536082474227,
|
| 176 |
-
"loss": 1.498,
|
| 177 |
-
"step": 28
|
| 178 |
-
},
|
| 179 |
-
{
|
| 180 |
-
"epoch": 0.01,
|
| 181 |
-
"learning_rate": 0.00014639175257731958,
|
| 182 |
-
"loss": 1.6517,
|
| 183 |
-
"step": 29
|
| 184 |
-
},
|
| 185 |
-
{
|
| 186 |
-
"epoch": 0.01,
|
| 187 |
-
"learning_rate": 0.0001443298969072165,
|
| 188 |
-
"loss": 1.5836,
|
| 189 |
-
"step": 30
|
| 190 |
-
},
|
| 191 |
-
{
|
| 192 |
-
"epoch": 0.01,
|
| 193 |
-
"learning_rate": 0.00014226804123711342,
|
| 194 |
-
"loss": 1.7213,
|
| 195 |
-
"step": 31
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 0.01,
|
| 199 |
-
"learning_rate": 0.0001402061855670103,
|
| 200 |
-
"loss": 1.4432,
|
| 201 |
-
"step": 32
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 0.01,
|
| 205 |
-
"learning_rate": 0.00013814432989690722,
|
| 206 |
-
"loss": 2.1706,
|
| 207 |
-
"step": 33
|
| 208 |
-
},
|
| 209 |
-
{
|
| 210 |
-
"epoch": 0.01,
|
| 211 |
-
"learning_rate": 0.00013608247422680414,
|
| 212 |
-
"loss": 1.613,
|
| 213 |
-
"step": 34
|
| 214 |
-
},
|
| 215 |
-
{
|
| 216 |
-
"epoch": 0.01,
|
| 217 |
-
"learning_rate": 0.00013402061855670103,
|
| 218 |
-
"loss": 1.7608,
|
| 219 |
-
"step": 35
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 0.01,
|
| 223 |
-
"learning_rate": 0.00013195876288659794,
|
| 224 |
-
"loss": 1.6158,
|
| 225 |
-
"step": 36
|
| 226 |
-
},
|
| 227 |
-
{
|
| 228 |
-
"epoch": 0.01,
|
| 229 |
-
"learning_rate": 0.00012989690721649486,
|
| 230 |
-
"loss": 1.5101,
|
| 231 |
-
"step": 37
|
| 232 |
-
},
|
| 233 |
-
{
|
| 234 |
-
"epoch": 0.02,
|
| 235 |
-
"learning_rate": 0.00012783505154639175,
|
| 236 |
-
"loss": 1.5626,
|
| 237 |
-
"step": 38
|
| 238 |
-
},
|
| 239 |
-
{
|
| 240 |
-
"epoch": 0.02,
|
| 241 |
-
"learning_rate": 0.00012577319587628866,
|
| 242 |
-
"loss": 2.1865,
|
| 243 |
-
"step": 39
|
| 244 |
-
},
|
| 245 |
-
{
|
| 246 |
-
"epoch": 0.02,
|
| 247 |
-
"learning_rate": 0.00012371134020618558,
|
| 248 |
-
"loss": 1.4066,
|
| 249 |
-
"step": 40
|
| 250 |
-
},
|
| 251 |
-
{
|
| 252 |
-
"epoch": 0.02,
|
| 253 |
-
"learning_rate": 0.00012164948453608247,
|
| 254 |
-
"loss": 1.4324,
|
| 255 |
-
"step": 41
|
| 256 |
-
},
|
| 257 |
-
{
|
| 258 |
-
"epoch": 0.02,
|
| 259 |
-
"learning_rate": 0.00011958762886597938,
|
| 260 |
-
"loss": 1.7029,
|
| 261 |
-
"step": 42
|
| 262 |
-
},
|
| 263 |
-
{
|
| 264 |
-
"epoch": 0.02,
|
| 265 |
-
"learning_rate": 0.0001175257731958763,
|
| 266 |
-
"loss": 1.9695,
|
| 267 |
-
"step": 43
|
| 268 |
-
},
|
| 269 |
-
{
|
| 270 |
-
"epoch": 0.02,
|
| 271 |
-
"learning_rate": 0.00011546391752577319,
|
| 272 |
-
"loss": 1.7016,
|
| 273 |
-
"step": 44
|
| 274 |
-
},
|
| 275 |
-
{
|
| 276 |
-
"epoch": 0.02,
|
| 277 |
-
"learning_rate": 0.0001134020618556701,
|
| 278 |
-
"loss": 1.768,
|
| 279 |
-
"step": 45
|
| 280 |
-
},
|
| 281 |
-
{
|
| 282 |
-
"epoch": 0.02,
|
| 283 |
-
"learning_rate": 0.00011134020618556702,
|
| 284 |
-
"loss": 1.3867,
|
| 285 |
-
"step": 46
|
| 286 |
-
},
|
| 287 |
-
{
|
| 288 |
-
"epoch": 0.02,
|
| 289 |
-
"learning_rate": 0.00010927835051546391,
|
| 290 |
-
"loss": 1.8445,
|
| 291 |
-
"step": 47
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 0.02,
|
| 295 |
-
"learning_rate": 0.00010721649484536083,
|
| 296 |
-
"loss": 1.5515,
|
| 297 |
-
"step": 48
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 0.02,
|
| 301 |
-
"learning_rate": 0.00010515463917525774,
|
| 302 |
-
"loss": 1.7586,
|
| 303 |
-
"step": 49
|
| 304 |
-
},
|
| 305 |
-
{
|
| 306 |
-
"epoch": 0.02,
|
| 307 |
-
"learning_rate": 0.00010309278350515463,
|
| 308 |
-
"loss": 1.4246,
|
| 309 |
-
"step": 50
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 0.02,
|
| 313 |
-
"learning_rate": 0.00010103092783505155,
|
| 314 |
-
"loss": 1.7533,
|
| 315 |
-
"step": 51
|
| 316 |
-
},
|
| 317 |
-
{
|
| 318 |
-
"epoch": 0.02,
|
| 319 |
-
"learning_rate": 9.896907216494846e-05,
|
| 320 |
-
"loss": 1.6105,
|
| 321 |
-
"step": 52
|
| 322 |
-
},
|
| 323 |
-
{
|
| 324 |
-
"epoch": 0.02,
|
| 325 |
-
"learning_rate": 9.690721649484537e-05,
|
| 326 |
-
"loss": 1.5276,
|
| 327 |
-
"step": 53
|
| 328 |
-
},
|
| 329 |
-
{
|
| 330 |
-
"epoch": 0.02,
|
| 331 |
-
"learning_rate": 9.484536082474227e-05,
|
| 332 |
-
"loss": 1.3974,
|
| 333 |
-
"step": 54
|
| 334 |
-
},
|
| 335 |
-
{
|
| 336 |
-
"epoch": 0.02,
|
| 337 |
-
"learning_rate": 9.278350515463918e-05,
|
| 338 |
-
"loss": 1.4832,
|
| 339 |
-
"step": 55
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 0.02,
|
| 343 |
-
"learning_rate": 9.072164948453609e-05,
|
| 344 |
-
"loss": 1.695,
|
| 345 |
-
"step": 56
|
| 346 |
-
},
|
| 347 |
-
{
|
| 348 |
-
"epoch": 0.02,
|
| 349 |
-
"learning_rate": 8.865979381443299e-05,
|
| 350 |
-
"loss": 1.6554,
|
| 351 |
-
"step": 57
|
| 352 |
-
},
|
| 353 |
-
{
|
| 354 |
-
"epoch": 0.02,
|
| 355 |
-
"learning_rate": 8.65979381443299e-05,
|
| 356 |
-
"loss": 1.5769,
|
| 357 |
-
"step": 58
|
| 358 |
-
},
|
| 359 |
-
{
|
| 360 |
-
"epoch": 0.02,
|
| 361 |
-
"learning_rate": 8.453608247422681e-05,
|
| 362 |
-
"loss": 1.8362,
|
| 363 |
-
"step": 59
|
| 364 |
-
},
|
| 365 |
-
{
|
| 366 |
-
"epoch": 0.02,
|
| 367 |
-
"learning_rate": 8.247422680412371e-05,
|
| 368 |
-
"loss": 1.5832,
|
| 369 |
-
"step": 60
|
| 370 |
-
},
|
| 371 |
-
{
|
| 372 |
-
"epoch": 0.02,
|
| 373 |
-
"learning_rate": 8.041237113402063e-05,
|
| 374 |
-
"loss": 1.6469,
|
| 375 |
-
"step": 61
|
| 376 |
-
},
|
| 377 |
-
{
|
| 378 |
-
"epoch": 0.02,
|
| 379 |
-
"learning_rate": 7.835051546391753e-05,
|
| 380 |
-
"loss": 1.5835,
|
| 381 |
-
"step": 62
|
| 382 |
-
},
|
| 383 |
-
{
|
| 384 |
-
"epoch": 0.03,
|
| 385 |
-
"learning_rate": 7.628865979381443e-05,
|
| 386 |
-
"loss": 1.5521,
|
| 387 |
-
"step": 63
|
| 388 |
-
},
|
| 389 |
-
{
|
| 390 |
-
"epoch": 0.03,
|
| 391 |
-
"learning_rate": 7.422680412371135e-05,
|
| 392 |
-
"loss": 1.3237,
|
| 393 |
-
"step": 64
|
| 394 |
-
},
|
| 395 |
-
{
|
| 396 |
-
"epoch": 0.03,
|
| 397 |
-
"learning_rate": 7.216494845360825e-05,
|
| 398 |
-
"loss": 1.413,
|
| 399 |
-
"step": 65
|
| 400 |
-
},
|
| 401 |
-
{
|
| 402 |
-
"epoch": 0.03,
|
| 403 |
-
"learning_rate": 7.010309278350515e-05,
|
| 404 |
-
"loss": 1.9074,
|
| 405 |
-
"step": 66
|
| 406 |
-
},
|
| 407 |
-
{
|
| 408 |
-
"epoch": 0.03,
|
| 409 |
-
"learning_rate": 6.804123711340207e-05,
|
| 410 |
-
"loss": 1.5583,
|
| 411 |
-
"step": 67
|
| 412 |
-
},
|
| 413 |
-
{
|
| 414 |
-
"epoch": 0.03,
|
| 415 |
-
"learning_rate": 6.597938144329897e-05,
|
| 416 |
-
"loss": 1.5104,
|
| 417 |
-
"step": 68
|
| 418 |
-
},
|
| 419 |
-
{
|
| 420 |
-
"epoch": 0.03,
|
| 421 |
-
"learning_rate": 6.391752577319587e-05,
|
| 422 |
-
"loss": 1.6876,
|
| 423 |
-
"step": 69
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 0.03,
|
| 427 |
-
"learning_rate": 6.185567010309279e-05,
|
| 428 |
-
"loss": 1.5563,
|
| 429 |
-
"step": 70
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 0.03,
|
| 433 |
-
"learning_rate": 5.979381443298969e-05,
|
| 434 |
-
"loss": 2.1585,
|
| 435 |
-
"step": 71
|
| 436 |
-
},
|
| 437 |
-
{
|
| 438 |
-
"epoch": 0.03,
|
| 439 |
-
"learning_rate": 5.7731958762886594e-05,
|
| 440 |
-
"loss": 1.8855,
|
| 441 |
-
"step": 72
|
| 442 |
-
},
|
| 443 |
-
{
|
| 444 |
-
"epoch": 0.03,
|
| 445 |
-
"learning_rate": 5.567010309278351e-05,
|
| 446 |
-
"loss": 1.3202,
|
| 447 |
-
"step": 73
|
| 448 |
-
},
|
| 449 |
-
{
|
| 450 |
-
"epoch": 0.03,
|
| 451 |
-
"learning_rate": 5.360824742268041e-05,
|
| 452 |
-
"loss": 1.3524,
|
| 453 |
-
"step": 74
|
| 454 |
-
},
|
| 455 |
-
{
|
| 456 |
-
"epoch": 0.03,
|
| 457 |
-
"learning_rate": 5.1546391752577315e-05,
|
| 458 |
-
"loss": 1.4767,
|
| 459 |
-
"step": 75
|
| 460 |
-
},
|
| 461 |
-
{
|
| 462 |
-
"epoch": 0.03,
|
| 463 |
-
"learning_rate": 4.948453608247423e-05,
|
| 464 |
-
"loss": 1.8857,
|
| 465 |
-
"step": 76
|
| 466 |
-
},
|
| 467 |
-
{
|
| 468 |
-
"epoch": 0.03,
|
| 469 |
-
"learning_rate": 4.7422680412371134e-05,
|
| 470 |
-
"loss": 1.6848,
|
| 471 |
-
"step": 77
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 0.03,
|
| 475 |
-
"learning_rate": 4.536082474226804e-05,
|
| 476 |
-
"loss": 1.3765,
|
| 477 |
-
"step": 78
|
| 478 |
-
},
|
| 479 |
-
{
|
| 480 |
-
"epoch": 0.03,
|
| 481 |
-
"learning_rate": 4.329896907216495e-05,
|
| 482 |
-
"loss": 1.4421,
|
| 483 |
-
"step": 79
|
| 484 |
-
},
|
| 485 |
-
{
|
| 486 |
-
"epoch": 0.03,
|
| 487 |
-
"learning_rate": 4.1237113402061855e-05,
|
| 488 |
-
"loss": 1.5687,
|
| 489 |
-
"step": 80
|
| 490 |
-
},
|
| 491 |
-
{
|
| 492 |
-
"epoch": 0.03,
|
| 493 |
-
"learning_rate": 3.9175257731958764e-05,
|
| 494 |
-
"loss": 1.4066,
|
| 495 |
-
"step": 81
|
| 496 |
-
},
|
| 497 |
-
{
|
| 498 |
-
"epoch": 0.03,
|
| 499 |
-
"learning_rate": 3.7113402061855674e-05,
|
| 500 |
-
"loss": 1.6676,
|
| 501 |
-
"step": 82
|
| 502 |
-
},
|
| 503 |
-
{
|
| 504 |
-
"epoch": 0.03,
|
| 505 |
-
"learning_rate": 3.5051546391752576e-05,
|
| 506 |
-
"loss": 1.3302,
|
| 507 |
-
"step": 83
|
| 508 |
-
},
|
| 509 |
-
{
|
| 510 |
-
"epoch": 0.03,
|
| 511 |
-
"learning_rate": 3.2989690721649485e-05,
|
| 512 |
-
"loss": 2.014,
|
| 513 |
-
"step": 84
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 0.03,
|
| 517 |
-
"learning_rate": 3.0927835051546395e-05,
|
| 518 |
-
"loss": 1.6257,
|
| 519 |
-
"step": 85
|
| 520 |
-
},
|
| 521 |
-
{
|
| 522 |
-
"epoch": 0.03,
|
| 523 |
-
"learning_rate": 2.8865979381443297e-05,
|
| 524 |
-
"loss": 1.5508,
|
| 525 |
-
"step": 86
|
| 526 |
-
},
|
| 527 |
-
{
|
| 528 |
-
"epoch": 0.03,
|
| 529 |
-
"learning_rate": 2.6804123711340206e-05,
|
| 530 |
-
"loss": 1.4105,
|
| 531 |
-
"step": 87
|
| 532 |
-
},
|
| 533 |
-
{
|
| 534 |
-
"epoch": 0.04,
|
| 535 |
-
"learning_rate": 2.4742268041237116e-05,
|
| 536 |
-
"loss": 1.789,
|
| 537 |
-
"step": 88
|
| 538 |
-
},
|
| 539 |
-
{
|
| 540 |
-
"epoch": 0.04,
|
| 541 |
-
"learning_rate": 2.268041237113402e-05,
|
| 542 |
-
"loss": 1.7302,
|
| 543 |
-
"step": 89
|
| 544 |
-
},
|
| 545 |
-
{
|
| 546 |
-
"epoch": 0.04,
|
| 547 |
-
"learning_rate": 2.0618556701030927e-05,
|
| 548 |
-
"loss": 1.7847,
|
| 549 |
-
"step": 90
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 0.04,
|
| 553 |
-
"learning_rate": 1.8556701030927837e-05,
|
| 554 |
-
"loss": 1.489,
|
| 555 |
-
"step": 91
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 0.04,
|
| 559 |
-
"learning_rate": 1.6494845360824743e-05,
|
| 560 |
-
"loss": 1.0728,
|
| 561 |
-
"step": 92
|
| 562 |
-
},
|
| 563 |
-
{
|
| 564 |
-
"epoch": 0.04,
|
| 565 |
-
"learning_rate": 1.4432989690721649e-05,
|
| 566 |
-
"loss": 1.5944,
|
| 567 |
-
"step": 93
|
| 568 |
-
},
|
| 569 |
-
{
|
| 570 |
-
"epoch": 0.04,
|
| 571 |
-
"learning_rate": 1.2371134020618558e-05,
|
| 572 |
-
"loss": 2.0057,
|
| 573 |
-
"step": 94
|
| 574 |
-
},
|
| 575 |
-
{
|
| 576 |
-
"epoch": 0.04,
|
| 577 |
-
"learning_rate": 1.0309278350515464e-05,
|
| 578 |
-
"loss": 1.481,
|
| 579 |
-
"step": 95
|
| 580 |
-
},
|
| 581 |
-
{
|
| 582 |
-
"epoch": 0.04,
|
| 583 |
-
"learning_rate": 8.247422680412371e-06,
|
| 584 |
-
"loss": 1.478,
|
| 585 |
-
"step": 96
|
| 586 |
-
},
|
| 587 |
-
{
|
| 588 |
-
"epoch": 0.04,
|
| 589 |
-
"learning_rate": 6.185567010309279e-06,
|
| 590 |
-
"loss": 1.6756,
|
| 591 |
-
"step": 97
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 0.04,
|
| 595 |
-
"learning_rate": 4.123711340206186e-06,
|
| 596 |
-
"loss": 1.6028,
|
| 597 |
-
"step": 98
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 0.04,
|
| 601 |
-
"learning_rate": 2.061855670103093e-06,
|
| 602 |
-
"loss": 1.6606,
|
| 603 |
-
"step": 99
|
| 604 |
-
},
|
| 605 |
-
{
|
| 606 |
-
"epoch": 0.04,
|
| 607 |
-
"learning_rate": 0.0,
|
| 608 |
-
"loss": 1.4798,
|
| 609 |
-
"step": 100
|
| 610 |
}
|
| 611 |
],
|
| 612 |
"logging_steps": 1,
|
| 613 |
-
"max_steps":
|
| 614 |
"num_train_epochs": 1,
|
| 615 |
"save_steps": 5,
|
| 616 |
-
"total_flos":
|
| 617 |
"trial_name": null,
|
| 618 |
"trial_params": null
|
| 619 |
}
|
|
|
|
| 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": 1.6718,
|
| 27 |
"step": 3
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 0.0,
|
| 31 |
+
"learning_rate": 0.00019722222222222225,
|
| 32 |
+
"loss": 1.9564,
|
| 33 |
"step": 4
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 0.0,
|
| 37 |
+
"learning_rate": 0.00019444444444444446,
|
| 38 |
+
"loss": 1.6946,
|
| 39 |
"step": 5
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 0.0,
|
| 43 |
+
"learning_rate": 0.00019166666666666667,
|
| 44 |
+
"loss": 2.2475,
|
| 45 |
"step": 6
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 0.0,
|
| 49 |
+
"learning_rate": 0.00018888888888888888,
|
| 50 |
+
"loss": 2.4487,
|
| 51 |
"step": 7
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.0,
|
| 55 |
+
"learning_rate": 0.00018611111111111112,
|
| 56 |
+
"loss": 1.4369,
|
| 57 |
"step": 8
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 0.0,
|
| 61 |
+
"learning_rate": 0.00018333333333333334,
|
| 62 |
+
"loss": 1.9525,
|
| 63 |
"step": 9
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 0.0,
|
| 67 |
+
"learning_rate": 0.00018055555555555557,
|
| 68 |
+
"loss": 1.9572,
|
| 69 |
"step": 10
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 0.0,
|
| 73 |
+
"learning_rate": 0.00017777777777777779,
|
| 74 |
+
"loss": 1.727,
|
| 75 |
"step": 11
|
| 76 |
},
|
| 77 |
{
|
| 78 |
"epoch": 0.0,
|
| 79 |
+
"learning_rate": 0.000175,
|
| 80 |
+
"loss": 1.9873,
|
| 81 |
"step": 12
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 0.01,
|
| 85 |
+
"learning_rate": 0.00017222222222222224,
|
| 86 |
+
"loss": 1.7861,
|
| 87 |
"step": 13
|
| 88 |
},
|
| 89 |
{
|
| 90 |
"epoch": 0.01,
|
| 91 |
+
"learning_rate": 0.00016944444444444445,
|
| 92 |
+
"loss": 1.7475,
|
| 93 |
"step": 14
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.01,
|
| 97 |
+
"learning_rate": 0.0001666666666666667,
|
| 98 |
+
"loss": 1.7548,
|
| 99 |
"step": 15
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
}
|
| 101 |
],
|
| 102 |
"logging_steps": 1,
|
| 103 |
+
"max_steps": 75,
|
| 104 |
"num_train_epochs": 1,
|
| 105 |
"save_steps": 5,
|
| 106 |
+
"total_flos": 960031721619456.0,
|
| 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:b27b54f4df6b3ea01a348f2fcdb61d482edb9033e83476fc3a25b4bfa991e681
|
| 3 |
size 4600
|