Wangpeng An commited on
Upload folder using huggingface_hub
Browse files- config.json +1 -1
- model-00001-of-00004.safetensors +1 -1
- model-00002-of-00004.safetensors +1 -1
- model-00003-of-00004.safetensors +1 -1
- model-00004-of-00004.safetensors +1 -1
- trainer_state.json +394 -394
- training_args.bin +1 -1
config.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
"X": [
|
| 3 |
"Video"
|
| 4 |
],
|
| 5 |
-
"_name_or_path": "/mnt/bn/tns-algo-video-public-my2/wangpeng.an/model/
|
| 6 |
"architectures": [
|
| 7 |
"LlavaLlamaForCausalLM"
|
| 8 |
],
|
|
|
|
| 2 |
"X": [
|
| 3 |
"Video"
|
| 4 |
],
|
| 5 |
+
"_name_or_path": "/mnt/bn/tns-algo-video-public-my2/wangpeng.an/model/pretrained_model/omnis3",
|
| 6 |
"architectures": [
|
| 7 |
"LlavaLlamaForCausalLM"
|
| 8 |
],
|
model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4976698672
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d3037fc3fcae7f29a27ca3389e8340e374e497ddf883f06aba385121871a91b
|
| 3 |
size 4976698672
|
model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999802720
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b3eb0fe3b8b62fdd8e7988d2f613e5208e9929b10a5e0977dd1a0c934845460c
|
| 3 |
size 4999802720
|
model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4915916176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3aed1673818f58b983377abded8f1bf2576b59aeed4e764d76eed55cbf9ca68a
|
| 3 |
size 4915916176
|
model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3851682320
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2d40974ba0e547116c5e6dd968d1e5618b8fcd7983d1e33feed37b804659034
|
| 3 |
size 3851682320
|
trainer_state.json
CHANGED
|
@@ -12,2732 +12,2732 @@
|
|
| 12 |
"epoch": 0.002570694087403599,
|
| 13 |
"grad_norm": 0.0,
|
| 14 |
"learning_rate": 1.6666666666666667e-06,
|
| 15 |
-
"loss":
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.005141388174807198,
|
| 20 |
"grad_norm": 0.0,
|
| 21 |
"learning_rate": 3.3333333333333333e-06,
|
| 22 |
-
"loss":
|
| 23 |
"step": 2
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.007712082262210797,
|
| 27 |
"grad_norm": 0.0,
|
| 28 |
"learning_rate": 5e-06,
|
| 29 |
-
"loss":
|
| 30 |
"step": 3
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.010282776349614395,
|
| 34 |
"grad_norm": 0.0,
|
| 35 |
"learning_rate": 6.666666666666667e-06,
|
| 36 |
-
"loss":
|
| 37 |
"step": 4
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.012853470437017995,
|
| 41 |
"grad_norm": 0.0,
|
| 42 |
"learning_rate": 8.333333333333334e-06,
|
| 43 |
-
"loss":
|
| 44 |
"step": 5
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.015424164524421594,
|
| 48 |
"grad_norm": 0.0,
|
| 49 |
"learning_rate": 1e-05,
|
| 50 |
-
"loss":
|
| 51 |
"step": 6
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.017994858611825194,
|
| 55 |
"grad_norm": 0.0,
|
| 56 |
"learning_rate": 1.1666666666666668e-05,
|
| 57 |
-
"loss":
|
| 58 |
"step": 7
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.02056555269922879,
|
| 62 |
"grad_norm": 0.0,
|
| 63 |
"learning_rate": 1.3333333333333333e-05,
|
| 64 |
-
"loss":
|
| 65 |
"step": 8
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.02313624678663239,
|
| 69 |
"grad_norm": 0.0,
|
| 70 |
"learning_rate": 1.5000000000000002e-05,
|
| 71 |
-
"loss":
|
| 72 |
"step": 9
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.02570694087403599,
|
| 76 |
"grad_norm": 0.0,
|
| 77 |
"learning_rate": 1.6666666666666667e-05,
|
| 78 |
-
"loss":
|
| 79 |
"step": 10
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.028277634961439587,
|
| 83 |
"grad_norm": 0.0,
|
| 84 |
"learning_rate": 1.8333333333333333e-05,
|
| 85 |
-
"loss":
|
| 86 |
"step": 11
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.030848329048843187,
|
| 90 |
"grad_norm": 0.0,
|
| 91 |
"learning_rate": 2e-05,
|
| 92 |
-
"loss":
|
| 93 |
"step": 12
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.033419023136246784,
|
| 97 |
"grad_norm": 0.0,
|
| 98 |
"learning_rate": 1.9999652796146877e-05,
|
| 99 |
-
"loss":
|
| 100 |
"step": 13
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.03598971722365039,
|
| 104 |
"grad_norm": 0.0,
|
| 105 |
"learning_rate": 1.9998611208697607e-05,
|
| 106 |
-
"loss":
|
| 107 |
"step": 14
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.038560411311053984,
|
| 111 |
"grad_norm": 0.0,
|
| 112 |
"learning_rate": 1.9996875309980824e-05,
|
| 113 |
-
"loss":
|
| 114 |
"step": 15
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.04113110539845758,
|
| 118 |
"grad_norm": 0.0,
|
| 119 |
"learning_rate": 1.9994445220538678e-05,
|
| 120 |
-
"loss":
|
| 121 |
"step": 16
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.043701799485861184,
|
| 125 |
"grad_norm": 0.0,
|
| 126 |
"learning_rate": 1.999132110911845e-05,
|
| 127 |
-
"loss":
|
| 128 |
"step": 17
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.04627249357326478,
|
| 132 |
"grad_norm": 0.0,
|
| 133 |
"learning_rate": 1.9987503192660842e-05,
|
| 134 |
-
"loss":
|
| 135 |
"step": 18
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.04884318766066838,
|
| 139 |
"grad_norm": 0.0,
|
| 140 |
"learning_rate": 1.9982991736284914e-05,
|
| 141 |
-
"loss":
|
| 142 |
"step": 19
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.05141388174807198,
|
| 146 |
"grad_norm": 0.0,
|
| 147 |
"learning_rate": 1.997778705326968e-05,
|
| 148 |
-
"loss":
|
| 149 |
"step": 20
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.05398457583547558,
|
| 153 |
"grad_norm": 0.0,
|
| 154 |
"learning_rate": 1.9971889505032337e-05,
|
| 155 |
-
"loss":
|
| 156 |
"step": 21
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.056555269922879174,
|
| 160 |
"grad_norm": 0.0,
|
| 161 |
"learning_rate": 1.9965299501103178e-05,
|
| 162 |
-
"loss":
|
| 163 |
"step": 22
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.05912596401028278,
|
| 167 |
"grad_norm": 0.0,
|
| 168 |
"learning_rate": 1.995801749909715e-05,
|
| 169 |
-
"loss":
|
| 170 |
"step": 23
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.061696658097686374,
|
| 174 |
"grad_norm": 0.0,
|
| 175 |
"learning_rate": 1.995004400468209e-05,
|
| 176 |
-
"loss":
|
| 177 |
"step": 24
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.06426735218508997,
|
| 181 |
"grad_norm": 0.0,
|
| 182 |
"learning_rate": 1.9941379571543597e-05,
|
| 183 |
-
"loss":
|
| 184 |
"step": 25
|
| 185 |
},
|
| 186 |
{
|
| 187 |
"epoch": 0.06683804627249357,
|
| 188 |
"grad_norm": 0.0,
|
| 189 |
"learning_rate": 1.9932024801346583e-05,
|
| 190 |
-
"loss":
|
| 191 |
"step": 26
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.06940874035989718,
|
| 195 |
"grad_norm": 0.0,
|
| 196 |
"learning_rate": 1.992198034369349e-05,
|
| 197 |
-
"loss":
|
| 198 |
"step": 27
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"epoch": 0.07197943444730077,
|
| 202 |
"grad_norm": 0.0,
|
| 203 |
"learning_rate": 1.991124689607921e-05,
|
| 204 |
-
"loss":
|
| 205 |
"step": 28
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.07455012853470437,
|
| 209 |
"grad_norm": 0.0,
|
| 210 |
"learning_rate": 1.9899825203842613e-05,
|
| 211 |
-
"loss":
|
| 212 |
"step": 29
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.07712082262210797,
|
| 216 |
"grad_norm": 0.0,
|
| 217 |
"learning_rate": 1.988771606011481e-05,
|
| 218 |
-
"loss":
|
| 219 |
"step": 30
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.07969151670951156,
|
| 223 |
"grad_norm": 0.0,
|
| 224 |
"learning_rate": 1.987492030576407e-05,
|
| 225 |
-
"loss":
|
| 226 |
"step": 31
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.08226221079691516,
|
| 230 |
"grad_norm": 0.0,
|
| 231 |
"learning_rate": 1.986143882933744e-05,
|
| 232 |
-
"loss":
|
| 233 |
"step": 32
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.08483290488431877,
|
| 237 |
"grad_norm": 0.0,
|
| 238 |
"learning_rate": 1.9847272566999026e-05,
|
| 239 |
-
"loss":
|
| 240 |
"step": 33
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 0.08740359897172237,
|
| 244 |
"grad_norm": 0.0,
|
| 245 |
"learning_rate": 1.9832422502465013e-05,
|
| 246 |
-
"loss":
|
| 247 |
"step": 34
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.08997429305912596,
|
| 251 |
"grad_norm": 0.0,
|
| 252 |
"learning_rate": 1.9816889666935318e-05,
|
| 253 |
-
"loss":
|
| 254 |
"step": 35
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.09254498714652956,
|
| 258 |
"grad_norm": 0.0,
|
| 259 |
"learning_rate": 1.9800675139022006e-05,
|
| 260 |
-
"loss":
|
| 261 |
"step": 36
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.09511568123393316,
|
| 265 |
"grad_norm": 0.0,
|
| 266 |
"learning_rate": 1.9783780044674402e-05,
|
| 267 |
-
"loss":
|
| 268 |
"step": 37
|
| 269 |
},
|
| 270 |
{
|
| 271 |
"epoch": 0.09768637532133675,
|
| 272 |
"grad_norm": 0.0,
|
| 273 |
"learning_rate": 1.976620555710087e-05,
|
| 274 |
-
"loss":
|
| 275 |
"step": 38
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.10025706940874037,
|
| 279 |
"grad_norm": 0.0,
|
| 280 |
"learning_rate": 1.974795289668737e-05,
|
| 281 |
-
"loss":
|
| 282 |
"step": 39
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 0.10282776349614396,
|
| 286 |
"grad_norm": 0.0,
|
| 287 |
"learning_rate": 1.972902333091271e-05,
|
| 288 |
-
"loss":
|
| 289 |
"step": 40
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.10539845758354756,
|
| 293 |
"grad_norm": 0.0,
|
| 294 |
"learning_rate": 1.9709418174260523e-05,
|
| 295 |
-
"loss": 1.
|
| 296 |
"step": 41
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.10796915167095116,
|
| 300 |
"grad_norm": 0.0,
|
| 301 |
"learning_rate": 1.9689138788127994e-05,
|
| 302 |
-
"loss": 1.
|
| 303 |
"step": 42
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.11053984575835475,
|
| 307 |
"grad_norm": 0.0,
|
| 308 |
"learning_rate": 1.966818658073133e-05,
|
| 309 |
-
"loss": 1.
|
| 310 |
"step": 43
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 0.11311053984575835,
|
| 314 |
"grad_norm": 0.0,
|
| 315 |
"learning_rate": 1.9646563007007952e-05,
|
| 316 |
-
"loss":
|
| 317 |
"step": 44
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 0.11568123393316196,
|
| 321 |
"grad_norm": 0.0,
|
| 322 |
"learning_rate": 1.9624269568515486e-05,
|
| 323 |
-
"loss":
|
| 324 |
"step": 45
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.11825192802056556,
|
| 328 |
"grad_norm": 0.0,
|
| 329 |
"learning_rate": 1.960130781332748e-05,
|
| 330 |
-
"loss": 1.
|
| 331 |
"step": 46
|
| 332 |
},
|
| 333 |
{
|
| 334 |
"epoch": 0.12082262210796915,
|
| 335 |
"grad_norm": 0.0,
|
| 336 |
"learning_rate": 1.957767933592591e-05,
|
| 337 |
-
"loss":
|
| 338 |
"step": 47
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.12339331619537275,
|
| 342 |
"grad_norm": 0.0,
|
| 343 |
"learning_rate": 1.955338577709046e-05,
|
| 344 |
-
"loss":
|
| 345 |
"step": 48
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.12596401028277635,
|
| 349 |
"grad_norm": 0.0,
|
| 350 |
"learning_rate": 1.9528428823784567e-05,
|
| 351 |
-
"loss": 1.
|
| 352 |
"step": 49
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.12853470437017994,
|
| 356 |
"grad_norm": 0.0,
|
| 357 |
"learning_rate": 1.9502810209038302e-05,
|
| 358 |
-
"loss": 1.
|
| 359 |
"step": 50
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 0.13110539845758354,
|
| 363 |
"grad_norm": 0.0,
|
| 364 |
"learning_rate": 1.9476531711828027e-05,
|
| 365 |
-
"loss": 1.
|
| 366 |
"step": 51
|
| 367 |
},
|
| 368 |
{
|
| 369 |
"epoch": 0.13367609254498714,
|
| 370 |
"grad_norm": 0.0,
|
| 371 |
"learning_rate": 1.9449595156952827e-05,
|
| 372 |
-
"loss": 1.
|
| 373 |
"step": 52
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 0.13624678663239073,
|
| 377 |
"grad_norm": 0.0,
|
| 378 |
"learning_rate": 1.9422002414907837e-05,
|
| 379 |
-
"loss": 1.
|
| 380 |
"step": 53
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.13881748071979436,
|
| 384 |
"grad_norm": 0.0,
|
| 385 |
"learning_rate": 1.9393755401754324e-05,
|
| 386 |
-
"loss": 1.
|
| 387 |
"step": 54
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.14138817480719795,
|
| 391 |
"grad_norm": 0.0,
|
| 392 |
"learning_rate": 1.936485607898665e-05,
|
| 393 |
-
"loss":
|
| 394 |
"step": 55
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 0.14395886889460155,
|
| 398 |
"grad_norm": 0.0,
|
| 399 |
"learning_rate": 1.9335306453396066e-05,
|
| 400 |
-
"loss": 1.
|
| 401 |
"step": 56
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 0.14652956298200515,
|
| 405 |
"grad_norm": 0.0,
|
| 406 |
"learning_rate": 1.9305108576931336e-05,
|
| 407 |
-
"loss": 1.
|
| 408 |
"step": 57
|
| 409 |
},
|
| 410 |
{
|
| 411 |
"epoch": 0.14910025706940874,
|
| 412 |
"grad_norm": 0.0,
|
| 413 |
"learning_rate": 1.927426454655627e-05,
|
| 414 |
-
"loss": 1.
|
| 415 |
"step": 58
|
| 416 |
},
|
| 417 |
{
|
| 418 |
"epoch": 0.15167095115681234,
|
| 419 |
"grad_norm": 0.0,
|
| 420 |
"learning_rate": 1.924277650410412e-05,
|
| 421 |
-
"loss": 1.
|
| 422 |
"step": 59
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.15424164524421594,
|
| 426 |
"grad_norm": 0.0,
|
| 427 |
"learning_rate": 1.9210646636128805e-05,
|
| 428 |
-
"loss":
|
| 429 |
"step": 60
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.15681233933161953,
|
| 433 |
"grad_norm": 0.0,
|
| 434 |
"learning_rate": 1.9177877173753127e-05,
|
| 435 |
-
"loss": 1.
|
| 436 |
"step": 61
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.15938303341902313,
|
| 440 |
"grad_norm": 0.0,
|
| 441 |
"learning_rate": 1.91444703925138e-05,
|
| 442 |
-
"loss": 1.
|
| 443 |
"step": 62
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 0.16195372750642673,
|
| 447 |
"grad_norm": 0.0,
|
| 448 |
"learning_rate": 1.9110428612203463e-05,
|
| 449 |
-
"loss": 1.
|
| 450 |
"step": 63
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"epoch": 0.16452442159383032,
|
| 454 |
"grad_norm": 0.0,
|
| 455 |
"learning_rate": 1.9075754196709574e-05,
|
| 456 |
-
"loss": 1.
|
| 457 |
"step": 64
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.16709511568123395,
|
| 461 |
"grad_norm": 0.0,
|
| 462 |
"learning_rate": 1.904044955385026e-05,
|
| 463 |
-
"loss": 1.
|
| 464 |
"step": 65
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.16966580976863754,
|
| 468 |
"grad_norm": 0.0,
|
| 469 |
"learning_rate": 1.9004517135207127e-05,
|
| 470 |
-
"loss": 1.
|
| 471 |
"step": 66
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.17223650385604114,
|
| 475 |
"grad_norm": 0.0,
|
| 476 |
"learning_rate": 1.8967959435955027e-05,
|
| 477 |
-
"loss": 1.
|
| 478 |
"step": 67
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.17480719794344474,
|
| 482 |
"grad_norm": 0.0,
|
| 483 |
"learning_rate": 1.893077899468876e-05,
|
| 484 |
-
"loss": 1.
|
| 485 |
"step": 68
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 0.17737789203084833,
|
| 489 |
"grad_norm": 0.0,
|
| 490 |
"learning_rate": 1.889297839324682e-05,
|
| 491 |
-
"loss": 1.
|
| 492 |
"step": 69
|
| 493 |
},
|
| 494 |
{
|
| 495 |
"epoch": 0.17994858611825193,
|
| 496 |
"grad_norm": 0.0,
|
| 497 |
"learning_rate": 1.8854560256532098e-05,
|
| 498 |
-
"loss": 1.
|
| 499 |
"step": 70
|
| 500 |
},
|
| 501 |
{
|
| 502 |
"epoch": 0.18251928020565553,
|
| 503 |
"grad_norm": 0.0,
|
| 504 |
"learning_rate": 1.8815527252329624e-05,
|
| 505 |
-
"loss": 1.
|
| 506 |
"step": 71
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.18508997429305912,
|
| 510 |
"grad_norm": 0.0,
|
| 511 |
"learning_rate": 1.8775882091121282e-05,
|
| 512 |
-
"loss": 1.
|
| 513 |
"step": 72
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.18766066838046272,
|
| 517 |
"grad_norm": 0.0,
|
| 518 |
"learning_rate": 1.8735627525897618e-05,
|
| 519 |
-
"loss": 1.
|
| 520 |
"step": 73
|
| 521 |
},
|
| 522 |
{
|
| 523 |
"epoch": 0.19023136246786632,
|
| 524 |
"grad_norm": 0.0,
|
| 525 |
"learning_rate": 1.8694766351966665e-05,
|
| 526 |
-
"loss": 1.
|
| 527 |
"step": 74
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 0.1928020565552699,
|
| 531 |
"grad_norm": 0.0,
|
| 532 |
"learning_rate": 1.8653301406759827e-05,
|
| 533 |
-
"loss": 1.
|
| 534 |
"step": 75
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.1953727506426735,
|
| 538 |
"grad_norm": 0.0,
|
| 539 |
"learning_rate": 1.8611235569634852e-05,
|
| 540 |
-
"loss": 1.
|
| 541 |
"step": 76
|
| 542 |
},
|
| 543 |
{
|
| 544 |
"epoch": 0.19794344473007713,
|
| 545 |
"grad_norm": 0.0,
|
| 546 |
"learning_rate": 1.8568571761675893e-05,
|
| 547 |
-
"loss": 1.
|
| 548 |
"step": 77
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.20051413881748073,
|
| 552 |
"grad_norm": 0.0,
|
| 553 |
"learning_rate": 1.8525312945490647e-05,
|
| 554 |
-
"loss": 1.
|
| 555 |
"step": 78
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.20308483290488433,
|
| 559 |
"grad_norm": 0.0,
|
| 560 |
"learning_rate": 1.8481462125004647e-05,
|
| 561 |
-
"loss": 1.
|
| 562 |
"step": 79
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 0.20565552699228792,
|
| 566 |
"grad_norm": 0.0,
|
| 567 |
"learning_rate": 1.8437022345252666e-05,
|
| 568 |
-
"loss": 1.
|
| 569 |
"step": 80
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.20822622107969152,
|
| 573 |
"grad_norm": 0.0,
|
| 574 |
"learning_rate": 1.8391996692167242e-05,
|
| 575 |
-
"loss": 1.
|
| 576 |
"step": 81
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"epoch": 0.21079691516709512,
|
| 580 |
"grad_norm": 0.0,
|
| 581 |
"learning_rate": 1.8346388292364438e-05,
|
| 582 |
-
"loss": 1.
|
| 583 |
"step": 82
|
| 584 |
},
|
| 585 |
{
|
| 586 |
"epoch": 0.2133676092544987,
|
| 587 |
"grad_norm": 0.0,
|
| 588 |
"learning_rate": 1.8300200312926674e-05,
|
| 589 |
-
"loss": 1.
|
| 590 |
"step": 83
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.2159383033419023,
|
| 594 |
"grad_norm": 0.0,
|
| 595 |
"learning_rate": 1.8253435961182844e-05,
|
| 596 |
-
"loss": 1.
|
| 597 |
"step": 84
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.2185089974293059,
|
| 601 |
"grad_norm": 0.0,
|
| 602 |
"learning_rate": 1.8206098484485563e-05,
|
| 603 |
-
"loss": 1.
|
| 604 |
"step": 85
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.2210796915167095,
|
| 608 |
"grad_norm": 0.0,
|
| 609 |
"learning_rate": 1.8158191169985696e-05,
|
| 610 |
-
"loss": 1.
|
| 611 |
"step": 86
|
| 612 |
},
|
| 613 |
{
|
| 614 |
"epoch": 0.2236503856041131,
|
| 615 |
"grad_norm": 0.0,
|
| 616 |
"learning_rate": 1.810971734440408e-05,
|
| 617 |
-
"loss": 1.
|
| 618 |
"step": 87
|
| 619 |
},
|
| 620 |
{
|
| 621 |
"epoch": 0.2262210796915167,
|
| 622 |
"grad_norm": 0.0,
|
| 623 |
"learning_rate": 1.806068037380052e-05,
|
| 624 |
-
"loss": 1.
|
| 625 |
"step": 88
|
| 626 |
},
|
| 627 |
{
|
| 628 |
"epoch": 0.22879177377892032,
|
| 629 |
"grad_norm": 0.0,
|
| 630 |
"learning_rate": 1.801108366334004e-05,
|
| 631 |
-
"loss": 1.
|
| 632 |
"step": 89
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.23136246786632392,
|
| 636 |
"grad_norm": 0.0,
|
| 637 |
"learning_rate": 1.796093065705644e-05,
|
| 638 |
-
"loss": 1.
|
| 639 |
"step": 90
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.23393316195372751,
|
| 643 |
"grad_norm": 0.0,
|
| 644 |
"learning_rate": 1.791022483761312e-05,
|
| 645 |
-
"loss": 1.
|
| 646 |
"step": 91
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 0.2365038560411311,
|
| 650 |
"grad_norm": 0.0,
|
| 651 |
"learning_rate": 1.7858969726061262e-05,
|
| 652 |
-
"loss": 1.
|
| 653 |
"step": 92
|
| 654 |
},
|
| 655 |
{
|
| 656 |
"epoch": 0.2390745501285347,
|
| 657 |
"grad_norm": 0.0,
|
| 658 |
"learning_rate": 1.7807168881595304e-05,
|
| 659 |
-
"loss": 1.
|
| 660 |
"step": 93
|
| 661 |
},
|
| 662 |
{
|
| 663 |
"epoch": 0.2416452442159383,
|
| 664 |
"grad_norm": 0.0,
|
| 665 |
"learning_rate": 1.7754825901305814e-05,
|
| 666 |
-
"loss": 1.
|
| 667 |
"step": 94
|
| 668 |
},
|
| 669 |
{
|
| 670 |
"epoch": 0.2442159383033419,
|
| 671 |
"grad_norm": 0.0,
|
| 672 |
"learning_rate": 1.7701944419929673e-05,
|
| 673 |
-
"loss": 1.
|
| 674 |
"step": 95
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 0.2467866323907455,
|
| 678 |
"grad_norm": 0.0,
|
| 679 |
"learning_rate": 1.7648528109597704e-05,
|
| 680 |
-
"loss": 1.
|
| 681 |
"step": 96
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.2493573264781491,
|
| 685 |
"grad_norm": 0.0,
|
| 686 |
"learning_rate": 1.7594580679579654e-05,
|
| 687 |
-
"loss": 1.
|
| 688 |
"step": 97
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 0.2519280205655527,
|
| 692 |
"grad_norm": 0.0,
|
| 693 |
"learning_rate": 1.7540105876026647e-05,
|
| 694 |
-
"loss": 1.
|
| 695 |
"step": 98
|
| 696 |
},
|
| 697 |
{
|
| 698 |
"epoch": 0.2544987146529563,
|
| 699 |
"grad_norm": 0.0,
|
| 700 |
"learning_rate": 1.7485107481711014e-05,
|
| 701 |
-
"loss": 1.
|
| 702 |
"step": 99
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"epoch": 0.2570694087403599,
|
| 706 |
"grad_norm": 0.0,
|
| 707 |
"learning_rate": 1.7429589315763637e-05,
|
| 708 |
-
"loss": 1.
|
| 709 |
"step": 100
|
| 710 |
},
|
| 711 |
{
|
| 712 |
"epoch": 0.2596401028277635,
|
| 713 |
"grad_norm": 0.0,
|
| 714 |
"learning_rate": 1.737355523340875e-05,
|
| 715 |
-
"loss": 1.
|
| 716 |
"step": 101
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 0.2622107969151671,
|
| 720 |
"grad_norm": 0.0,
|
| 721 |
"learning_rate": 1.7317009125696208e-05,
|
| 722 |
-
"loss": 1.
|
| 723 |
"step": 102
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.2647814910025707,
|
| 727 |
"grad_norm": 0.0,
|
| 728 |
"learning_rate": 1.725995491923131e-05,
|
| 729 |
-
"loss": 1.
|
| 730 |
"step": 103
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.26735218508997427,
|
| 734 |
"grad_norm": 0.0,
|
| 735 |
"learning_rate": 1.7202396575902118e-05,
|
| 736 |
-
"loss": 1.
|
| 737 |
"step": 104
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 0.2699228791773779,
|
| 741 |
"grad_norm": 0.0,
|
| 742 |
"learning_rate": 1.714433809260435e-05,
|
| 743 |
-
"loss": 1.
|
| 744 |
"step": 105
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.27249357326478146,
|
| 748 |
"grad_norm": 0.0,
|
| 749 |
"learning_rate": 1.7085783500963825e-05,
|
| 750 |
-
"loss": 1.
|
| 751 |
"step": 106
|
| 752 |
},
|
| 753 |
{
|
| 754 |
"epoch": 0.2750642673521851,
|
| 755 |
"grad_norm": 0.0,
|
| 756 |
"learning_rate": 1.702673686705651e-05,
|
| 757 |
-
"loss": 1.
|
| 758 |
"step": 107
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 0.2776349614395887,
|
| 762 |
"grad_norm": 0.0,
|
| 763 |
"learning_rate": 1.6967202291126174e-05,
|
| 764 |
-
"loss": 1.
|
| 765 |
"step": 108
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.2802056555269923,
|
| 769 |
"grad_norm": 0.0,
|
| 770 |
"learning_rate": 1.690718390729964e-05,
|
| 771 |
-
"loss": 1.
|
| 772 |
"step": 109
|
| 773 |
},
|
| 774 |
{
|
| 775 |
"epoch": 0.2827763496143959,
|
| 776 |
"grad_norm": 0.0,
|
| 777 |
"learning_rate": 1.684668588329973e-05,
|
| 778 |
-
"loss": 1.
|
| 779 |
"step": 110
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 0.2853470437017995,
|
| 783 |
"grad_norm": 0.0,
|
| 784 |
"learning_rate": 1.6785712420155864e-05,
|
| 785 |
-
"loss": 1.
|
| 786 |
"step": 111
|
| 787 |
},
|
| 788 |
{
|
| 789 |
"epoch": 0.2879177377892031,
|
| 790 |
"grad_norm": 0.0,
|
| 791 |
"learning_rate": 1.67242677519123e-05,
|
| 792 |
-
"loss": 1.
|
| 793 |
"step": 112
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.29048843187660667,
|
| 797 |
"grad_norm": 0.0,
|
| 798 |
"learning_rate": 1.6662356145334158e-05,
|
| 799 |
-
"loss": 1.
|
| 800 |
"step": 113
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 0.2930591259640103,
|
| 804 |
"grad_norm": 0.0,
|
| 805 |
"learning_rate": 1.6599981899611103e-05,
|
| 806 |
-
"loss": 1.
|
| 807 |
"step": 114
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.29562982005141386,
|
| 811 |
"grad_norm": 0.0,
|
| 812 |
"learning_rate": 1.653714934605883e-05,
|
| 813 |
-
"loss": 1.
|
| 814 |
"step": 115
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 0.2982005141388175,
|
| 818 |
"grad_norm": 0.0,
|
| 819 |
"learning_rate": 1.647386284781828e-05,
|
| 820 |
-
"loss": 1.
|
| 821 |
"step": 116
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 0.30077120822622105,
|
| 825 |
"grad_norm": 0.0,
|
| 826 |
"learning_rate": 1.6410126799552653e-05,
|
| 827 |
-
"loss": 1.
|
| 828 |
"step": 117
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 0.3033419023136247,
|
| 832 |
"grad_norm": 0.0,
|
| 833 |
"learning_rate": 1.6345945627142264e-05,
|
| 834 |
-
"loss": 1.
|
| 835 |
"step": 118
|
| 836 |
},
|
| 837 |
{
|
| 838 |
"epoch": 0.3059125964010283,
|
| 839 |
"grad_norm": 0.0,
|
| 840 |
"learning_rate": 1.628132378737718e-05,
|
| 841 |
-
"loss": 1.
|
| 842 |
"step": 119
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 0.30848329048843187,
|
| 846 |
"grad_norm": 0.0,
|
| 847 |
"learning_rate": 1.6216265767647756e-05,
|
| 848 |
-
"loss": 1.
|
| 849 |
"step": 120
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.3110539845758355,
|
| 853 |
"grad_norm": 0.0,
|
| 854 |
"learning_rate": 1.615077608563302e-05,
|
| 855 |
-
"loss": 1.
|
| 856 |
"step": 121
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.31362467866323906,
|
| 860 |
"grad_norm": 0.0,
|
| 861 |
"learning_rate": 1.6084859288986957e-05,
|
| 862 |
-
"loss": 1.
|
| 863 |
"step": 122
|
| 864 |
},
|
| 865 |
{
|
| 866 |
"epoch": 0.3161953727506427,
|
| 867 |
"grad_norm": 0.0,
|
| 868 |
"learning_rate": 1.601851995502272e-05,
|
| 869 |
-
"loss": 1.
|
| 870 |
"step": 123
|
| 871 |
},
|
| 872 |
{
|
| 873 |
"epoch": 0.31876606683804626,
|
| 874 |
"grad_norm": 0.0,
|
| 875 |
"learning_rate": 1.5951762690394788e-05,
|
| 876 |
-
"loss": 1.
|
| 877 |
"step": 124
|
| 878 |
},
|
| 879 |
{
|
| 880 |
"epoch": 0.3213367609254499,
|
| 881 |
"grad_norm": 0.0,
|
| 882 |
"learning_rate": 1.5884592130779056e-05,
|
| 883 |
-
"loss": 1.
|
| 884 |
"step": 125
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 0.32390745501285345,
|
| 888 |
"grad_norm": 0.0,
|
| 889 |
"learning_rate": 1.581701294055095e-05,
|
| 890 |
-
"loss": 1.
|
| 891 |
"step": 126
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.3264781491002571,
|
| 895 |
"grad_norm": 0.0,
|
| 896 |
"learning_rate": 1.5749029812461515e-05,
|
| 897 |
-
"loss": 1.
|
| 898 |
"step": 127
|
| 899 |
},
|
| 900 |
{
|
| 901 |
"epoch": 0.32904884318766064,
|
| 902 |
"grad_norm": 0.0,
|
| 903 |
"learning_rate": 1.568064746731156e-05,
|
| 904 |
-
"loss": 1.
|
| 905 |
"step": 128
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.33161953727506427,
|
| 909 |
"grad_norm": 0.0,
|
| 910 |
"learning_rate": 1.5611870653623826e-05,
|
| 911 |
-
"loss": 1.
|
| 912 |
"step": 129
|
| 913 |
},
|
| 914 |
{
|
| 915 |
"epoch": 0.3341902313624679,
|
| 916 |
"grad_norm": 0.0,
|
| 917 |
"learning_rate": 1.5542704147313257e-05,
|
| 918 |
-
"loss": 1.
|
| 919 |
"step": 130
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 0.33676092544987146,
|
| 923 |
"grad_norm": 0.0,
|
| 924 |
"learning_rate": 1.5473152751355353e-05,
|
| 925 |
-
"loss": 1.
|
| 926 |
"step": 131
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 0.3393316195372751,
|
| 930 |
"grad_norm": 0.0,
|
| 931 |
"learning_rate": 1.5403221295452647e-05,
|
| 932 |
-
"loss": 1.
|
| 933 |
"step": 132
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.34190231362467866,
|
| 937 |
"grad_norm": 0.0,
|
| 938 |
"learning_rate": 1.5332914635699327e-05,
|
| 939 |
-
"loss": 1.
|
| 940 |
"step": 133
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 0.3444730077120823,
|
| 944 |
"grad_norm": 0.0,
|
| 945 |
"learning_rate": 1.5262237654244026e-05,
|
| 946 |
-
"loss": 1.
|
| 947 |
"step": 134
|
| 948 |
},
|
| 949 |
{
|
| 950 |
"epoch": 0.34704370179948585,
|
| 951 |
"grad_norm": 0.0,
|
| 952 |
"learning_rate": 1.5191195258950804e-05,
|
| 953 |
-
"loss": 1.
|
| 954 |
"step": 135
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.3496143958868895,
|
| 958 |
"grad_norm": 0.0,
|
| 959 |
"learning_rate": 1.5119792383058338e-05,
|
| 960 |
-
"loss": 1.
|
| 961 |
"step": 136
|
| 962 |
},
|
| 963 |
{
|
| 964 |
"epoch": 0.35218508997429304,
|
| 965 |
"grad_norm": 0.0,
|
| 966 |
"learning_rate": 1.5048033984837352e-05,
|
| 967 |
-
"loss": 1.
|
| 968 |
"step": 137
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 0.35475578406169667,
|
| 972 |
"grad_norm": 0.0,
|
| 973 |
"learning_rate": 1.4975925047246319e-05,
|
| 974 |
-
"loss": 1.
|
| 975 |
"step": 138
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.35732647814910024,
|
| 979 |
"grad_norm": 0.0,
|
| 980 |
"learning_rate": 1.4903470577585433e-05,
|
| 981 |
-
"loss": 1.
|
| 982 |
"step": 139
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.35989717223650386,
|
| 986 |
"grad_norm": 0.0,
|
| 987 |
"learning_rate": 1.4830675607148899e-05,
|
| 988 |
-
"loss": 1.
|
| 989 |
"step": 140
|
| 990 |
},
|
| 991 |
{
|
| 992 |
"epoch": 0.36246786632390743,
|
| 993 |
"grad_norm": 0.0,
|
| 994 |
"learning_rate": 1.475754519087557e-05,
|
| 995 |
-
"loss": 1.
|
| 996 |
"step": 141
|
| 997 |
},
|
| 998 |
{
|
| 999 |
"epoch": 0.36503856041131105,
|
| 1000 |
"grad_norm": 0.0,
|
| 1001 |
"learning_rate": 1.4684084406997903e-05,
|
| 1002 |
-
"loss": 1.
|
| 1003 |
"step": 142
|
| 1004 |
},
|
| 1005 |
{
|
| 1006 |
"epoch": 0.3676092544987147,
|
| 1007 |
"grad_norm": 0.0,
|
| 1008 |
"learning_rate": 1.4610298356689341e-05,
|
| 1009 |
-
"loss": 1.
|
| 1010 |
"step": 143
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 0.37017994858611825,
|
| 1014 |
"grad_norm": 0.0,
|
| 1015 |
"learning_rate": 1.453619216371008e-05,
|
| 1016 |
-
"loss": 1.
|
| 1017 |
"step": 144
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.37275064267352187,
|
| 1021 |
"grad_norm": 0.0,
|
| 1022 |
"learning_rate": 1.446177097405127e-05,
|
| 1023 |
-
"loss": 1.
|
| 1024 |
"step": 145
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
"epoch": 0.37532133676092544,
|
| 1028 |
"grad_norm": 0.0,
|
| 1029 |
"learning_rate": 1.4387039955577668e-05,
|
| 1030 |
-
"loss": 1.
|
| 1031 |
"step": 146
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
"epoch": 0.37789203084832906,
|
| 1035 |
"grad_norm": 0.0,
|
| 1036 |
"learning_rate": 1.4312004297668791e-05,
|
| 1037 |
-
"loss": 1.
|
| 1038 |
"step": 147
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
"epoch": 0.38046272493573263,
|
| 1042 |
"grad_norm": 0.0,
|
| 1043 |
"learning_rate": 1.4236669210858544e-05,
|
| 1044 |
-
"loss": 1.
|
| 1045 |
"step": 148
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
"epoch": 0.38303341902313626,
|
| 1049 |
"grad_norm": 0.0,
|
| 1050 |
"learning_rate": 1.4161039926473412e-05,
|
| 1051 |
-
"loss": 1.
|
| 1052 |
"step": 149
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 0.3856041131105398,
|
| 1056 |
"grad_norm": 0.0,
|
| 1057 |
"learning_rate": 1.4085121696269185e-05,
|
| 1058 |
-
"loss": 1.
|
| 1059 |
"step": 150
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.38817480719794345,
|
| 1063 |
"grad_norm": 0.0,
|
| 1064 |
"learning_rate": 1.4008919792066273e-05,
|
| 1065 |
-
"loss": 1.
|
| 1066 |
"step": 151
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 0.390745501285347,
|
| 1070 |
"grad_norm": 0.0,
|
| 1071 |
"learning_rate": 1.3932439505383628e-05,
|
| 1072 |
-
"loss": 1.
|
| 1073 |
"step": 152
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 0.39331619537275064,
|
| 1077 |
"grad_norm": 0.0,
|
| 1078 |
"learning_rate": 1.385568614707129e-05,
|
| 1079 |
-
"loss": 1.
|
| 1080 |
"step": 153
|
| 1081 |
},
|
| 1082 |
{
|
| 1083 |
"epoch": 0.39588688946015427,
|
| 1084 |
"grad_norm": 0.0,
|
| 1085 |
"learning_rate": 1.3778665046941616e-05,
|
| 1086 |
-
"loss": 1.
|
| 1087 |
"step": 154
|
| 1088 |
},
|
| 1089 |
{
|
| 1090 |
"epoch": 0.39845758354755784,
|
| 1091 |
"grad_norm": 0.0,
|
| 1092 |
"learning_rate": 1.3701381553399147e-05,
|
| 1093 |
-
"loss": 1.
|
| 1094 |
"step": 155
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 0.40102827763496146,
|
| 1098 |
"grad_norm": 0.0,
|
| 1099 |
"learning_rate": 1.3623841033069232e-05,
|
| 1100 |
-
"loss": 1.
|
| 1101 |
"step": 156
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.40359897172236503,
|
| 1105 |
"grad_norm": 0.0,
|
| 1106 |
"learning_rate": 1.3546048870425356e-05,
|
| 1107 |
-
"loss": 1.
|
| 1108 |
"step": 157
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.40616966580976865,
|
| 1112 |
"grad_norm": 0.0,
|
| 1113 |
"learning_rate": 1.3468010467415248e-05,
|
| 1114 |
-
"loss": 1.
|
| 1115 |
"step": 158
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
"epoch": 0.4087403598971722,
|
| 1119 |
"grad_norm": 0.0,
|
| 1120 |
"learning_rate": 1.3389731243085747e-05,
|
| 1121 |
-
"loss": 1.
|
| 1122 |
"step": 159
|
| 1123 |
},
|
| 1124 |
{
|
| 1125 |
"epoch": 0.41131105398457585,
|
| 1126 |
"grad_norm": 0.0,
|
| 1127 |
"learning_rate": 1.3311216633206514e-05,
|
| 1128 |
-
"loss": 1.
|
| 1129 |
"step": 160
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.4138817480719794,
|
| 1133 |
"grad_norm": 0.0,
|
| 1134 |
"learning_rate": 1.3232472089892567e-05,
|
| 1135 |
-
"loss": 1.
|
| 1136 |
"step": 161
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 0.41645244215938304,
|
| 1140 |
"grad_norm": 0.0,
|
| 1141 |
"learning_rate": 1.315350308122567e-05,
|
| 1142 |
-
"loss": 1.
|
| 1143 |
"step": 162
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.4190231362467866,
|
| 1147 |
"grad_norm": 0.0,
|
| 1148 |
"learning_rate": 1.3074315090874652e-05,
|
| 1149 |
-
"loss": 1.
|
| 1150 |
"step": 163
|
| 1151 |
},
|
| 1152 |
{
|
| 1153 |
"epoch": 0.42159383033419023,
|
| 1154 |
"grad_norm": 0.0,
|
| 1155 |
"learning_rate": 1.2994913617714573e-05,
|
| 1156 |
-
"loss": 1.
|
| 1157 |
"step": 164
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
"epoch": 0.4241645244215938,
|
| 1161 |
"grad_norm": 0.0,
|
| 1162 |
"learning_rate": 1.2915304175444929e-05,
|
| 1163 |
-
"loss": 1.
|
| 1164 |
"step": 165
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.4267352185089974,
|
| 1168 |
"grad_norm": 0.0,
|
| 1169 |
"learning_rate": 1.2835492292206735e-05,
|
| 1170 |
-
"loss": 1.
|
| 1171 |
"step": 166
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
"epoch": 0.42930591259640105,
|
| 1175 |
"grad_norm": 0.0,
|
| 1176 |
"learning_rate": 1.2755483510198668e-05,
|
| 1177 |
-
"loss": 1.
|
| 1178 |
"step": 167
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 0.4318766066838046,
|
| 1182 |
"grad_norm": 0.0,
|
| 1183 |
"learning_rate": 1.2675283385292212e-05,
|
| 1184 |
-
"loss": 1.
|
| 1185 |
"step": 168
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.43444730077120824,
|
| 1189 |
"grad_norm": 0.0,
|
| 1190 |
"learning_rate": 1.2594897486645836e-05,
|
| 1191 |
-
"loss": 1.
|
| 1192 |
"step": 169
|
| 1193 |
},
|
| 1194 |
{
|
| 1195 |
"epoch": 0.4370179948586118,
|
| 1196 |
"grad_norm": 0.0,
|
| 1197 |
"learning_rate": 1.2514331396318298e-05,
|
| 1198 |
-
"loss": 1.
|
| 1199 |
"step": 170
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
"epoch": 0.43958868894601544,
|
| 1203 |
"grad_norm": 0.0,
|
| 1204 |
"learning_rate": 1.2433590708880991e-05,
|
| 1205 |
-
"loss": 1.
|
| 1206 |
"step": 171
|
| 1207 |
},
|
| 1208 |
{
|
| 1209 |
"epoch": 0.442159383033419,
|
| 1210 |
"grad_norm": 0.0,
|
| 1211 |
"learning_rate": 1.2352681031029476e-05,
|
| 1212 |
-
"loss": 1.
|
| 1213 |
"step": 172
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
"epoch": 0.44473007712082263,
|
| 1217 |
"grad_norm": 0.0,
|
| 1218 |
"learning_rate": 1.2271607981194132e-05,
|
| 1219 |
-
"loss": 1.
|
| 1220 |
"step": 173
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 0.4473007712082262,
|
| 1224 |
"grad_norm": 0.0,
|
| 1225 |
"learning_rate": 1.2190377189150016e-05,
|
| 1226 |
-
"loss": 1.
|
| 1227 |
"step": 174
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.4498714652956298,
|
| 1231 |
"grad_norm": 0.0,
|
| 1232 |
"learning_rate": 1.2108994295625924e-05,
|
| 1233 |
-
"loss": 1.
|
| 1234 |
"step": 175
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.4524421593830334,
|
| 1238 |
"grad_norm": 0.0,
|
| 1239 |
"learning_rate": 1.2027464951912703e-05,
|
| 1240 |
-
"loss": 1.
|
| 1241 |
"step": 176
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.455012853470437,
|
| 1245 |
"grad_norm": 0.0,
|
| 1246 |
"learning_rate": 1.1945794819470805e-05,
|
| 1247 |
-
"loss": 1.
|
| 1248 |
"step": 177
|
| 1249 |
},
|
| 1250 |
{
|
| 1251 |
"epoch": 0.45758354755784064,
|
| 1252 |
"grad_norm": 0.0,
|
| 1253 |
"learning_rate": 1.1863989569537165e-05,
|
| 1254 |
-
"loss": 1.
|
| 1255 |
"step": 178
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 0.4601542416452442,
|
| 1259 |
"grad_norm": 0.0,
|
| 1260 |
"learning_rate": 1.1782054882731377e-05,
|
| 1261 |
-
"loss": 1.
|
| 1262 |
"step": 179
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 0.46272493573264784,
|
| 1266 |
"grad_norm": 0.0,
|
| 1267 |
"learning_rate": 1.1699996448661242e-05,
|
| 1268 |
-
"loss": 1.
|
| 1269 |
"step": 180
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.4652956298200514,
|
| 1273 |
"grad_norm": 0.0,
|
| 1274 |
"learning_rate": 1.161781996552765e-05,
|
| 1275 |
-
"loss": 1.
|
| 1276 |
"step": 181
|
| 1277 |
},
|
| 1278 |
{
|
| 1279 |
"epoch": 0.46786632390745503,
|
| 1280 |
"grad_norm": 0.0,
|
| 1281 |
"learning_rate": 1.1535531139728918e-05,
|
| 1282 |
-
"loss": 1.
|
| 1283 |
"step": 182
|
| 1284 |
},
|
| 1285 |
{
|
| 1286 |
"epoch": 0.4704370179948586,
|
| 1287 |
"grad_norm": 0.0,
|
| 1288 |
"learning_rate": 1.1453135685464524e-05,
|
| 1289 |
-
"loss": 1.
|
| 1290 |
"step": 183
|
| 1291 |
},
|
| 1292 |
{
|
| 1293 |
"epoch": 0.4730077120822622,
|
| 1294 |
"grad_norm": 0.0,
|
| 1295 |
"learning_rate": 1.1370639324338313e-05,
|
| 1296 |
-
"loss": 1.
|
| 1297 |
"step": 184
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 0.4755784061696658,
|
| 1301 |
"grad_norm": 0.0,
|
| 1302 |
"learning_rate": 1.1288047784961166e-05,
|
| 1303 |
-
"loss": 1.
|
| 1304 |
"step": 185
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 0.4781491002570694,
|
| 1308 |
"grad_norm": 0.0,
|
| 1309 |
"learning_rate": 1.1205366802553231e-05,
|
| 1310 |
-
"loss": 1.
|
| 1311 |
"step": 186
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.480719794344473,
|
| 1315 |
"grad_norm": 0.0,
|
| 1316 |
"learning_rate": 1.1122602118545642e-05,
|
| 1317 |
-
"loss": 1.
|
| 1318 |
"step": 187
|
| 1319 |
},
|
| 1320 |
{
|
| 1321 |
"epoch": 0.4832904884318766,
|
| 1322 |
"grad_norm": 0.0,
|
| 1323 |
"learning_rate": 1.1039759480181836e-05,
|
| 1324 |
-
"loss": 1.
|
| 1325 |
"step": 188
|
| 1326 |
},
|
| 1327 |
{
|
| 1328 |
"epoch": 0.48586118251928023,
|
| 1329 |
"grad_norm": 0.0,
|
| 1330 |
"learning_rate": 1.0956844640118462e-05,
|
| 1331 |
-
"loss": 1.
|
| 1332 |
"step": 189
|
| 1333 |
},
|
| 1334 |
{
|
| 1335 |
"epoch": 0.4884318766066838,
|
| 1336 |
"grad_norm": 0.0,
|
| 1337 |
"learning_rate": 1.0873863356025911e-05,
|
| 1338 |
-
"loss": 1.
|
| 1339 |
"step": 190
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
"epoch": 0.4910025706940874,
|
| 1343 |
"grad_norm": 0.0,
|
| 1344 |
"learning_rate": 1.0790821390188493e-05,
|
| 1345 |
-
"loss": 1.
|
| 1346 |
"step": 191
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 0.493573264781491,
|
| 1350 |
"grad_norm": 0.0,
|
| 1351 |
"learning_rate": 1.0707724509104318e-05,
|
| 1352 |
-
"loss": 1.
|
| 1353 |
"step": 192
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.4961439588688946,
|
| 1357 |
"grad_norm": 0.0,
|
| 1358 |
"learning_rate": 1.062457848308484e-05,
|
| 1359 |
-
"loss": 1.
|
| 1360 |
"step": 193
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.4987146529562982,
|
| 1364 |
"grad_norm": 0.0,
|
| 1365 |
"learning_rate": 1.0541389085854177e-05,
|
| 1366 |
-
"loss": 1.
|
| 1367 |
"step": 194
|
| 1368 |
},
|
| 1369 |
{
|
| 1370 |
"epoch": 0.5012853470437018,
|
| 1371 |
"grad_norm": 0.0,
|
| 1372 |
"learning_rate": 1.0458162094148185e-05,
|
| 1373 |
-
"loss": 1.
|
| 1374 |
"step": 195
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.5038560411311054,
|
| 1378 |
"grad_norm": 0.0,
|
| 1379 |
"learning_rate": 1.0374903287313307e-05,
|
| 1380 |
-
"loss": 1.
|
| 1381 |
"step": 196
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 0.506426735218509,
|
| 1385 |
"grad_norm": 0.0,
|
| 1386 |
"learning_rate": 1.029161844690525e-05,
|
| 1387 |
-
"loss": 1.
|
| 1388 |
"step": 197
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 0.5089974293059126,
|
| 1392 |
"grad_norm": 0.0,
|
| 1393 |
"learning_rate": 1.0208313356287505e-05,
|
| 1394 |
-
"loss": 1.
|
| 1395 |
"step": 198
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.5115681233933161,
|
| 1399 |
"grad_norm": 0.0,
|
| 1400 |
"learning_rate": 1.0124993800229774e-05,
|
| 1401 |
-
"loss": 1.
|
| 1402 |
"step": 199
|
| 1403 |
},
|
| 1404 |
{
|
| 1405 |
"epoch": 0.5141388174807198,
|
| 1406 |
"grad_norm": 0.0,
|
| 1407 |
"learning_rate": 1.004166556450623e-05,
|
| 1408 |
-
"loss": 1.
|
| 1409 |
"step": 200
|
| 1410 |
},
|
| 1411 |
{
|
| 1412 |
"epoch": 0.5167095115681234,
|
| 1413 |
"grad_norm": 0.0,
|
| 1414 |
"learning_rate": 9.958334435493776e-06,
|
| 1415 |
-
"loss": 1.
|
| 1416 |
"step": 201
|
| 1417 |
},
|
| 1418 |
{
|
| 1419 |
"epoch": 0.519280205655527,
|
| 1420 |
"grad_norm": 0.0,
|
| 1421 |
"learning_rate": 9.87500619977023e-06,
|
| 1422 |
-
"loss": 1.
|
| 1423 |
"step": 202
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 0.5218508997429306,
|
| 1427 |
"grad_norm": 0.0,
|
| 1428 |
"learning_rate": 9.791686643712498e-06,
|
| 1429 |
-
"loss": 1.
|
| 1430 |
"step": 203
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 0.5244215938303342,
|
| 1434 |
"grad_norm": 0.0,
|
| 1435 |
"learning_rate": 9.708381553094754e-06,
|
| 1436 |
-
"loss": 1.
|
| 1437 |
"step": 204
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.5269922879177378,
|
| 1441 |
"grad_norm": 0.0,
|
| 1442 |
"learning_rate": 9.625096712686694e-06,
|
| 1443 |
-
"loss": 1.
|
| 1444 |
"step": 205
|
| 1445 |
},
|
| 1446 |
{
|
| 1447 |
"epoch": 0.5295629820051414,
|
| 1448 |
"grad_norm": 0.0,
|
| 1449 |
"learning_rate": 9.541837905851817e-06,
|
| 1450 |
-
"loss": 1.
|
| 1451 |
"step": 206
|
| 1452 |
},
|
| 1453 |
{
|
| 1454 |
"epoch": 0.532133676092545,
|
| 1455 |
"grad_norm": 0.0,
|
| 1456 |
"learning_rate": 9.458610914145826e-06,
|
| 1457 |
-
"loss": 1.
|
| 1458 |
"step": 207
|
| 1459 |
},
|
| 1460 |
{
|
| 1461 |
"epoch": 0.5347043701799485,
|
| 1462 |
"grad_norm": 0.0,
|
| 1463 |
"learning_rate": 9.375421516915165e-06,
|
| 1464 |
-
"loss": 1.
|
| 1465 |
"step": 208
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.5372750642673522,
|
| 1469 |
"grad_norm": 0.0,
|
| 1470 |
"learning_rate": 9.292275490895685e-06,
|
| 1471 |
-
"loss": 1.
|
| 1472 |
"step": 209
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 0.5398457583547558,
|
| 1476 |
"grad_norm": 0.0,
|
| 1477 |
"learning_rate": 9.209178609811509e-06,
|
| 1478 |
-
"loss": 1.
|
| 1479 |
"step": 210
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.5424164524421594,
|
| 1483 |
"grad_norm": 0.0,
|
| 1484 |
"learning_rate": 9.126136643974094e-06,
|
| 1485 |
-
"loss": 1.
|
| 1486 |
"step": 211
|
| 1487 |
},
|
| 1488 |
{
|
| 1489 |
"epoch": 0.5449871465295629,
|
| 1490 |
"grad_norm": 0.0,
|
| 1491 |
"learning_rate": 9.043155359881538e-06,
|
| 1492 |
-
"loss": 1.
|
| 1493 |
"step": 212
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"epoch": 0.5475578406169666,
|
| 1497 |
"grad_norm": 0.0,
|
| 1498 |
"learning_rate": 8.960240519818167e-06,
|
| 1499 |
-
"loss": 1.
|
| 1500 |
"step": 213
|
| 1501 |
},
|
| 1502 |
{
|
| 1503 |
"epoch": 0.5501285347043702,
|
| 1504 |
"grad_norm": 0.0,
|
| 1505 |
"learning_rate": 8.877397881454358e-06,
|
| 1506 |
-
"loss": 1.
|
| 1507 |
"step": 214
|
| 1508 |
},
|
| 1509 |
{
|
| 1510 |
"epoch": 0.5526992287917738,
|
| 1511 |
"grad_norm": 0.0,
|
| 1512 |
"learning_rate": 8.79463319744677e-06,
|
| 1513 |
-
"loss": 1.
|
| 1514 |
"step": 215
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 0.5552699228791774,
|
| 1518 |
"grad_norm": 0.0,
|
| 1519 |
"learning_rate": 8.711952215038837e-06,
|
| 1520 |
-
"loss": 1.
|
| 1521 |
"step": 216
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
"epoch": 0.5578406169665809,
|
| 1525 |
"grad_norm": 0.0,
|
| 1526 |
"learning_rate": 8.629360675661693e-06,
|
| 1527 |
-
"loss": 1.
|
| 1528 |
"step": 217
|
| 1529 |
},
|
| 1530 |
{
|
| 1531 |
"epoch": 0.5604113110539846,
|
| 1532 |
"grad_norm": 0.0,
|
| 1533 |
"learning_rate": 8.546864314535478e-06,
|
| 1534 |
-
"loss": 1.
|
| 1535 |
"step": 218
|
| 1536 |
},
|
| 1537 |
{
|
| 1538 |
"epoch": 0.5629820051413882,
|
| 1539 |
"grad_norm": 0.0,
|
| 1540 |
"learning_rate": 8.464468860271084e-06,
|
| 1541 |
-
"loss": 1.
|
| 1542 |
"step": 219
|
| 1543 |
},
|
| 1544 |
{
|
| 1545 |
"epoch": 0.5655526992287918,
|
| 1546 |
"grad_norm": 0.0,
|
| 1547 |
"learning_rate": 8.382180034472353e-06,
|
| 1548 |
-
"loss": 1.
|
| 1549 |
"step": 220
|
| 1550 |
},
|
| 1551 |
{
|
| 1552 |
"epoch": 0.5681233933161953,
|
| 1553 |
"grad_norm": 0.0,
|
| 1554 |
"learning_rate": 8.30000355133876e-06,
|
| 1555 |
-
"loss": 1.
|
| 1556 |
"step": 221
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
"epoch": 0.570694087403599,
|
| 1560 |
"grad_norm": 0.0,
|
| 1561 |
"learning_rate": 8.217945117268624e-06,
|
| 1562 |
-
"loss": 1.
|
| 1563 |
"step": 222
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 0.5732647814910026,
|
| 1567 |
"grad_norm": 0.0,
|
| 1568 |
"learning_rate": 8.136010430462837e-06,
|
| 1569 |
-
"loss": 1.
|
| 1570 |
"step": 223
|
| 1571 |
},
|
| 1572 |
{
|
| 1573 |
"epoch": 0.5758354755784062,
|
| 1574 |
"grad_norm": 0.0,
|
| 1575 |
"learning_rate": 8.0542051805292e-06,
|
| 1576 |
-
"loss": 1.
|
| 1577 |
"step": 224
|
| 1578 |
},
|
| 1579 |
{
|
| 1580 |
"epoch": 0.5784061696658098,
|
| 1581 |
"grad_norm": 0.0,
|
| 1582 |
"learning_rate": 7.9725350480873e-06,
|
| 1583 |
-
"loss": 1.
|
| 1584 |
"step": 225
|
| 1585 |
},
|
| 1586 |
{
|
| 1587 |
"epoch": 0.5809768637532133,
|
| 1588 |
"grad_norm": 0.0,
|
| 1589 |
"learning_rate": 7.89100570437408e-06,
|
| 1590 |
-
"loss": 1.
|
| 1591 |
"step": 226
|
| 1592 |
},
|
| 1593 |
{
|
| 1594 |
"epoch": 0.583547557840617,
|
| 1595 |
"grad_norm": 0.0,
|
| 1596 |
"learning_rate": 7.809622810849986e-06,
|
| 1597 |
-
"loss": 1.
|
| 1598 |
"step": 227
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
"epoch": 0.5861182519280206,
|
| 1602 |
"grad_norm": 0.0,
|
| 1603 |
"learning_rate": 7.72839201880587e-06,
|
| 1604 |
-
"loss": 1.
|
| 1605 |
"step": 228
|
| 1606 |
},
|
| 1607 |
{
|
| 1608 |
"epoch": 0.5886889460154242,
|
| 1609 |
"grad_norm": 0.0,
|
| 1610 |
"learning_rate": 7.647318968970528e-06,
|
| 1611 |
-
"loss": 1.
|
| 1612 |
"step": 229
|
| 1613 |
},
|
| 1614 |
{
|
| 1615 |
"epoch": 0.5912596401028277,
|
| 1616 |
"grad_norm": 0.0,
|
| 1617 |
"learning_rate": 7.566409291119008e-06,
|
| 1618 |
-
"loss": 1.
|
| 1619 |
"step": 230
|
| 1620 |
},
|
| 1621 |
{
|
| 1622 |
"epoch": 0.5938303341902313,
|
| 1623 |
"grad_norm": 0.0,
|
| 1624 |
"learning_rate": 7.485668603681706e-06,
|
| 1625 |
-
"loss": 1.
|
| 1626 |
"step": 231
|
| 1627 |
},
|
| 1628 |
{
|
| 1629 |
"epoch": 0.596401028277635,
|
| 1630 |
"grad_norm": 0.0,
|
| 1631 |
"learning_rate": 7.405102513354166e-06,
|
| 1632 |
-
"loss": 1.
|
| 1633 |
"step": 232
|
| 1634 |
},
|
| 1635 |
{
|
| 1636 |
"epoch": 0.5989717223650386,
|
| 1637 |
"grad_norm": 0.0,
|
| 1638 |
"learning_rate": 7.324716614707794e-06,
|
| 1639 |
-
"loss": 1.
|
| 1640 |
"step": 233
|
| 1641 |
},
|
| 1642 |
{
|
| 1643 |
"epoch": 0.6015424164524421,
|
| 1644 |
"grad_norm": 0.0,
|
| 1645 |
"learning_rate": 7.2445164898013345e-06,
|
| 1646 |
-
"loss": 1.
|
| 1647 |
"step": 234
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
"epoch": 0.6041131105398457,
|
| 1651 |
"grad_norm": 0.0,
|
| 1652 |
"learning_rate": 7.1645077077932666e-06,
|
| 1653 |
-
"loss": 1.
|
| 1654 |
"step": 235
|
| 1655 |
},
|
| 1656 |
{
|
| 1657 |
"epoch": 0.6066838046272494,
|
| 1658 |
"grad_norm": 0.0,
|
| 1659 |
"learning_rate": 7.084695824555074e-06,
|
| 1660 |
-
"loss": 1.
|
| 1661 |
"step": 236
|
| 1662 |
},
|
| 1663 |
{
|
| 1664 |
"epoch": 0.609254498714653,
|
| 1665 |
"grad_norm": 0.0,
|
| 1666 |
"learning_rate": 7.005086382285426e-06,
|
| 1667 |
-
"loss": 1.
|
| 1668 |
"step": 237
|
| 1669 |
},
|
| 1670 |
{
|
| 1671 |
"epoch": 0.6118251928020566,
|
| 1672 |
"grad_norm": 0.0,
|
| 1673 |
"learning_rate": 6.925684909125354e-06,
|
| 1674 |
-
"loss": 1.
|
| 1675 |
"step": 238
|
| 1676 |
},
|
| 1677 |
{
|
| 1678 |
"epoch": 0.6143958868894601,
|
| 1679 |
"grad_norm": 0.0,
|
| 1680 |
"learning_rate": 6.84649691877433e-06,
|
| 1681 |
-
"loss": 1.
|
| 1682 |
"step": 239
|
| 1683 |
},
|
| 1684 |
{
|
| 1685 |
"epoch": 0.6169665809768637,
|
| 1686 |
"grad_norm": 0.0,
|
| 1687 |
"learning_rate": 6.767527910107437e-06,
|
| 1688 |
-
"loss": 1.
|
| 1689 |
"step": 240
|
| 1690 |
},
|
| 1691 |
{
|
| 1692 |
"epoch": 0.6195372750642674,
|
| 1693 |
"grad_norm": 0.0,
|
| 1694 |
"learning_rate": 6.688783366793488e-06,
|
| 1695 |
-
"loss": 1.
|
| 1696 |
"step": 241
|
| 1697 |
},
|
| 1698 |
{
|
| 1699 |
"epoch": 0.622107969151671,
|
| 1700 |
"grad_norm": 0.0,
|
| 1701 |
"learning_rate": 6.610268756914254e-06,
|
| 1702 |
-
"loss": 1.
|
| 1703 |
"step": 242
|
| 1704 |
},
|
| 1705 |
{
|
| 1706 |
"epoch": 0.6246786632390745,
|
| 1707 |
"grad_norm": 0.0,
|
| 1708 |
"learning_rate": 6.5319895325847535e-06,
|
| 1709 |
-
"loss": 1.
|
| 1710 |
"step": 243
|
| 1711 |
},
|
| 1712 |
{
|
| 1713 |
"epoch": 0.6272493573264781,
|
| 1714 |
"grad_norm": 0.0,
|
| 1715 |
"learning_rate": 6.453951129574644e-06,
|
| 1716 |
-
"loss": 1.
|
| 1717 |
"step": 244
|
| 1718 |
},
|
| 1719 |
{
|
| 1720 |
"epoch": 0.6298200514138818,
|
| 1721 |
"grad_norm": 0.0,
|
| 1722 |
"learning_rate": 6.3761589669307745e-06,
|
| 1723 |
-
"loss": 1.
|
| 1724 |
"step": 245
|
| 1725 |
},
|
| 1726 |
{
|
| 1727 |
"epoch": 0.6323907455012854,
|
| 1728 |
"grad_norm": 0.0,
|
| 1729 |
"learning_rate": 6.298618446600856e-06,
|
| 1730 |
-
"loss": 1.
|
| 1731 |
"step": 246
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
"epoch": 0.6349614395886889,
|
| 1735 |
"grad_norm": 0.0,
|
| 1736 |
"learning_rate": 6.221334953058389e-06,
|
| 1737 |
-
"loss": 1.
|
| 1738 |
"step": 247
|
| 1739 |
},
|
| 1740 |
{
|
| 1741 |
"epoch": 0.6375321336760925,
|
| 1742 |
"grad_norm": 0.0,
|
| 1743 |
"learning_rate": 6.144313852928712e-06,
|
| 1744 |
-
"loss": 1.
|
| 1745 |
"step": 248
|
| 1746 |
},
|
| 1747 |
{
|
| 1748 |
"epoch": 0.6401028277634961,
|
| 1749 |
"grad_norm": 0.0,
|
| 1750 |
"learning_rate": 6.067560494616374e-06,
|
| 1751 |
-
"loss": 1.
|
| 1752 |
"step": 249
|
| 1753 |
},
|
| 1754 |
{
|
| 1755 |
"epoch": 0.6426735218508998,
|
| 1756 |
"grad_norm": 0.0,
|
| 1757 |
"learning_rate": 5.9910802079337285e-06,
|
| 1758 |
-
"loss": 1.
|
| 1759 |
"step": 250
|
| 1760 |
},
|
| 1761 |
{
|
| 1762 |
"epoch": 0.6452442159383034,
|
| 1763 |
"grad_norm": 0.0,
|
| 1764 |
"learning_rate": 5.9148783037308154e-06,
|
| 1765 |
-
"loss": 1.
|
| 1766 |
"step": 251
|
| 1767 |
},
|
| 1768 |
{
|
| 1769 |
"epoch": 0.6478149100257069,
|
| 1770 |
"grad_norm": 0.0,
|
| 1771 |
"learning_rate": 5.838960073526589e-06,
|
| 1772 |
-
"loss": 1.
|
| 1773 |
"step": 252
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
"epoch": 0.6503856041131105,
|
| 1777 |
"grad_norm": 0.0,
|
| 1778 |
"learning_rate": 5.763330789141457e-06,
|
| 1779 |
-
"loss": 1.
|
| 1780 |
"step": 253
|
| 1781 |
},
|
| 1782 |
{
|
| 1783 |
"epoch": 0.6529562982005142,
|
| 1784 |
"grad_norm": 0.0,
|
| 1785 |
"learning_rate": 5.687995702331211e-06,
|
| 1786 |
-
"loss": 1.
|
| 1787 |
"step": 254
|
| 1788 |
},
|
| 1789 |
{
|
| 1790 |
"epoch": 0.6555269922879178,
|
| 1791 |
"grad_norm": 0.0,
|
| 1792 |
"learning_rate": 5.612960044422335e-06,
|
| 1793 |
-
"loss": 1.
|
| 1794 |
"step": 255
|
| 1795 |
},
|
| 1796 |
{
|
| 1797 |
"epoch": 0.6580976863753213,
|
| 1798 |
"grad_norm": 0.0,
|
| 1799 |
"learning_rate": 5.538229025948729e-06,
|
| 1800 |
-
"loss": 1.
|
| 1801 |
"step": 256
|
| 1802 |
},
|
| 1803 |
{
|
| 1804 |
"epoch": 0.6606683804627249,
|
| 1805 |
"grad_norm": 0.0,
|
| 1806 |
"learning_rate": 5.463807836289921e-06,
|
| 1807 |
-
"loss": 1.
|
| 1808 |
"step": 257
|
| 1809 |
},
|
| 1810 |
{
|
| 1811 |
"epoch": 0.6632390745501285,
|
| 1812 |
"grad_norm": 0.0,
|
| 1813 |
"learning_rate": 5.389701643310661e-06,
|
| 1814 |
-
"loss": 1.
|
| 1815 |
"step": 258
|
| 1816 |
},
|
| 1817 |
{
|
| 1818 |
"epoch": 0.6658097686375322,
|
| 1819 |
"grad_norm": 0.0,
|
| 1820 |
"learning_rate": 5.3159155930021e-06,
|
| 1821 |
-
"loss": 1.
|
| 1822 |
"step": 259
|
| 1823 |
},
|
| 1824 |
{
|
| 1825 |
"epoch": 0.6683804627249358,
|
| 1826 |
"grad_norm": 0.0,
|
| 1827 |
"learning_rate": 5.2424548091244334e-06,
|
| 1828 |
-
"loss": 1.
|
| 1829 |
"step": 260
|
| 1830 |
},
|
| 1831 |
{
|
| 1832 |
"epoch": 0.6709511568123393,
|
| 1833 |
"grad_norm": 0.0,
|
| 1834 |
"learning_rate": 5.169324392851105e-06,
|
| 1835 |
-
"loss": 1.
|
| 1836 |
"step": 261
|
| 1837 |
},
|
| 1838 |
{
|
| 1839 |
"epoch": 0.6735218508997429,
|
| 1840 |
"grad_norm": 0.0,
|
| 1841 |
"learning_rate": 5.096529422414571e-06,
|
| 1842 |
-
"loss": 1.
|
| 1843 |
"step": 262
|
| 1844 |
},
|
| 1845 |
{
|
| 1846 |
"epoch": 0.6760925449871465,
|
| 1847 |
"grad_norm": 0.0,
|
| 1848 |
"learning_rate": 5.0240749527536845e-06,
|
| 1849 |
-
"loss": 1.
|
| 1850 |
"step": 263
|
| 1851 |
},
|
| 1852 |
{
|
| 1853 |
"epoch": 0.6786632390745502,
|
| 1854 |
"grad_norm": 0.0,
|
| 1855 |
"learning_rate": 4.951966015162652e-06,
|
| 1856 |
-
"loss": 1.
|
| 1857 |
"step": 264
|
| 1858 |
},
|
| 1859 |
{
|
| 1860 |
"epoch": 0.6812339331619537,
|
| 1861 |
"grad_norm": 0.0,
|
| 1862 |
"learning_rate": 4.880207616941663e-06,
|
| 1863 |
-
"loss": 1.
|
| 1864 |
"step": 265
|
| 1865 |
},
|
| 1866 |
{
|
| 1867 |
"epoch": 0.6838046272493573,
|
| 1868 |
"grad_norm": 0.0,
|
| 1869 |
"learning_rate": 4.8088047410492e-06,
|
| 1870 |
-
"loss": 1.
|
| 1871 |
"step": 266
|
| 1872 |
},
|
| 1873 |
{
|
| 1874 |
"epoch": 0.6863753213367609,
|
| 1875 |
"grad_norm": 0.0,
|
| 1876 |
"learning_rate": 4.737762345755975e-06,
|
| 1877 |
-
"loss": 1.
|
| 1878 |
"step": 267
|
| 1879 |
},
|
| 1880 |
{
|
| 1881 |
"epoch": 0.6889460154241646,
|
| 1882 |
"grad_norm": 0.0,
|
| 1883 |
"learning_rate": 4.667085364300678e-06,
|
| 1884 |
-
"loss": 1.
|
| 1885 |
"step": 268
|
| 1886 |
},
|
| 1887 |
{
|
| 1888 |
"epoch": 0.6915167095115681,
|
| 1889 |
"grad_norm": 0.0,
|
| 1890 |
"learning_rate": 4.596778704547359e-06,
|
| 1891 |
-
"loss": 1.
|
| 1892 |
"step": 269
|
| 1893 |
},
|
| 1894 |
{
|
| 1895 |
"epoch": 0.6940874035989717,
|
| 1896 |
"grad_norm": 0.0,
|
| 1897 |
"learning_rate": 4.526847248644652e-06,
|
| 1898 |
-
"loss": 1.
|
| 1899 |
"step": 270
|
| 1900 |
},
|
| 1901 |
{
|
| 1902 |
"epoch": 0.6966580976863753,
|
| 1903 |
"grad_norm": 0.0,
|
| 1904 |
"learning_rate": 4.457295852686746e-06,
|
| 1905 |
-
"loss": 1.
|
| 1906 |
"step": 271
|
| 1907 |
},
|
| 1908 |
{
|
| 1909 |
"epoch": 0.699228791773779,
|
| 1910 |
"grad_norm": 0.0,
|
| 1911 |
"learning_rate": 4.388129346376177e-06,
|
| 1912 |
-
"loss": 1.
|
| 1913 |
"step": 272
|
| 1914 |
},
|
| 1915 |
{
|
| 1916 |
"epoch": 0.7017994858611826,
|
| 1917 |
"grad_norm": 0.0,
|
| 1918 |
"learning_rate": 4.319352532688444e-06,
|
| 1919 |
-
"loss": 1.
|
| 1920 |
"step": 273
|
| 1921 |
},
|
| 1922 |
{
|
| 1923 |
"epoch": 0.7043701799485861,
|
| 1924 |
"grad_norm": 0.0,
|
| 1925 |
"learning_rate": 4.250970187538484e-06,
|
| 1926 |
-
"loss": 1.
|
| 1927 |
"step": 274
|
| 1928 |
},
|
| 1929 |
{
|
| 1930 |
"epoch": 0.7069408740359897,
|
| 1931 |
"grad_norm": 0.0,
|
| 1932 |
"learning_rate": 4.182987059449056e-06,
|
| 1933 |
-
"loss": 1.
|
| 1934 |
"step": 275
|
| 1935 |
},
|
| 1936 |
{
|
| 1937 |
"epoch": 0.7095115681233933,
|
| 1938 |
"grad_norm": 0.0,
|
| 1939 |
"learning_rate": 4.115407869220948e-06,
|
| 1940 |
-
"loss": 1.
|
| 1941 |
"step": 276
|
| 1942 |
},
|
| 1943 |
{
|
| 1944 |
"epoch": 0.712082262210797,
|
| 1945 |
"grad_norm": 0.0,
|
| 1946 |
"learning_rate": 4.048237309605216e-06,
|
| 1947 |
-
"loss": 1.
|
| 1948 |
"step": 277
|
| 1949 |
},
|
| 1950 |
{
|
| 1951 |
"epoch": 0.7146529562982005,
|
| 1952 |
"grad_norm": 0.0,
|
| 1953 |
"learning_rate": 3.981480044977284e-06,
|
| 1954 |
-
"loss": 1.
|
| 1955 |
"step": 278
|
| 1956 |
},
|
| 1957 |
{
|
| 1958 |
"epoch": 0.7172236503856041,
|
| 1959 |
"grad_norm": 0.0,
|
| 1960 |
"learning_rate": 3.915140711013044e-06,
|
| 1961 |
-
"loss": 1.
|
| 1962 |
"step": 279
|
| 1963 |
},
|
| 1964 |
{
|
| 1965 |
"epoch": 0.7197943444730077,
|
| 1966 |
"grad_norm": 0.0,
|
| 1967 |
"learning_rate": 3.849223914366981e-06,
|
| 1968 |
-
"loss": 1.
|
| 1969 |
"step": 280
|
| 1970 |
},
|
| 1971 |
{
|
| 1972 |
"epoch": 0.7223650385604113,
|
| 1973 |
"grad_norm": 0.0,
|
| 1974 |
"learning_rate": 3.7837342323522454e-06,
|
| 1975 |
-
"loss": 1.
|
| 1976 |
"step": 281
|
| 1977 |
},
|
| 1978 |
{
|
| 1979 |
"epoch": 0.7249357326478149,
|
| 1980 |
"grad_norm": 0.0,
|
| 1981 |
"learning_rate": 3.7186762126228227e-06,
|
| 1982 |
-
"loss": 1.
|
| 1983 |
"step": 282
|
| 1984 |
},
|
| 1985 |
{
|
| 1986 |
"epoch": 0.7275064267352185,
|
| 1987 |
"grad_norm": 0.0,
|
| 1988 |
"learning_rate": 3.654054372857738e-06,
|
| 1989 |
-
"loss": 1.
|
| 1990 |
"step": 283
|
| 1991 |
},
|
| 1992 |
{
|
| 1993 |
"epoch": 0.7300771208226221,
|
| 1994 |
"grad_norm": 0.0,
|
| 1995 |
"learning_rate": 3.5898732004473523e-06,
|
| 1996 |
-
"loss": 1.
|
| 1997 |
"step": 284
|
| 1998 |
},
|
| 1999 |
{
|
| 2000 |
"epoch": 0.7326478149100257,
|
| 2001 |
"grad_norm": 0.0,
|
| 2002 |
"learning_rate": 3.5261371521817247e-06,
|
| 2003 |
-
"loss": 1.
|
| 2004 |
"step": 285
|
| 2005 |
},
|
| 2006 |
{
|
| 2007 |
"epoch": 0.7352185089974294,
|
| 2008 |
"grad_norm": 0.0,
|
| 2009 |
"learning_rate": 3.462850653941171e-06,
|
| 2010 |
-
"loss": 1.
|
| 2011 |
"step": 286
|
| 2012 |
},
|
| 2013 |
{
|
| 2014 |
"epoch": 0.7377892030848329,
|
| 2015 |
"grad_norm": 0.0,
|
| 2016 |
"learning_rate": 3.4000181003889e-06,
|
| 2017 |
-
"loss": 1.
|
| 2018 |
"step": 287
|
| 2019 |
},
|
| 2020 |
{
|
| 2021 |
"epoch": 0.7403598971722365,
|
| 2022 |
"grad_norm": 0.0,
|
| 2023 |
"learning_rate": 3.337643854665843e-06,
|
| 2024 |
-
"loss": 1.
|
| 2025 |
"step": 288
|
| 2026 |
},
|
| 2027 |
{
|
| 2028 |
"epoch": 0.7429305912596401,
|
| 2029 |
"grad_norm": 0.0,
|
| 2030 |
"learning_rate": 3.2757322480876996e-06,
|
| 2031 |
-
"loss": 1.
|
| 2032 |
"step": 289
|
| 2033 |
},
|
| 2034 |
{
|
| 2035 |
"epoch": 0.7455012853470437,
|
| 2036 |
"grad_norm": 0.0,
|
| 2037 |
"learning_rate": 3.2142875798441376e-06,
|
| 2038 |
-
"loss": 1.
|
| 2039 |
"step": 290
|
| 2040 |
},
|
| 2041 |
{
|
| 2042 |
"epoch": 0.7480719794344473,
|
| 2043 |
"grad_norm": 0.0,
|
| 2044 |
"learning_rate": 3.15331411670027e-06,
|
| 2045 |
-
"loss": 1.
|
| 2046 |
"step": 291
|
| 2047 |
},
|
| 2048 |
{
|
| 2049 |
"epoch": 0.7506426735218509,
|
| 2050 |
"grad_norm": 0.0,
|
| 2051 |
"learning_rate": 3.092816092700366e-06,
|
| 2052 |
-
"loss": 1.
|
| 2053 |
"step": 292
|
| 2054 |
},
|
| 2055 |
{
|
| 2056 |
"epoch": 0.7532133676092545,
|
| 2057 |
"grad_norm": 0.0,
|
| 2058 |
"learning_rate": 3.032797708873828e-06,
|
| 2059 |
-
"loss": 1.
|
| 2060 |
"step": 293
|
| 2061 |
},
|
| 2062 |
{
|
| 2063 |
"epoch": 0.7557840616966581,
|
| 2064 |
"grad_norm": 0.0,
|
| 2065 |
"learning_rate": 2.97326313294349e-06,
|
| 2066 |
-
"loss": 1.
|
| 2067 |
"step": 294
|
| 2068 |
},
|
| 2069 |
{
|
| 2070 |
"epoch": 0.7583547557840618,
|
| 2071 |
"grad_norm": 0.0,
|
| 2072 |
"learning_rate": 2.914216499036178e-06,
|
| 2073 |
-
"loss": 1.
|
| 2074 |
"step": 295
|
| 2075 |
},
|
| 2076 |
{
|
| 2077 |
"epoch": 0.7609254498714653,
|
| 2078 |
"grad_norm": 0.0,
|
| 2079 |
"learning_rate": 2.855661907395655e-06,
|
| 2080 |
-
"loss": 1.
|
| 2081 |
"step": 296
|
| 2082 |
},
|
| 2083 |
{
|
| 2084 |
"epoch": 0.7634961439588689,
|
| 2085 |
"grad_norm": 0.0,
|
| 2086 |
"learning_rate": 2.7976034240978834e-06,
|
| 2087 |
-
"loss": 1.
|
| 2088 |
"step": 297
|
| 2089 |
},
|
| 2090 |
{
|
| 2091 |
"epoch": 0.7660668380462725,
|
| 2092 |
"grad_norm": 0.0,
|
| 2093 |
"learning_rate": 2.740045080768694e-06,
|
| 2094 |
-
"loss": 1.
|
| 2095 |
"step": 298
|
| 2096 |
},
|
| 2097 |
{
|
| 2098 |
"epoch": 0.7686375321336761,
|
| 2099 |
"grad_norm": 0.0,
|
| 2100 |
"learning_rate": 2.6829908743037936e-06,
|
| 2101 |
-
"loss": 1.
|
| 2102 |
"step": 299
|
| 2103 |
},
|
| 2104 |
{
|
| 2105 |
"epoch": 0.7712082262210797,
|
| 2106 |
"grad_norm": 0.0,
|
| 2107 |
"learning_rate": 2.626444766591253e-06,
|
| 2108 |
-
"loss": 1.
|
| 2109 |
"step": 300
|
| 2110 |
},
|
| 2111 |
{
|
| 2112 |
"epoch": 0.7737789203084833,
|
| 2113 |
"grad_norm": 0.0,
|
| 2114 |
"learning_rate": 2.570410684236365e-06,
|
| 2115 |
-
"loss": 1.
|
| 2116 |
"step": 301
|
| 2117 |
},
|
| 2118 |
{
|
| 2119 |
"epoch": 0.7763496143958869,
|
| 2120 |
"grad_norm": 0.0,
|
| 2121 |
"learning_rate": 2.514892518288988e-06,
|
| 2122 |
-
"loss": 1.
|
| 2123 |
"step": 302
|
| 2124 |
},
|
| 2125 |
{
|
| 2126 |
"epoch": 0.7789203084832905,
|
| 2127 |
"grad_norm": 0.0,
|
| 2128 |
"learning_rate": 2.4598941239733555e-06,
|
| 2129 |
-
"loss": 1.
|
| 2130 |
"step": 303
|
| 2131 |
},
|
| 2132 |
{
|
| 2133 |
"epoch": 0.781491002570694,
|
| 2134 |
"grad_norm": 0.0,
|
| 2135 |
"learning_rate": 2.4054193204203457e-06,
|
| 2136 |
-
"loss": 1.
|
| 2137 |
"step": 304
|
| 2138 |
},
|
| 2139 |
{
|
| 2140 |
"epoch": 0.7840616966580977,
|
| 2141 |
"grad_norm": 0.0,
|
| 2142 |
"learning_rate": 2.3514718904022993e-06,
|
| 2143 |
-
"loss": 1.
|
| 2144 |
"step": 305
|
| 2145 |
},
|
| 2146 |
{
|
| 2147 |
"epoch": 0.7866323907455013,
|
| 2148 |
"grad_norm": 0.0,
|
| 2149 |
"learning_rate": 2.2980555800703273e-06,
|
| 2150 |
-
"loss": 1.
|
| 2151 |
"step": 306
|
| 2152 |
},
|
| 2153 |
{
|
| 2154 |
"epoch": 0.7892030848329049,
|
| 2155 |
"grad_norm": 0.0,
|
| 2156 |
"learning_rate": 2.2451740986941905e-06,
|
| 2157 |
-
"loss": 1.
|
| 2158 |
"step": 307
|
| 2159 |
},
|
| 2160 |
{
|
| 2161 |
"epoch": 0.7917737789203085,
|
| 2162 |
"grad_norm": 0.0,
|
| 2163 |
"learning_rate": 2.1928311184046967e-06,
|
| 2164 |
-
"loss": 1.
|
| 2165 |
"step": 308
|
| 2166 |
},
|
| 2167 |
{
|
| 2168 |
"epoch": 0.794344473007712,
|
| 2169 |
"grad_norm": 0.0,
|
| 2170 |
"learning_rate": 2.1410302739387424e-06,
|
| 2171 |
-
"loss": 1.
|
| 2172 |
"step": 309
|
| 2173 |
},
|
| 2174 |
{
|
| 2175 |
"epoch": 0.7969151670951157,
|
| 2176 |
"grad_norm": 0.0,
|
| 2177 |
"learning_rate": 2.0897751623868833e-06,
|
| 2178 |
-
"loss": 1.
|
| 2179 |
"step": 310
|
| 2180 |
},
|
| 2181 |
{
|
| 2182 |
"epoch": 0.7994858611825193,
|
| 2183 |
"grad_norm": 0.0,
|
| 2184 |
"learning_rate": 2.0390693429435626e-06,
|
| 2185 |
-
"loss": 1.
|
| 2186 |
"step": 311
|
| 2187 |
},
|
| 2188 |
{
|
| 2189 |
"epoch": 0.8020565552699229,
|
| 2190 |
"grad_norm": 0.0,
|
| 2191 |
"learning_rate": 1.9889163366599607e-06,
|
| 2192 |
-
"loss": 1.
|
| 2193 |
"step": 312
|
| 2194 |
},
|
| 2195 |
{
|
| 2196 |
"epoch": 0.8046272493573264,
|
| 2197 |
"grad_norm": 0.0,
|
| 2198 |
"learning_rate": 1.939319626199483e-06,
|
| 2199 |
-
"loss": 1.
|
| 2200 |
"step": 313
|
| 2201 |
},
|
| 2202 |
{
|
| 2203 |
"epoch": 0.8071979434447301,
|
| 2204 |
"grad_norm": 0.0,
|
| 2205 |
"learning_rate": 1.890282655595922e-06,
|
| 2206 |
-
"loss": 1.
|
| 2207 |
"step": 314
|
| 2208 |
},
|
| 2209 |
{
|
| 2210 |
"epoch": 0.8097686375321337,
|
| 2211 |
"grad_norm": 0.0,
|
| 2212 |
"learning_rate": 1.8418088300143044e-06,
|
| 2213 |
-
"loss": 1.
|
| 2214 |
"step": 315
|
| 2215 |
},
|
| 2216 |
{
|
| 2217 |
"epoch": 0.8123393316195373,
|
| 2218 |
"grad_norm": 0.0,
|
| 2219 |
"learning_rate": 1.7939015155144378e-06,
|
| 2220 |
-
"loss": 1.
|
| 2221 |
"step": 316
|
| 2222 |
},
|
| 2223 |
{
|
| 2224 |
"epoch": 0.8149100257069408,
|
| 2225 |
"grad_norm": 0.0,
|
| 2226 |
"learning_rate": 1.7465640388171589e-06,
|
| 2227 |
-
"loss": 1.
|
| 2228 |
"step": 317
|
| 2229 |
},
|
| 2230 |
{
|
| 2231 |
"epoch": 0.8174807197943444,
|
| 2232 |
"grad_norm": 0.0,
|
| 2233 |
"learning_rate": 1.6997996870733268e-06,
|
| 2234 |
-
"loss": 1.
|
| 2235 |
"step": 318
|
| 2236 |
},
|
| 2237 |
{
|
| 2238 |
"epoch": 0.8200514138817481,
|
| 2239 |
"grad_norm": 0.0,
|
| 2240 |
"learning_rate": 1.6536117076355652e-06,
|
| 2241 |
-
"loss": 1.
|
| 2242 |
"step": 319
|
| 2243 |
},
|
| 2244 |
{
|
| 2245 |
"epoch": 0.8226221079691517,
|
| 2246 |
"grad_norm": 0.0,
|
| 2247 |
"learning_rate": 1.6080033078327585e-06,
|
| 2248 |
-
"loss": 1.
|
| 2249 |
"step": 320
|
| 2250 |
},
|
| 2251 |
{
|
| 2252 |
"epoch": 0.8251928020565553,
|
| 2253 |
"grad_norm": 0.0,
|
| 2254 |
"learning_rate": 1.5629776547473397e-06,
|
| 2255 |
-
"loss": 1.
|
| 2256 |
"step": 321
|
| 2257 |
},
|
| 2258 |
{
|
| 2259 |
"epoch": 0.8277634961439588,
|
| 2260 |
"grad_norm": 0.0,
|
| 2261 |
"learning_rate": 1.5185378749953538e-06,
|
| 2262 |
-
"loss": 1.
|
| 2263 |
"step": 322
|
| 2264 |
},
|
| 2265 |
{
|
| 2266 |
"epoch": 0.8303341902313625,
|
| 2267 |
"grad_norm": 0.0,
|
| 2268 |
"learning_rate": 1.4746870545093528e-06,
|
| 2269 |
-
"loss": 1.
|
| 2270 |
"step": 323
|
| 2271 |
},
|
| 2272 |
{
|
| 2273 |
"epoch": 0.8329048843187661,
|
| 2274 |
"grad_norm": 0.0,
|
| 2275 |
"learning_rate": 1.4314282383241097e-06,
|
| 2276 |
-
"loss": 1.
|
| 2277 |
"step": 324
|
| 2278 |
},
|
| 2279 |
{
|
| 2280 |
"epoch": 0.8354755784061697,
|
| 2281 |
"grad_norm": 0.0,
|
| 2282 |
"learning_rate": 1.388764430365147e-06,
|
| 2283 |
-
"loss": 1.
|
| 2284 |
"step": 325
|
| 2285 |
},
|
| 2286 |
{
|
| 2287 |
"epoch": 0.8380462724935732,
|
| 2288 |
"grad_norm": 0.0,
|
| 2289 |
"learning_rate": 1.3466985932401743e-06,
|
| 2290 |
-
"loss": 1.
|
| 2291 |
"step": 326
|
| 2292 |
},
|
| 2293 |
{
|
| 2294 |
"epoch": 0.8406169665809768,
|
| 2295 |
"grad_norm": 0.0,
|
| 2296 |
"learning_rate": 1.3052336480333372e-06,
|
| 2297 |
-
"loss": 1.
|
| 2298 |
"step": 327
|
| 2299 |
},
|
| 2300 |
{
|
| 2301 |
"epoch": 0.8431876606683805,
|
| 2302 |
"grad_norm": 0.0,
|
| 2303 |
"learning_rate": 1.2643724741023845e-06,
|
| 2304 |
-
"loss": 1.
|
| 2305 |
"step": 328
|
| 2306 |
},
|
| 2307 |
{
|
| 2308 |
"epoch": 0.8457583547557841,
|
| 2309 |
"grad_norm": 0.0,
|
| 2310 |
"learning_rate": 1.2241179088787192e-06,
|
| 2311 |
-
"loss": 1.
|
| 2312 |
"step": 329
|
| 2313 |
},
|
| 2314 |
{
|
| 2315 |
"epoch": 0.8483290488431876,
|
| 2316 |
"grad_norm": 0.0,
|
| 2317 |
"learning_rate": 1.1844727476703776e-06,
|
| 2318 |
-
"loss": 1.
|
| 2319 |
"step": 330
|
| 2320 |
},
|
| 2321 |
{
|
| 2322 |
"epoch": 0.8508997429305912,
|
| 2323 |
"grad_norm": 0.0,
|
| 2324 |
"learning_rate": 1.1454397434679022e-06,
|
| 2325 |
-
"loss": 1.
|
| 2326 |
"step": 331
|
| 2327 |
},
|
| 2328 |
{
|
| 2329 |
"epoch": 0.8534704370179949,
|
| 2330 |
"grad_norm": 0.0,
|
| 2331 |
"learning_rate": 1.1070216067531825e-06,
|
| 2332 |
-
"loss": 1.
|
| 2333 |
"step": 332
|
| 2334 |
},
|
| 2335 |
{
|
| 2336 |
"epoch": 0.8560411311053985,
|
| 2337 |
"grad_norm": 0.0,
|
| 2338 |
"learning_rate": 1.0692210053112451e-06,
|
| 2339 |
-
"loss": 1.
|
| 2340 |
"step": 333
|
| 2341 |
},
|
| 2342 |
{
|
| 2343 |
"epoch": 0.8586118251928021,
|
| 2344 |
"grad_norm": 0.0,
|
| 2345 |
"learning_rate": 1.032040564044975e-06,
|
| 2346 |
-
"loss": 1.
|
| 2347 |
"step": 334
|
| 2348 |
},
|
| 2349 |
{
|
| 2350 |
"epoch": 0.8611825192802056,
|
| 2351 |
"grad_norm": 0.0,
|
| 2352 |
"learning_rate": 9.954828647928727e-07,
|
| 2353 |
-
"loss": 1.
|
| 2354 |
"step": 335
|
| 2355 |
},
|
| 2356 |
{
|
| 2357 |
"epoch": 0.8637532133676092,
|
| 2358 |
"grad_norm": 0.0,
|
| 2359 |
"learning_rate": 9.595504461497441e-07,
|
| 2360 |
-
"loss": 1.
|
| 2361 |
"step": 336
|
| 2362 |
},
|
| 2363 |
{
|
| 2364 |
"epoch": 0.8663239074550129,
|
| 2365 |
"grad_norm": 0.0,
|
| 2366 |
"learning_rate": 9.242458032904311e-07,
|
| 2367 |
-
"loss": 1.
|
| 2368 |
"step": 337
|
| 2369 |
},
|
| 2370 |
{
|
| 2371 |
"epoch": 0.8688946015424165,
|
| 2372 |
"grad_norm": 0.0,
|
| 2373 |
"learning_rate": 8.895713877965373e-07,
|
| 2374 |
-
"loss": 1.
|
| 2375 |
"step": 338
|
| 2376 |
},
|
| 2377 |
{
|
| 2378 |
"epoch": 0.87146529562982,
|
| 2379 |
"grad_norm": 0.0,
|
| 2380 |
"learning_rate": 8.555296074861996e-07,
|
| 2381 |
-
"loss": 1.
|
| 2382 |
"step": 339
|
| 2383 |
},
|
| 2384 |
{
|
| 2385 |
"epoch": 0.8740359897172236,
|
| 2386 |
"grad_norm": 0.0,
|
| 2387 |
"learning_rate": 8.22122826246875e-07,
|
| 2388 |
-
"loss": 1.
|
| 2389 |
"step": 340
|
| 2390 |
},
|
| 2391 |
{
|
| 2392 |
"epoch": 0.8766066838046273,
|
| 2393 |
"grad_norm": 0.0,
|
| 2394 |
"learning_rate": 7.89353363871197e-07,
|
| 2395 |
-
"loss": 1.
|
| 2396 |
"step": 341
|
| 2397 |
},
|
| 2398 |
{
|
| 2399 |
"epoch": 0.8791773778920309,
|
| 2400 |
"grad_norm": 0.0,
|
| 2401 |
"learning_rate": 7.572234958958846e-07,
|
| 2402 |
-
"loss": 1.
|
| 2403 |
"step": 342
|
| 2404 |
},
|
| 2405 |
{
|
| 2406 |
"epoch": 0.8817480719794345,
|
| 2407 |
"grad_norm": 0.0,
|
| 2408 |
"learning_rate": 7.2573545344373e-07,
|
| 2409 |
-
"loss": 1.
|
| 2410 |
"step": 343
|
| 2411 |
},
|
| 2412 |
{
|
| 2413 |
"epoch": 0.884318766066838,
|
| 2414 |
"grad_norm": 0.0,
|
| 2415 |
"learning_rate": 6.948914230686688e-07,
|
| 2416 |
-
"loss": 1.
|
| 2417 |
"step": 344
|
| 2418 |
},
|
| 2419 |
{
|
| 2420 |
"epoch": 0.8868894601542416,
|
| 2421 |
"grad_norm": 0.0,
|
| 2422 |
"learning_rate": 6.646935466039373e-07,
|
| 2423 |
-
"loss": 1.
|
| 2424 |
"step": 345
|
| 2425 |
},
|
| 2426 |
{
|
| 2427 |
"epoch": 0.8894601542416453,
|
| 2428 |
"grad_norm": 0.0,
|
| 2429 |
"learning_rate": 6.351439210133492e-07,
|
| 2430 |
-
"loss": 1.
|
| 2431 |
"step": 346
|
| 2432 |
},
|
| 2433 |
{
|
| 2434 |
"epoch": 0.8920308483290489,
|
| 2435 |
"grad_norm": 0.0,
|
| 2436 |
"learning_rate": 6.062445982456777e-07,
|
| 2437 |
-
"loss": 1.
|
| 2438 |
"step": 347
|
| 2439 |
},
|
| 2440 |
{
|
| 2441 |
"epoch": 0.8946015424164524,
|
| 2442 |
"grad_norm": 0.0,
|
| 2443 |
"learning_rate": 5.77997585092166e-07,
|
| 2444 |
-
"loss": 1.
|
| 2445 |
"step": 348
|
| 2446 |
},
|
| 2447 |
{
|
| 2448 |
"epoch": 0.897172236503856,
|
| 2449 |
"grad_norm": 0.0,
|
| 2450 |
"learning_rate": 5.504048430471753e-07,
|
| 2451 |
-
"loss": 1.
|
| 2452 |
"step": 349
|
| 2453 |
},
|
| 2454 |
{
|
| 2455 |
"epoch": 0.8997429305912596,
|
| 2456 |
"grad_norm": 0.0,
|
| 2457 |
"learning_rate": 5.234682881719766e-07,
|
| 2458 |
-
"loss": 1.
|
| 2459 |
"step": 350
|
| 2460 |
},
|
| 2461 |
{
|
| 2462 |
"epoch": 0.9023136246786633,
|
| 2463 |
"grad_norm": 0.0,
|
| 2464 |
"learning_rate": 4.971897909616985e-07,
|
| 2465 |
-
"loss": 1.
|
| 2466 |
"step": 351
|
| 2467 |
},
|
| 2468 |
{
|
| 2469 |
"epoch": 0.9048843187660668,
|
| 2470 |
"grad_norm": 0.0,
|
| 2471 |
"learning_rate": 4.715711762154362e-07,
|
| 2472 |
-
"loss": 1.
|
| 2473 |
"step": 352
|
| 2474 |
},
|
| 2475 |
{
|
| 2476 |
"epoch": 0.9074550128534704,
|
| 2477 |
"grad_norm": 0.0,
|
| 2478 |
"learning_rate": 4.4661422290954495e-07,
|
| 2479 |
-
"loss": 1.
|
| 2480 |
"step": 353
|
| 2481 |
},
|
| 2482 |
{
|
| 2483 |
"epoch": 0.910025706940874,
|
| 2484 |
"grad_norm": 0.0,
|
| 2485 |
"learning_rate": 4.2232066407409067e-07,
|
| 2486 |
-
"loss": 1.
|
| 2487 |
"step": 354
|
| 2488 |
},
|
| 2489 |
{
|
| 2490 |
"epoch": 0.9125964010282777,
|
| 2491 |
"grad_norm": 0.0,
|
| 2492 |
"learning_rate": 3.986921866725202e-07,
|
| 2493 |
-
"loss": 1.
|
| 2494 |
"step": 355
|
| 2495 |
},
|
| 2496 |
{
|
| 2497 |
"epoch": 0.9151670951156813,
|
| 2498 |
"grad_norm": 0.0,
|
| 2499 |
"learning_rate": 3.7573043148451673e-07,
|
| 2500 |
-
"loss": 1.
|
| 2501 |
"step": 356
|
| 2502 |
},
|
| 2503 |
{
|
| 2504 |
"epoch": 0.9177377892030848,
|
| 2505 |
"grad_norm": 0.0,
|
| 2506 |
"learning_rate": 3.5343699299205003e-07,
|
| 2507 |
-
"loss": 1.
|
| 2508 |
"step": 357
|
| 2509 |
},
|
| 2510 |
{
|
| 2511 |
"epoch": 0.9203084832904884,
|
| 2512 |
"grad_norm": 0.0,
|
| 2513 |
"learning_rate": 3.3181341926867283e-07,
|
| 2514 |
-
"loss": 1.
|
| 2515 |
"step": 358
|
| 2516 |
},
|
| 2517 |
{
|
| 2518 |
"epoch": 0.922879177377892,
|
| 2519 |
"grad_norm": 0.0,
|
| 2520 |
"learning_rate": 3.1086121187200667e-07,
|
| 2521 |
-
"loss": 1.
|
| 2522 |
"step": 359
|
| 2523 |
},
|
| 2524 |
{
|
| 2525 |
"epoch": 0.9254498714652957,
|
| 2526 |
"grad_norm": 0.0,
|
| 2527 |
"learning_rate": 2.905818257394799e-07,
|
| 2528 |
-
"loss": 1.
|
| 2529 |
"step": 360
|
| 2530 |
},
|
| 2531 |
{
|
| 2532 |
"epoch": 0.9280205655526992,
|
| 2533 |
"grad_norm": 0.0,
|
| 2534 |
"learning_rate": 2.7097666908729283e-07,
|
| 2535 |
-
"loss": 1.
|
| 2536 |
"step": 361
|
| 2537 |
},
|
| 2538 |
{
|
| 2539 |
"epoch": 0.9305912596401028,
|
| 2540 |
"grad_norm": 0.0,
|
| 2541 |
"learning_rate": 2.520471033126326e-07,
|
| 2542 |
-
"loss": 1.
|
| 2543 |
"step": 362
|
| 2544 |
},
|
| 2545 |
{
|
| 2546 |
"epoch": 0.9331619537275064,
|
| 2547 |
"grad_norm": 0.0,
|
| 2548 |
"learning_rate": 2.3379444289913344e-07,
|
| 2549 |
-
"loss": 1.
|
| 2550 |
"step": 363
|
| 2551 |
},
|
| 2552 |
{
|
| 2553 |
"epoch": 0.9357326478149101,
|
| 2554 |
"grad_norm": 0.0,
|
| 2555 |
"learning_rate": 2.1621995532559947e-07,
|
| 2556 |
-
"loss": 1.
|
| 2557 |
"step": 364
|
| 2558 |
},
|
| 2559 |
{
|
| 2560 |
"epoch": 0.9383033419023136,
|
| 2561 |
"grad_norm": 0.0,
|
| 2562 |
"learning_rate": 1.9932486097799408e-07,
|
| 2563 |
-
"loss": 1.
|
| 2564 |
"step": 365
|
| 2565 |
},
|
| 2566 |
{
|
| 2567 |
"epoch": 0.9408740359897172,
|
| 2568 |
"grad_norm": 0.0,
|
| 2569 |
"learning_rate": 1.8311033306468552e-07,
|
| 2570 |
-
"loss": 1.
|
| 2571 |
"step": 366
|
| 2572 |
},
|
| 2573 |
{
|
| 2574 |
"epoch": 0.9434447300771208,
|
| 2575 |
"grad_norm": 0.0,
|
| 2576 |
"learning_rate": 1.6757749753498865e-07,
|
| 2577 |
-
"loss": 1.
|
| 2578 |
"step": 367
|
| 2579 |
},
|
| 2580 |
{
|
| 2581 |
"epoch": 0.9460154241645244,
|
| 2582 |
"grad_norm": 0.0,
|
| 2583 |
"learning_rate": 1.5272743300097316e-07,
|
| 2584 |
-
"loss": 1.
|
| 2585 |
"step": 368
|
| 2586 |
},
|
| 2587 |
{
|
| 2588 |
"epoch": 0.9485861182519281,
|
| 2589 |
"grad_norm": 0.0,
|
| 2590 |
"learning_rate": 1.3856117066256225e-07,
|
| 2591 |
-
"loss": 1.
|
| 2592 |
"step": 369
|
| 2593 |
},
|
| 2594 |
{
|
| 2595 |
"epoch": 0.9511568123393316,
|
| 2596 |
"grad_norm": 0.0,
|
| 2597 |
"learning_rate": 1.2507969423593225e-07,
|
| 2598 |
-
"loss": 1.
|
| 2599 |
"step": 370
|
| 2600 |
},
|
| 2601 |
{
|
| 2602 |
"epoch": 0.9537275064267352,
|
| 2603 |
"grad_norm": 0.0,
|
| 2604 |
"learning_rate": 1.1228393988519381e-07,
|
| 2605 |
-
"loss": 1.
|
| 2606 |
"step": 371
|
| 2607 |
},
|
| 2608 |
{
|
| 2609 |
"epoch": 0.9562982005141388,
|
| 2610 |
"grad_norm": 0.0,
|
| 2611 |
"learning_rate": 1.0017479615738957e-07,
|
| 2612 |
-
"loss": 1.
|
| 2613 |
"step": 372
|
| 2614 |
},
|
| 2615 |
{
|
| 2616 |
"epoch": 0.9588688946015425,
|
| 2617 |
"grad_norm": 0.0,
|
| 2618 |
"learning_rate": 8.875310392079118e-08,
|
| 2619 |
-
"loss": 1.
|
| 2620 |
"step": 373
|
| 2621 |
},
|
| 2622 |
{
|
| 2623 |
"epoch": 0.961439588688946,
|
| 2624 |
"grad_norm": 0.0,
|
| 2625 |
"learning_rate": 7.801965630651165e-08,
|
| 2626 |
-
"loss": 1.
|
| 2627 |
"step": 374
|
| 2628 |
},
|
| 2629 |
{
|
| 2630 |
"epoch": 0.9640102827763496,
|
| 2631 |
"grad_norm": 0.0,
|
| 2632 |
"learning_rate": 6.797519865342161e-08,
|
| 2633 |
-
"loss": 1.
|
| 2634 |
"step": 375
|
| 2635 |
},
|
| 2636 |
{
|
| 2637 |
"epoch": 0.9665809768637532,
|
| 2638 |
"grad_norm": 0.0,
|
| 2639 |
"learning_rate": 5.862042845640403e-08,
|
| 2640 |
-
"loss": 1.
|
| 2641 |
"step": 376
|
| 2642 |
},
|
| 2643 |
{
|
| 2644 |
"epoch": 0.9691516709511568,
|
| 2645 |
"grad_norm": 0.0,
|
| 2646 |
"learning_rate": 4.9955995317908514e-08,
|
| 2647 |
-
"loss": 1.
|
| 2648 |
"step": 377
|
| 2649 |
},
|
| 2650 |
{
|
| 2651 |
"epoch": 0.9717223650385605,
|
| 2652 |
"grad_norm": 0.0,
|
| 2653 |
"learning_rate": 4.198250090284961e-08,
|
| 2654 |
-
"loss": 1.
|
| 2655 |
"step": 378
|
| 2656 |
},
|
| 2657 |
{
|
| 2658 |
"epoch": 0.974293059125964,
|
| 2659 |
"grad_norm": 0.0,
|
| 2660 |
"learning_rate": 3.47004988968247e-08,
|
| 2661 |
-
"loss": 1.
|
| 2662 |
"step": 379
|
| 2663 |
},
|
| 2664 |
{
|
| 2665 |
"epoch": 0.9768637532133676,
|
| 2666 |
"grad_norm": 0.0,
|
| 2667 |
"learning_rate": 2.8110494967664713e-08,
|
| 2668 |
-
"loss": 1.
|
| 2669 |
"step": 380
|
| 2670 |
},
|
| 2671 |
{
|
| 2672 |
"epoch": 0.9794344473007712,
|
| 2673 |
"grad_norm": 0.0,
|
| 2674 |
"learning_rate": 2.221294673032004e-08,
|
| 2675 |
-
"loss": 1.
|
| 2676 |
"step": 381
|
| 2677 |
},
|
| 2678 |
{
|
| 2679 |
"epoch": 0.9820051413881749,
|
| 2680 |
"grad_norm": 0.0,
|
| 2681 |
"learning_rate": 1.7008263715085904e-08,
|
| 2682 |
-
"loss": 1.
|
| 2683 |
"step": 382
|
| 2684 |
},
|
| 2685 |
{
|
| 2686 |
"epoch": 0.9845758354755784,
|
| 2687 |
"grad_norm": 0.0,
|
| 2688 |
"learning_rate": 1.24968073391607e-08,
|
| 2689 |
-
"loss": 1.
|
| 2690 |
"step": 383
|
| 2691 |
},
|
| 2692 |
{
|
| 2693 |
"epoch": 0.987146529562982,
|
| 2694 |
"grad_norm": 0.0,
|
| 2695 |
"learning_rate": 8.678890881552715e-09,
|
| 2696 |
-
"loss": 1.
|
| 2697 |
"step": 384
|
| 2698 |
},
|
| 2699 |
{
|
| 2700 |
"epoch": 0.9897172236503856,
|
| 2701 |
"grad_norm": 0.0,
|
| 2702 |
"learning_rate": 5.554779461323101e-09,
|
| 2703 |
-
"loss": 1.
|
| 2704 |
"step": 385
|
| 2705 |
},
|
| 2706 |
{
|
| 2707 |
"epoch": 0.9922879177377892,
|
| 2708 |
"grad_norm": 0.0,
|
| 2709 |
"learning_rate": 3.1246900191761463e-09,
|
| 2710 |
-
"loss": 1.
|
| 2711 |
"step": 386
|
| 2712 |
},
|
| 2713 |
{
|
| 2714 |
"epoch": 0.9948586118251928,
|
| 2715 |
"grad_norm": 0.0,
|
| 2716 |
"learning_rate": 1.3887913023946652e-09,
|
| 2717 |
-
"loss": 1.
|
| 2718 |
"step": 387
|
| 2719 |
},
|
| 2720 |
{
|
| 2721 |
"epoch": 0.9974293059125964,
|
| 2722 |
"grad_norm": 0.0,
|
| 2723 |
"learning_rate": 3.4720385312492223e-10,
|
| 2724 |
-
"loss": 1.
|
| 2725 |
"step": 388
|
| 2726 |
},
|
| 2727 |
{
|
| 2728 |
"epoch": 1.0,
|
| 2729 |
"grad_norm": 0.0,
|
| 2730 |
"learning_rate": 0.0,
|
| 2731 |
-
"loss": 1.
|
| 2732 |
"step": 389
|
| 2733 |
},
|
| 2734 |
{
|
| 2735 |
"epoch": 1.0,
|
| 2736 |
"step": 389,
|
| 2737 |
-
"total_flos": 1.
|
| 2738 |
-
"train_loss": 1.
|
| 2739 |
-
"train_runtime":
|
| 2740 |
-
"train_samples_per_second": 17.
|
| 2741 |
"train_steps_per_second": 0.067
|
| 2742 |
}
|
| 2743 |
],
|
|
@@ -2758,7 +2758,7 @@
|
|
| 2758 |
"attributes": {}
|
| 2759 |
}
|
| 2760 |
},
|
| 2761 |
-
"total_flos": 1.
|
| 2762 |
"train_batch_size": 8,
|
| 2763 |
"trial_name": null,
|
| 2764 |
"trial_params": null
|
|
|
|
| 12 |
"epoch": 0.002570694087403599,
|
| 13 |
"grad_norm": 0.0,
|
| 14 |
"learning_rate": 1.6666666666666667e-06,
|
| 15 |
+
"loss": 1.7749,
|
| 16 |
"step": 1
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.005141388174807198,
|
| 20 |
"grad_norm": 0.0,
|
| 21 |
"learning_rate": 3.3333333333333333e-06,
|
| 22 |
+
"loss": 1.7549,
|
| 23 |
"step": 2
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.007712082262210797,
|
| 27 |
"grad_norm": 0.0,
|
| 28 |
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 1.7659,
|
| 30 |
"step": 3
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.010282776349614395,
|
| 34 |
"grad_norm": 0.0,
|
| 35 |
"learning_rate": 6.666666666666667e-06,
|
| 36 |
+
"loss": 1.7842,
|
| 37 |
"step": 4
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.012853470437017995,
|
| 41 |
"grad_norm": 0.0,
|
| 42 |
"learning_rate": 8.333333333333334e-06,
|
| 43 |
+
"loss": 1.8015,
|
| 44 |
"step": 5
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.015424164524421594,
|
| 48 |
"grad_norm": 0.0,
|
| 49 |
"learning_rate": 1e-05,
|
| 50 |
+
"loss": 1.7534,
|
| 51 |
"step": 6
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.017994858611825194,
|
| 55 |
"grad_norm": 0.0,
|
| 56 |
"learning_rate": 1.1666666666666668e-05,
|
| 57 |
+
"loss": 1.7747,
|
| 58 |
"step": 7
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.02056555269922879,
|
| 62 |
"grad_norm": 0.0,
|
| 63 |
"learning_rate": 1.3333333333333333e-05,
|
| 64 |
+
"loss": 1.7783,
|
| 65 |
"step": 8
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.02313624678663239,
|
| 69 |
"grad_norm": 0.0,
|
| 70 |
"learning_rate": 1.5000000000000002e-05,
|
| 71 |
+
"loss": 1.803,
|
| 72 |
"step": 9
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.02570694087403599,
|
| 76 |
"grad_norm": 0.0,
|
| 77 |
"learning_rate": 1.6666666666666667e-05,
|
| 78 |
+
"loss": 1.7859,
|
| 79 |
"step": 10
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.028277634961439587,
|
| 83 |
"grad_norm": 0.0,
|
| 84 |
"learning_rate": 1.8333333333333333e-05,
|
| 85 |
+
"loss": 1.7856,
|
| 86 |
"step": 11
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.030848329048843187,
|
| 90 |
"grad_norm": 0.0,
|
| 91 |
"learning_rate": 2e-05,
|
| 92 |
+
"loss": 1.7825,
|
| 93 |
"step": 12
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.033419023136246784,
|
| 97 |
"grad_norm": 0.0,
|
| 98 |
"learning_rate": 1.9999652796146877e-05,
|
| 99 |
+
"loss": 1.8127,
|
| 100 |
"step": 13
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.03598971722365039,
|
| 104 |
"grad_norm": 0.0,
|
| 105 |
"learning_rate": 1.9998611208697607e-05,
|
| 106 |
+
"loss": 1.8115,
|
| 107 |
"step": 14
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.038560411311053984,
|
| 111 |
"grad_norm": 0.0,
|
| 112 |
"learning_rate": 1.9996875309980824e-05,
|
| 113 |
+
"loss": 1.7544,
|
| 114 |
"step": 15
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.04113110539845758,
|
| 118 |
"grad_norm": 0.0,
|
| 119 |
"learning_rate": 1.9994445220538678e-05,
|
| 120 |
+
"loss": 1.7886,
|
| 121 |
"step": 16
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.043701799485861184,
|
| 125 |
"grad_norm": 0.0,
|
| 126 |
"learning_rate": 1.999132110911845e-05,
|
| 127 |
+
"loss": 1.7866,
|
| 128 |
"step": 17
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.04627249357326478,
|
| 132 |
"grad_norm": 0.0,
|
| 133 |
"learning_rate": 1.9987503192660842e-05,
|
| 134 |
+
"loss": 1.7612,
|
| 135 |
"step": 18
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.04884318766066838,
|
| 139 |
"grad_norm": 0.0,
|
| 140 |
"learning_rate": 1.9982991736284914e-05,
|
| 141 |
+
"loss": 1.7944,
|
| 142 |
"step": 19
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.05141388174807198,
|
| 146 |
"grad_norm": 0.0,
|
| 147 |
"learning_rate": 1.997778705326968e-05,
|
| 148 |
+
"loss": 1.7656,
|
| 149 |
"step": 20
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.05398457583547558,
|
| 153 |
"grad_norm": 0.0,
|
| 154 |
"learning_rate": 1.9971889505032337e-05,
|
| 155 |
+
"loss": 1.7554,
|
| 156 |
"step": 21
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.056555269922879174,
|
| 160 |
"grad_norm": 0.0,
|
| 161 |
"learning_rate": 1.9965299501103178e-05,
|
| 162 |
+
"loss": 1.7637,
|
| 163 |
"step": 22
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.05912596401028278,
|
| 167 |
"grad_norm": 0.0,
|
| 168 |
"learning_rate": 1.995801749909715e-05,
|
| 169 |
+
"loss": 1.7803,
|
| 170 |
"step": 23
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.061696658097686374,
|
| 174 |
"grad_norm": 0.0,
|
| 175 |
"learning_rate": 1.995004400468209e-05,
|
| 176 |
+
"loss": 1.7402,
|
| 177 |
"step": 24
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 0.06426735218508997,
|
| 181 |
"grad_norm": 0.0,
|
| 182 |
"learning_rate": 1.9941379571543597e-05,
|
| 183 |
+
"loss": 1.7017,
|
| 184 |
"step": 25
|
| 185 |
},
|
| 186 |
{
|
| 187 |
"epoch": 0.06683804627249357,
|
| 188 |
"grad_norm": 0.0,
|
| 189 |
"learning_rate": 1.9932024801346583e-05,
|
| 190 |
+
"loss": 1.7671,
|
| 191 |
"step": 26
|
| 192 |
},
|
| 193 |
{
|
| 194 |
"epoch": 0.06940874035989718,
|
| 195 |
"grad_norm": 0.0,
|
| 196 |
"learning_rate": 1.992198034369349e-05,
|
| 197 |
+
"loss": 1.7014,
|
| 198 |
"step": 27
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"epoch": 0.07197943444730077,
|
| 202 |
"grad_norm": 0.0,
|
| 203 |
"learning_rate": 1.991124689607921e-05,
|
| 204 |
+
"loss": 1.7532,
|
| 205 |
"step": 28
|
| 206 |
},
|
| 207 |
{
|
| 208 |
"epoch": 0.07455012853470437,
|
| 209 |
"grad_norm": 0.0,
|
| 210 |
"learning_rate": 1.9899825203842613e-05,
|
| 211 |
+
"loss": 1.7129,
|
| 212 |
"step": 29
|
| 213 |
},
|
| 214 |
{
|
| 215 |
"epoch": 0.07712082262210797,
|
| 216 |
"grad_norm": 0.0,
|
| 217 |
"learning_rate": 1.988771606011481e-05,
|
| 218 |
+
"loss": 1.7126,
|
| 219 |
"step": 30
|
| 220 |
},
|
| 221 |
{
|
| 222 |
"epoch": 0.07969151670951156,
|
| 223 |
"grad_norm": 0.0,
|
| 224 |
"learning_rate": 1.987492030576407e-05,
|
| 225 |
+
"loss": 1.7393,
|
| 226 |
"step": 31
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.08226221079691516,
|
| 230 |
"grad_norm": 0.0,
|
| 231 |
"learning_rate": 1.986143882933744e-05,
|
| 232 |
+
"loss": 1.7742,
|
| 233 |
"step": 32
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.08483290488431877,
|
| 237 |
"grad_norm": 0.0,
|
| 238 |
"learning_rate": 1.9847272566999026e-05,
|
| 239 |
+
"loss": 1.7483,
|
| 240 |
"step": 33
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 0.08740359897172237,
|
| 244 |
"grad_norm": 0.0,
|
| 245 |
"learning_rate": 1.9832422502465013e-05,
|
| 246 |
+
"loss": 1.707,
|
| 247 |
"step": 34
|
| 248 |
},
|
| 249 |
{
|
| 250 |
"epoch": 0.08997429305912596,
|
| 251 |
"grad_norm": 0.0,
|
| 252 |
"learning_rate": 1.9816889666935318e-05,
|
| 253 |
+
"loss": 1.7507,
|
| 254 |
"step": 35
|
| 255 |
},
|
| 256 |
{
|
| 257 |
"epoch": 0.09254498714652956,
|
| 258 |
"grad_norm": 0.0,
|
| 259 |
"learning_rate": 1.9800675139022006e-05,
|
| 260 |
+
"loss": 1.7339,
|
| 261 |
"step": 36
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"epoch": 0.09511568123393316,
|
| 265 |
"grad_norm": 0.0,
|
| 266 |
"learning_rate": 1.9783780044674402e-05,
|
| 267 |
+
"loss": 1.748,
|
| 268 |
"step": 37
|
| 269 |
},
|
| 270 |
{
|
| 271 |
"epoch": 0.09768637532133675,
|
| 272 |
"grad_norm": 0.0,
|
| 273 |
"learning_rate": 1.976620555710087e-05,
|
| 274 |
+
"loss": 1.686,
|
| 275 |
"step": 38
|
| 276 |
},
|
| 277 |
{
|
| 278 |
"epoch": 0.10025706940874037,
|
| 279 |
"grad_norm": 0.0,
|
| 280 |
"learning_rate": 1.974795289668737e-05,
|
| 281 |
+
"loss": 1.7043,
|
| 282 |
"step": 39
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 0.10282776349614396,
|
| 286 |
"grad_norm": 0.0,
|
| 287 |
"learning_rate": 1.972902333091271e-05,
|
| 288 |
+
"loss": 1.7646,
|
| 289 |
"step": 40
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 0.10539845758354756,
|
| 293 |
"grad_norm": 0.0,
|
| 294 |
"learning_rate": 1.9709418174260523e-05,
|
| 295 |
+
"loss": 1.6802,
|
| 296 |
"step": 41
|
| 297 |
},
|
| 298 |
{
|
| 299 |
"epoch": 0.10796915167095116,
|
| 300 |
"grad_norm": 0.0,
|
| 301 |
"learning_rate": 1.9689138788127994e-05,
|
| 302 |
+
"loss": 1.6775,
|
| 303 |
"step": 42
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 0.11053984575835475,
|
| 307 |
"grad_norm": 0.0,
|
| 308 |
"learning_rate": 1.966818658073133e-05,
|
| 309 |
+
"loss": 1.6633,
|
| 310 |
"step": 43
|
| 311 |
},
|
| 312 |
{
|
| 313 |
"epoch": 0.11311053984575835,
|
| 314 |
"grad_norm": 0.0,
|
| 315 |
"learning_rate": 1.9646563007007952e-05,
|
| 316 |
+
"loss": 1.7637,
|
| 317 |
"step": 44
|
| 318 |
},
|
| 319 |
{
|
| 320 |
"epoch": 0.11568123393316196,
|
| 321 |
"grad_norm": 0.0,
|
| 322 |
"learning_rate": 1.9624269568515486e-05,
|
| 323 |
+
"loss": 1.7087,
|
| 324 |
"step": 45
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"epoch": 0.11825192802056556,
|
| 328 |
"grad_norm": 0.0,
|
| 329 |
"learning_rate": 1.960130781332748e-05,
|
| 330 |
+
"loss": 1.6562,
|
| 331 |
"step": 46
|
| 332 |
},
|
| 333 |
{
|
| 334 |
"epoch": 0.12082262210796915,
|
| 335 |
"grad_norm": 0.0,
|
| 336 |
"learning_rate": 1.957767933592591e-05,
|
| 337 |
+
"loss": 1.698,
|
| 338 |
"step": 47
|
| 339 |
},
|
| 340 |
{
|
| 341 |
"epoch": 0.12339331619537275,
|
| 342 |
"grad_norm": 0.0,
|
| 343 |
"learning_rate": 1.955338577709046e-05,
|
| 344 |
+
"loss": 1.7444,
|
| 345 |
"step": 48
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.12596401028277635,
|
| 349 |
"grad_norm": 0.0,
|
| 350 |
"learning_rate": 1.9528428823784567e-05,
|
| 351 |
+
"loss": 1.6743,
|
| 352 |
"step": 49
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.12853470437017994,
|
| 356 |
"grad_norm": 0.0,
|
| 357 |
"learning_rate": 1.9502810209038302e-05,
|
| 358 |
+
"loss": 1.6741,
|
| 359 |
"step": 50
|
| 360 |
},
|
| 361 |
{
|
| 362 |
"epoch": 0.13110539845758354,
|
| 363 |
"grad_norm": 0.0,
|
| 364 |
"learning_rate": 1.9476531711828027e-05,
|
| 365 |
+
"loss": 1.708,
|
| 366 |
"step": 51
|
| 367 |
},
|
| 368 |
{
|
| 369 |
"epoch": 0.13367609254498714,
|
| 370 |
"grad_norm": 0.0,
|
| 371 |
"learning_rate": 1.9449595156952827e-05,
|
| 372 |
+
"loss": 1.6587,
|
| 373 |
"step": 52
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 0.13624678663239073,
|
| 377 |
"grad_norm": 0.0,
|
| 378 |
"learning_rate": 1.9422002414907837e-05,
|
| 379 |
+
"loss": 1.6887,
|
| 380 |
"step": 53
|
| 381 |
},
|
| 382 |
{
|
| 383 |
"epoch": 0.13881748071979436,
|
| 384 |
"grad_norm": 0.0,
|
| 385 |
"learning_rate": 1.9393755401754324e-05,
|
| 386 |
+
"loss": 1.6714,
|
| 387 |
"step": 54
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"epoch": 0.14138817480719795,
|
| 391 |
"grad_norm": 0.0,
|
| 392 |
"learning_rate": 1.936485607898665e-05,
|
| 393 |
+
"loss": 1.7432,
|
| 394 |
"step": 55
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 0.14395886889460155,
|
| 398 |
"grad_norm": 0.0,
|
| 399 |
"learning_rate": 1.9335306453396066e-05,
|
| 400 |
+
"loss": 1.6675,
|
| 401 |
"step": 56
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 0.14652956298200515,
|
| 405 |
"grad_norm": 0.0,
|
| 406 |
"learning_rate": 1.9305108576931336e-05,
|
| 407 |
+
"loss": 1.6436,
|
| 408 |
"step": 57
|
| 409 |
},
|
| 410 |
{
|
| 411 |
"epoch": 0.14910025706940874,
|
| 412 |
"grad_norm": 0.0,
|
| 413 |
"learning_rate": 1.927426454655627e-05,
|
| 414 |
+
"loss": 1.6853,
|
| 415 |
"step": 58
|
| 416 |
},
|
| 417 |
{
|
| 418 |
"epoch": 0.15167095115681234,
|
| 419 |
"grad_norm": 0.0,
|
| 420 |
"learning_rate": 1.924277650410412e-05,
|
| 421 |
+
"loss": 1.6641,
|
| 422 |
"step": 59
|
| 423 |
},
|
| 424 |
{
|
| 425 |
"epoch": 0.15424164524421594,
|
| 426 |
"grad_norm": 0.0,
|
| 427 |
"learning_rate": 1.9210646636128805e-05,
|
| 428 |
+
"loss": 1.7385,
|
| 429 |
"step": 60
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 0.15681233933161953,
|
| 433 |
"grad_norm": 0.0,
|
| 434 |
"learning_rate": 1.9177877173753127e-05,
|
| 435 |
+
"loss": 1.7178,
|
| 436 |
"step": 61
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.15938303341902313,
|
| 440 |
"grad_norm": 0.0,
|
| 441 |
"learning_rate": 1.91444703925138e-05,
|
| 442 |
+
"loss": 1.6785,
|
| 443 |
"step": 62
|
| 444 |
},
|
| 445 |
{
|
| 446 |
"epoch": 0.16195372750642673,
|
| 447 |
"grad_norm": 0.0,
|
| 448 |
"learning_rate": 1.9110428612203463e-05,
|
| 449 |
+
"loss": 1.6799,
|
| 450 |
"step": 63
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"epoch": 0.16452442159383032,
|
| 454 |
"grad_norm": 0.0,
|
| 455 |
"learning_rate": 1.9075754196709574e-05,
|
| 456 |
+
"loss": 1.7075,
|
| 457 |
"step": 64
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.16709511568123395,
|
| 461 |
"grad_norm": 0.0,
|
| 462 |
"learning_rate": 1.904044955385026e-05,
|
| 463 |
+
"loss": 1.6621,
|
| 464 |
"step": 65
|
| 465 |
},
|
| 466 |
{
|
| 467 |
"epoch": 0.16966580976863754,
|
| 468 |
"grad_norm": 0.0,
|
| 469 |
"learning_rate": 1.9004517135207127e-05,
|
| 470 |
+
"loss": 1.6492,
|
| 471 |
"step": 66
|
| 472 |
},
|
| 473 |
{
|
| 474 |
"epoch": 0.17223650385604114,
|
| 475 |
"grad_norm": 0.0,
|
| 476 |
"learning_rate": 1.8967959435955027e-05,
|
| 477 |
+
"loss": 1.7297,
|
| 478 |
"step": 67
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.17480719794344474,
|
| 482 |
"grad_norm": 0.0,
|
| 483 |
"learning_rate": 1.893077899468876e-05,
|
| 484 |
+
"loss": 1.6882,
|
| 485 |
"step": 68
|
| 486 |
},
|
| 487 |
{
|
| 488 |
"epoch": 0.17737789203084833,
|
| 489 |
"grad_norm": 0.0,
|
| 490 |
"learning_rate": 1.889297839324682e-05,
|
| 491 |
+
"loss": 1.6714,
|
| 492 |
"step": 69
|
| 493 |
},
|
| 494 |
{
|
| 495 |
"epoch": 0.17994858611825193,
|
| 496 |
"grad_norm": 0.0,
|
| 497 |
"learning_rate": 1.8854560256532098e-05,
|
| 498 |
+
"loss": 1.6489,
|
| 499 |
"step": 70
|
| 500 |
},
|
| 501 |
{
|
| 502 |
"epoch": 0.18251928020565553,
|
| 503 |
"grad_norm": 0.0,
|
| 504 |
"learning_rate": 1.8815527252329624e-05,
|
| 505 |
+
"loss": 1.6721,
|
| 506 |
"step": 71
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 0.18508997429305912,
|
| 510 |
"grad_norm": 0.0,
|
| 511 |
"learning_rate": 1.8775882091121282e-05,
|
| 512 |
+
"loss": 1.6533,
|
| 513 |
"step": 72
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 0.18766066838046272,
|
| 517 |
"grad_norm": 0.0,
|
| 518 |
"learning_rate": 1.8735627525897618e-05,
|
| 519 |
+
"loss": 1.6443,
|
| 520 |
"step": 73
|
| 521 |
},
|
| 522 |
{
|
| 523 |
"epoch": 0.19023136246786632,
|
| 524 |
"grad_norm": 0.0,
|
| 525 |
"learning_rate": 1.8694766351966665e-05,
|
| 526 |
+
"loss": 1.6631,
|
| 527 |
"step": 74
|
| 528 |
},
|
| 529 |
{
|
| 530 |
"epoch": 0.1928020565552699,
|
| 531 |
"grad_norm": 0.0,
|
| 532 |
"learning_rate": 1.8653301406759827e-05,
|
| 533 |
+
"loss": 1.6873,
|
| 534 |
"step": 75
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 0.1953727506426735,
|
| 538 |
"grad_norm": 0.0,
|
| 539 |
"learning_rate": 1.8611235569634852e-05,
|
| 540 |
+
"loss": 1.7046,
|
| 541 |
"step": 76
|
| 542 |
},
|
| 543 |
{
|
| 544 |
"epoch": 0.19794344473007713,
|
| 545 |
"grad_norm": 0.0,
|
| 546 |
"learning_rate": 1.8568571761675893e-05,
|
| 547 |
+
"loss": 1.7002,
|
| 548 |
"step": 77
|
| 549 |
},
|
| 550 |
{
|
| 551 |
"epoch": 0.20051413881748073,
|
| 552 |
"grad_norm": 0.0,
|
| 553 |
"learning_rate": 1.8525312945490647e-05,
|
| 554 |
+
"loss": 1.698,
|
| 555 |
"step": 78
|
| 556 |
},
|
| 557 |
{
|
| 558 |
"epoch": 0.20308483290488433,
|
| 559 |
"grad_norm": 0.0,
|
| 560 |
"learning_rate": 1.8481462125004647e-05,
|
| 561 |
+
"loss": 1.6765,
|
| 562 |
"step": 79
|
| 563 |
},
|
| 564 |
{
|
| 565 |
"epoch": 0.20565552699228792,
|
| 566 |
"grad_norm": 0.0,
|
| 567 |
"learning_rate": 1.8437022345252666e-05,
|
| 568 |
+
"loss": 1.7185,
|
| 569 |
"step": 80
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.20822622107969152,
|
| 573 |
"grad_norm": 0.0,
|
| 574 |
"learning_rate": 1.8391996692167242e-05,
|
| 575 |
+
"loss": 1.6653,
|
| 576 |
"step": 81
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"epoch": 0.21079691516709512,
|
| 580 |
"grad_norm": 0.0,
|
| 581 |
"learning_rate": 1.8346388292364438e-05,
|
| 582 |
+
"loss": 1.7129,
|
| 583 |
"step": 82
|
| 584 |
},
|
| 585 |
{
|
| 586 |
"epoch": 0.2133676092544987,
|
| 587 |
"grad_norm": 0.0,
|
| 588 |
"learning_rate": 1.8300200312926674e-05,
|
| 589 |
+
"loss": 1.6709,
|
| 590 |
"step": 83
|
| 591 |
},
|
| 592 |
{
|
| 593 |
"epoch": 0.2159383033419023,
|
| 594 |
"grad_norm": 0.0,
|
| 595 |
"learning_rate": 1.8253435961182844e-05,
|
| 596 |
+
"loss": 1.6597,
|
| 597 |
"step": 84
|
| 598 |
},
|
| 599 |
{
|
| 600 |
"epoch": 0.2185089974293059,
|
| 601 |
"grad_norm": 0.0,
|
| 602 |
"learning_rate": 1.8206098484485563e-05,
|
| 603 |
+
"loss": 1.6812,
|
| 604 |
"step": 85
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.2210796915167095,
|
| 608 |
"grad_norm": 0.0,
|
| 609 |
"learning_rate": 1.8158191169985696e-05,
|
| 610 |
+
"loss": 1.6792,
|
| 611 |
"step": 86
|
| 612 |
},
|
| 613 |
{
|
| 614 |
"epoch": 0.2236503856041131,
|
| 615 |
"grad_norm": 0.0,
|
| 616 |
"learning_rate": 1.810971734440408e-05,
|
| 617 |
+
"loss": 1.6404,
|
| 618 |
"step": 87
|
| 619 |
},
|
| 620 |
{
|
| 621 |
"epoch": 0.2262210796915167,
|
| 622 |
"grad_norm": 0.0,
|
| 623 |
"learning_rate": 1.806068037380052e-05,
|
| 624 |
+
"loss": 1.6528,
|
| 625 |
"step": 88
|
| 626 |
},
|
| 627 |
{
|
| 628 |
"epoch": 0.22879177377892032,
|
| 629 |
"grad_norm": 0.0,
|
| 630 |
"learning_rate": 1.801108366334004e-05,
|
| 631 |
+
"loss": 1.6775,
|
| 632 |
"step": 89
|
| 633 |
},
|
| 634 |
{
|
| 635 |
"epoch": 0.23136246786632392,
|
| 636 |
"grad_norm": 0.0,
|
| 637 |
"learning_rate": 1.796093065705644e-05,
|
| 638 |
+
"loss": 1.679,
|
| 639 |
"step": 90
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"epoch": 0.23393316195372751,
|
| 643 |
"grad_norm": 0.0,
|
| 644 |
"learning_rate": 1.791022483761312e-05,
|
| 645 |
+
"loss": 1.658,
|
| 646 |
"step": 91
|
| 647 |
},
|
| 648 |
{
|
| 649 |
"epoch": 0.2365038560411311,
|
| 650 |
"grad_norm": 0.0,
|
| 651 |
"learning_rate": 1.7858969726061262e-05,
|
| 652 |
+
"loss": 1.6277,
|
| 653 |
"step": 92
|
| 654 |
},
|
| 655 |
{
|
| 656 |
"epoch": 0.2390745501285347,
|
| 657 |
"grad_norm": 0.0,
|
| 658 |
"learning_rate": 1.7807168881595304e-05,
|
| 659 |
+
"loss": 1.6602,
|
| 660 |
"step": 93
|
| 661 |
},
|
| 662 |
{
|
| 663 |
"epoch": 0.2416452442159383,
|
| 664 |
"grad_norm": 0.0,
|
| 665 |
"learning_rate": 1.7754825901305814e-05,
|
| 666 |
+
"loss": 1.6758,
|
| 667 |
"step": 94
|
| 668 |
},
|
| 669 |
{
|
| 670 |
"epoch": 0.2442159383033419,
|
| 671 |
"grad_norm": 0.0,
|
| 672 |
"learning_rate": 1.7701944419929673e-05,
|
| 673 |
+
"loss": 1.6353,
|
| 674 |
"step": 95
|
| 675 |
},
|
| 676 |
{
|
| 677 |
"epoch": 0.2467866323907455,
|
| 678 |
"grad_norm": 0.0,
|
| 679 |
"learning_rate": 1.7648528109597704e-05,
|
| 680 |
+
"loss": 1.6602,
|
| 681 |
"step": 96
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.2493573264781491,
|
| 685 |
"grad_norm": 0.0,
|
| 686 |
"learning_rate": 1.7594580679579654e-05,
|
| 687 |
+
"loss": 1.6404,
|
| 688 |
"step": 97
|
| 689 |
},
|
| 690 |
{
|
| 691 |
"epoch": 0.2519280205655527,
|
| 692 |
"grad_norm": 0.0,
|
| 693 |
"learning_rate": 1.7540105876026647e-05,
|
| 694 |
+
"loss": 1.6365,
|
| 695 |
"step": 98
|
| 696 |
},
|
| 697 |
{
|
| 698 |
"epoch": 0.2544987146529563,
|
| 699 |
"grad_norm": 0.0,
|
| 700 |
"learning_rate": 1.7485107481711014e-05,
|
| 701 |
+
"loss": 1.6353,
|
| 702 |
"step": 99
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"epoch": 0.2570694087403599,
|
| 706 |
"grad_norm": 0.0,
|
| 707 |
"learning_rate": 1.7429589315763637e-05,
|
| 708 |
+
"loss": 1.6541,
|
| 709 |
"step": 100
|
| 710 |
},
|
| 711 |
{
|
| 712 |
"epoch": 0.2596401028277635,
|
| 713 |
"grad_norm": 0.0,
|
| 714 |
"learning_rate": 1.737355523340875e-05,
|
| 715 |
+
"loss": 1.6133,
|
| 716 |
"step": 101
|
| 717 |
},
|
| 718 |
{
|
| 719 |
"epoch": 0.2622107969151671,
|
| 720 |
"grad_norm": 0.0,
|
| 721 |
"learning_rate": 1.7317009125696208e-05,
|
| 722 |
+
"loss": 1.6687,
|
| 723 |
"step": 102
|
| 724 |
},
|
| 725 |
{
|
| 726 |
"epoch": 0.2647814910025707,
|
| 727 |
"grad_norm": 0.0,
|
| 728 |
"learning_rate": 1.725995491923131e-05,
|
| 729 |
+
"loss": 1.636,
|
| 730 |
"step": 103
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.26735218508997427,
|
| 734 |
"grad_norm": 0.0,
|
| 735 |
"learning_rate": 1.7202396575902118e-05,
|
| 736 |
+
"loss": 1.6497,
|
| 737 |
"step": 104
|
| 738 |
},
|
| 739 |
{
|
| 740 |
"epoch": 0.2699228791773779,
|
| 741 |
"grad_norm": 0.0,
|
| 742 |
"learning_rate": 1.714433809260435e-05,
|
| 743 |
+
"loss": 1.6458,
|
| 744 |
"step": 105
|
| 745 |
},
|
| 746 |
{
|
| 747 |
"epoch": 0.27249357326478146,
|
| 748 |
"grad_norm": 0.0,
|
| 749 |
"learning_rate": 1.7085783500963825e-05,
|
| 750 |
+
"loss": 1.624,
|
| 751 |
"step": 106
|
| 752 |
},
|
| 753 |
{
|
| 754 |
"epoch": 0.2750642673521851,
|
| 755 |
"grad_norm": 0.0,
|
| 756 |
"learning_rate": 1.702673686705651e-05,
|
| 757 |
+
"loss": 1.6353,
|
| 758 |
"step": 107
|
| 759 |
},
|
| 760 |
{
|
| 761 |
"epoch": 0.2776349614395887,
|
| 762 |
"grad_norm": 0.0,
|
| 763 |
"learning_rate": 1.6967202291126174e-05,
|
| 764 |
+
"loss": 1.6406,
|
| 765 |
"step": 108
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"epoch": 0.2802056555269923,
|
| 769 |
"grad_norm": 0.0,
|
| 770 |
"learning_rate": 1.690718390729964e-05,
|
| 771 |
+
"loss": 1.6323,
|
| 772 |
"step": 109
|
| 773 |
},
|
| 774 |
{
|
| 775 |
"epoch": 0.2827763496143959,
|
| 776 |
"grad_norm": 0.0,
|
| 777 |
"learning_rate": 1.684668588329973e-05,
|
| 778 |
+
"loss": 1.665,
|
| 779 |
"step": 110
|
| 780 |
},
|
| 781 |
{
|
| 782 |
"epoch": 0.2853470437017995,
|
| 783 |
"grad_norm": 0.0,
|
| 784 |
"learning_rate": 1.6785712420155864e-05,
|
| 785 |
+
"loss": 1.635,
|
| 786 |
"step": 111
|
| 787 |
},
|
| 788 |
{
|
| 789 |
"epoch": 0.2879177377892031,
|
| 790 |
"grad_norm": 0.0,
|
| 791 |
"learning_rate": 1.67242677519123e-05,
|
| 792 |
+
"loss": 1.6335,
|
| 793 |
"step": 112
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.29048843187660667,
|
| 797 |
"grad_norm": 0.0,
|
| 798 |
"learning_rate": 1.6662356145334158e-05,
|
| 799 |
+
"loss": 1.6846,
|
| 800 |
"step": 113
|
| 801 |
},
|
| 802 |
{
|
| 803 |
"epoch": 0.2930591259640103,
|
| 804 |
"grad_norm": 0.0,
|
| 805 |
"learning_rate": 1.6599981899611103e-05,
|
| 806 |
+
"loss": 1.6353,
|
| 807 |
"step": 114
|
| 808 |
},
|
| 809 |
{
|
| 810 |
"epoch": 0.29562982005141386,
|
| 811 |
"grad_norm": 0.0,
|
| 812 |
"learning_rate": 1.653714934605883e-05,
|
| 813 |
+
"loss": 1.6189,
|
| 814 |
"step": 115
|
| 815 |
},
|
| 816 |
{
|
| 817 |
"epoch": 0.2982005141388175,
|
| 818 |
"grad_norm": 0.0,
|
| 819 |
"learning_rate": 1.647386284781828e-05,
|
| 820 |
+
"loss": 1.7021,
|
| 821 |
"step": 116
|
| 822 |
},
|
| 823 |
{
|
| 824 |
"epoch": 0.30077120822622105,
|
| 825 |
"grad_norm": 0.0,
|
| 826 |
"learning_rate": 1.6410126799552653e-05,
|
| 827 |
+
"loss": 1.6777,
|
| 828 |
"step": 117
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"epoch": 0.3033419023136247,
|
| 832 |
"grad_norm": 0.0,
|
| 833 |
"learning_rate": 1.6345945627142264e-05,
|
| 834 |
+
"loss": 1.6377,
|
| 835 |
"step": 118
|
| 836 |
},
|
| 837 |
{
|
| 838 |
"epoch": 0.3059125964010283,
|
| 839 |
"grad_norm": 0.0,
|
| 840 |
"learning_rate": 1.628132378737718e-05,
|
| 841 |
+
"loss": 1.6616,
|
| 842 |
"step": 119
|
| 843 |
},
|
| 844 |
{
|
| 845 |
"epoch": 0.30848329048843187,
|
| 846 |
"grad_norm": 0.0,
|
| 847 |
"learning_rate": 1.6216265767647756e-05,
|
| 848 |
+
"loss": 1.616,
|
| 849 |
"step": 120
|
| 850 |
},
|
| 851 |
{
|
| 852 |
"epoch": 0.3110539845758355,
|
| 853 |
"grad_norm": 0.0,
|
| 854 |
"learning_rate": 1.615077608563302e-05,
|
| 855 |
+
"loss": 1.6816,
|
| 856 |
"step": 121
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.31362467866323906,
|
| 860 |
"grad_norm": 0.0,
|
| 861 |
"learning_rate": 1.6084859288986957e-05,
|
| 862 |
+
"loss": 1.6099,
|
| 863 |
"step": 122
|
| 864 |
},
|
| 865 |
{
|
| 866 |
"epoch": 0.3161953727506427,
|
| 867 |
"grad_norm": 0.0,
|
| 868 |
"learning_rate": 1.601851995502272e-05,
|
| 869 |
+
"loss": 1.6274,
|
| 870 |
"step": 123
|
| 871 |
},
|
| 872 |
{
|
| 873 |
"epoch": 0.31876606683804626,
|
| 874 |
"grad_norm": 0.0,
|
| 875 |
"learning_rate": 1.5951762690394788e-05,
|
| 876 |
+
"loss": 1.6663,
|
| 877 |
"step": 124
|
| 878 |
},
|
| 879 |
{
|
| 880 |
"epoch": 0.3213367609254499,
|
| 881 |
"grad_norm": 0.0,
|
| 882 |
"learning_rate": 1.5884592130779056e-05,
|
| 883 |
+
"loss": 1.6494,
|
| 884 |
"step": 125
|
| 885 |
},
|
| 886 |
{
|
| 887 |
"epoch": 0.32390745501285345,
|
| 888 |
"grad_norm": 0.0,
|
| 889 |
"learning_rate": 1.581701294055095e-05,
|
| 890 |
+
"loss": 1.614,
|
| 891 |
"step": 126
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"epoch": 0.3264781491002571,
|
| 895 |
"grad_norm": 0.0,
|
| 896 |
"learning_rate": 1.5749029812461515e-05,
|
| 897 |
+
"loss": 1.6265,
|
| 898 |
"step": 127
|
| 899 |
},
|
| 900 |
{
|
| 901 |
"epoch": 0.32904884318766064,
|
| 902 |
"grad_norm": 0.0,
|
| 903 |
"learning_rate": 1.568064746731156e-05,
|
| 904 |
+
"loss": 1.5913,
|
| 905 |
"step": 128
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.33161953727506427,
|
| 909 |
"grad_norm": 0.0,
|
| 910 |
"learning_rate": 1.5611870653623826e-05,
|
| 911 |
+
"loss": 1.5984,
|
| 912 |
"step": 129
|
| 913 |
},
|
| 914 |
{
|
| 915 |
"epoch": 0.3341902313624679,
|
| 916 |
"grad_norm": 0.0,
|
| 917 |
"learning_rate": 1.5542704147313257e-05,
|
| 918 |
+
"loss": 1.6343,
|
| 919 |
"step": 130
|
| 920 |
},
|
| 921 |
{
|
| 922 |
"epoch": 0.33676092544987146,
|
| 923 |
"grad_norm": 0.0,
|
| 924 |
"learning_rate": 1.5473152751355353e-05,
|
| 925 |
+
"loss": 1.6355,
|
| 926 |
"step": 131
|
| 927 |
},
|
| 928 |
{
|
| 929 |
"epoch": 0.3393316195372751,
|
| 930 |
"grad_norm": 0.0,
|
| 931 |
"learning_rate": 1.5403221295452647e-05,
|
| 932 |
+
"loss": 1.647,
|
| 933 |
"step": 132
|
| 934 |
},
|
| 935 |
{
|
| 936 |
"epoch": 0.34190231362467866,
|
| 937 |
"grad_norm": 0.0,
|
| 938 |
"learning_rate": 1.5332914635699327e-05,
|
| 939 |
+
"loss": 1.6191,
|
| 940 |
"step": 133
|
| 941 |
},
|
| 942 |
{
|
| 943 |
"epoch": 0.3444730077120823,
|
| 944 |
"grad_norm": 0.0,
|
| 945 |
"learning_rate": 1.5262237654244026e-05,
|
| 946 |
+
"loss": 1.624,
|
| 947 |
"step": 134
|
| 948 |
},
|
| 949 |
{
|
| 950 |
"epoch": 0.34704370179948585,
|
| 951 |
"grad_norm": 0.0,
|
| 952 |
"learning_rate": 1.5191195258950804e-05,
|
| 953 |
+
"loss": 1.6055,
|
| 954 |
"step": 135
|
| 955 |
},
|
| 956 |
{
|
| 957 |
"epoch": 0.3496143958868895,
|
| 958 |
"grad_norm": 0.0,
|
| 959 |
"learning_rate": 1.5119792383058338e-05,
|
| 960 |
+
"loss": 1.6492,
|
| 961 |
"step": 136
|
| 962 |
},
|
| 963 |
{
|
| 964 |
"epoch": 0.35218508997429304,
|
| 965 |
"grad_norm": 0.0,
|
| 966 |
"learning_rate": 1.5048033984837352e-05,
|
| 967 |
+
"loss": 1.6155,
|
| 968 |
"step": 137
|
| 969 |
},
|
| 970 |
{
|
| 971 |
"epoch": 0.35475578406169667,
|
| 972 |
"grad_norm": 0.0,
|
| 973 |
"learning_rate": 1.4975925047246319e-05,
|
| 974 |
+
"loss": 1.6042,
|
| 975 |
"step": 138
|
| 976 |
},
|
| 977 |
{
|
| 978 |
"epoch": 0.35732647814910024,
|
| 979 |
"grad_norm": 0.0,
|
| 980 |
"learning_rate": 1.4903470577585433e-05,
|
| 981 |
+
"loss": 1.6367,
|
| 982 |
"step": 139
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.35989717223650386,
|
| 986 |
"grad_norm": 0.0,
|
| 987 |
"learning_rate": 1.4830675607148899e-05,
|
| 988 |
+
"loss": 1.5928,
|
| 989 |
"step": 140
|
| 990 |
},
|
| 991 |
{
|
| 992 |
"epoch": 0.36246786632390743,
|
| 993 |
"grad_norm": 0.0,
|
| 994 |
"learning_rate": 1.475754519087557e-05,
|
| 995 |
+
"loss": 1.6526,
|
| 996 |
"step": 141
|
| 997 |
},
|
| 998 |
{
|
| 999 |
"epoch": 0.36503856041131105,
|
| 1000 |
"grad_norm": 0.0,
|
| 1001 |
"learning_rate": 1.4684084406997903e-05,
|
| 1002 |
+
"loss": 1.6362,
|
| 1003 |
"step": 142
|
| 1004 |
},
|
| 1005 |
{
|
| 1006 |
"epoch": 0.3676092544987147,
|
| 1007 |
"grad_norm": 0.0,
|
| 1008 |
"learning_rate": 1.4610298356689341e-05,
|
| 1009 |
+
"loss": 1.6201,
|
| 1010 |
"step": 143
|
| 1011 |
},
|
| 1012 |
{
|
| 1013 |
"epoch": 0.37017994858611825,
|
| 1014 |
"grad_norm": 0.0,
|
| 1015 |
"learning_rate": 1.453619216371008e-05,
|
| 1016 |
+
"loss": 1.6162,
|
| 1017 |
"step": 144
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.37275064267352187,
|
| 1021 |
"grad_norm": 0.0,
|
| 1022 |
"learning_rate": 1.446177097405127e-05,
|
| 1023 |
+
"loss": 1.6172,
|
| 1024 |
"step": 145
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
"epoch": 0.37532133676092544,
|
| 1028 |
"grad_norm": 0.0,
|
| 1029 |
"learning_rate": 1.4387039955577668e-05,
|
| 1030 |
+
"loss": 1.6301,
|
| 1031 |
"step": 146
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
"epoch": 0.37789203084832906,
|
| 1035 |
"grad_norm": 0.0,
|
| 1036 |
"learning_rate": 1.4312004297668791e-05,
|
| 1037 |
+
"loss": 1.6096,
|
| 1038 |
"step": 147
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
"epoch": 0.38046272493573263,
|
| 1042 |
"grad_norm": 0.0,
|
| 1043 |
"learning_rate": 1.4236669210858544e-05,
|
| 1044 |
+
"loss": 1.6152,
|
| 1045 |
"step": 148
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
"epoch": 0.38303341902313626,
|
| 1049 |
"grad_norm": 0.0,
|
| 1050 |
"learning_rate": 1.4161039926473412e-05,
|
| 1051 |
+
"loss": 1.6321,
|
| 1052 |
"step": 149
|
| 1053 |
},
|
| 1054 |
{
|
| 1055 |
"epoch": 0.3856041131105398,
|
| 1056 |
"grad_norm": 0.0,
|
| 1057 |
"learning_rate": 1.4085121696269185e-05,
|
| 1058 |
+
"loss": 1.5957,
|
| 1059 |
"step": 150
|
| 1060 |
},
|
| 1061 |
{
|
| 1062 |
"epoch": 0.38817480719794345,
|
| 1063 |
"grad_norm": 0.0,
|
| 1064 |
"learning_rate": 1.4008919792066273e-05,
|
| 1065 |
+
"loss": 1.6421,
|
| 1066 |
"step": 151
|
| 1067 |
},
|
| 1068 |
{
|
| 1069 |
"epoch": 0.390745501285347,
|
| 1070 |
"grad_norm": 0.0,
|
| 1071 |
"learning_rate": 1.3932439505383628e-05,
|
| 1072 |
+
"loss": 1.6189,
|
| 1073 |
"step": 152
|
| 1074 |
},
|
| 1075 |
{
|
| 1076 |
"epoch": 0.39331619537275064,
|
| 1077 |
"grad_norm": 0.0,
|
| 1078 |
"learning_rate": 1.385568614707129e-05,
|
| 1079 |
+
"loss": 1.6106,
|
| 1080 |
"step": 153
|
| 1081 |
},
|
| 1082 |
{
|
| 1083 |
"epoch": 0.39588688946015427,
|
| 1084 |
"grad_norm": 0.0,
|
| 1085 |
"learning_rate": 1.3778665046941616e-05,
|
| 1086 |
+
"loss": 1.6321,
|
| 1087 |
"step": 154
|
| 1088 |
},
|
| 1089 |
{
|
| 1090 |
"epoch": 0.39845758354755784,
|
| 1091 |
"grad_norm": 0.0,
|
| 1092 |
"learning_rate": 1.3701381553399147e-05,
|
| 1093 |
+
"loss": 1.5796,
|
| 1094 |
"step": 155
|
| 1095 |
},
|
| 1096 |
{
|
| 1097 |
"epoch": 0.40102827763496146,
|
| 1098 |
"grad_norm": 0.0,
|
| 1099 |
"learning_rate": 1.3623841033069232e-05,
|
| 1100 |
+
"loss": 1.6555,
|
| 1101 |
"step": 156
|
| 1102 |
},
|
| 1103 |
{
|
| 1104 |
"epoch": 0.40359897172236503,
|
| 1105 |
"grad_norm": 0.0,
|
| 1106 |
"learning_rate": 1.3546048870425356e-05,
|
| 1107 |
+
"loss": 1.6028,
|
| 1108 |
"step": 157
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.40616966580976865,
|
| 1112 |
"grad_norm": 0.0,
|
| 1113 |
"learning_rate": 1.3468010467415248e-05,
|
| 1114 |
+
"loss": 1.5969,
|
| 1115 |
"step": 158
|
| 1116 |
},
|
| 1117 |
{
|
| 1118 |
"epoch": 0.4087403598971722,
|
| 1119 |
"grad_norm": 0.0,
|
| 1120 |
"learning_rate": 1.3389731243085747e-05,
|
| 1121 |
+
"loss": 1.6077,
|
| 1122 |
"step": 159
|
| 1123 |
},
|
| 1124 |
{
|
| 1125 |
"epoch": 0.41131105398457585,
|
| 1126 |
"grad_norm": 0.0,
|
| 1127 |
"learning_rate": 1.3311216633206514e-05,
|
| 1128 |
+
"loss": 1.5762,
|
| 1129 |
"step": 160
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.4138817480719794,
|
| 1133 |
"grad_norm": 0.0,
|
| 1134 |
"learning_rate": 1.3232472089892567e-05,
|
| 1135 |
+
"loss": 1.6079,
|
| 1136 |
"step": 161
|
| 1137 |
},
|
| 1138 |
{
|
| 1139 |
"epoch": 0.41645244215938304,
|
| 1140 |
"grad_norm": 0.0,
|
| 1141 |
"learning_rate": 1.315350308122567e-05,
|
| 1142 |
+
"loss": 1.5994,
|
| 1143 |
"step": 162
|
| 1144 |
},
|
| 1145 |
{
|
| 1146 |
"epoch": 0.4190231362467866,
|
| 1147 |
"grad_norm": 0.0,
|
| 1148 |
"learning_rate": 1.3074315090874652e-05,
|
| 1149 |
+
"loss": 1.5732,
|
| 1150 |
"step": 163
|
| 1151 |
},
|
| 1152 |
{
|
| 1153 |
"epoch": 0.42159383033419023,
|
| 1154 |
"grad_norm": 0.0,
|
| 1155 |
"learning_rate": 1.2994913617714573e-05,
|
| 1156 |
+
"loss": 1.5901,
|
| 1157 |
"step": 164
|
| 1158 |
},
|
| 1159 |
{
|
| 1160 |
"epoch": 0.4241645244215938,
|
| 1161 |
"grad_norm": 0.0,
|
| 1162 |
"learning_rate": 1.2915304175444929e-05,
|
| 1163 |
+
"loss": 1.6138,
|
| 1164 |
"step": 165
|
| 1165 |
},
|
| 1166 |
{
|
| 1167 |
"epoch": 0.4267352185089974,
|
| 1168 |
"grad_norm": 0.0,
|
| 1169 |
"learning_rate": 1.2835492292206735e-05,
|
| 1170 |
+
"loss": 1.5945,
|
| 1171 |
"step": 166
|
| 1172 |
},
|
| 1173 |
{
|
| 1174 |
"epoch": 0.42930591259640105,
|
| 1175 |
"grad_norm": 0.0,
|
| 1176 |
"learning_rate": 1.2755483510198668e-05,
|
| 1177 |
+
"loss": 1.6067,
|
| 1178 |
"step": 167
|
| 1179 |
},
|
| 1180 |
{
|
| 1181 |
"epoch": 0.4318766066838046,
|
| 1182 |
"grad_norm": 0.0,
|
| 1183 |
"learning_rate": 1.2675283385292212e-05,
|
| 1184 |
+
"loss": 1.5957,
|
| 1185 |
"step": 168
|
| 1186 |
},
|
| 1187 |
{
|
| 1188 |
"epoch": 0.43444730077120824,
|
| 1189 |
"grad_norm": 0.0,
|
| 1190 |
"learning_rate": 1.2594897486645836e-05,
|
| 1191 |
+
"loss": 1.6089,
|
| 1192 |
"step": 169
|
| 1193 |
},
|
| 1194 |
{
|
| 1195 |
"epoch": 0.4370179948586118,
|
| 1196 |
"grad_norm": 0.0,
|
| 1197 |
"learning_rate": 1.2514331396318298e-05,
|
| 1198 |
+
"loss": 1.6335,
|
| 1199 |
"step": 170
|
| 1200 |
},
|
| 1201 |
{
|
| 1202 |
"epoch": 0.43958868894601544,
|
| 1203 |
"grad_norm": 0.0,
|
| 1204 |
"learning_rate": 1.2433590708880991e-05,
|
| 1205 |
+
"loss": 1.6406,
|
| 1206 |
"step": 171
|
| 1207 |
},
|
| 1208 |
{
|
| 1209 |
"epoch": 0.442159383033419,
|
| 1210 |
"grad_norm": 0.0,
|
| 1211 |
"learning_rate": 1.2352681031029476e-05,
|
| 1212 |
+
"loss": 1.5759,
|
| 1213 |
"step": 172
|
| 1214 |
},
|
| 1215 |
{
|
| 1216 |
"epoch": 0.44473007712082263,
|
| 1217 |
"grad_norm": 0.0,
|
| 1218 |
"learning_rate": 1.2271607981194132e-05,
|
| 1219 |
+
"loss": 1.5955,
|
| 1220 |
"step": 173
|
| 1221 |
},
|
| 1222 |
{
|
| 1223 |
"epoch": 0.4473007712082262,
|
| 1224 |
"grad_norm": 0.0,
|
| 1225 |
"learning_rate": 1.2190377189150016e-05,
|
| 1226 |
+
"loss": 1.6069,
|
| 1227 |
"step": 174
|
| 1228 |
},
|
| 1229 |
{
|
| 1230 |
"epoch": 0.4498714652956298,
|
| 1231 |
"grad_norm": 0.0,
|
| 1232 |
"learning_rate": 1.2108994295625924e-05,
|
| 1233 |
+
"loss": 1.5796,
|
| 1234 |
"step": 175
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.4524421593830334,
|
| 1238 |
"grad_norm": 0.0,
|
| 1239 |
"learning_rate": 1.2027464951912703e-05,
|
| 1240 |
+
"loss": 1.5952,
|
| 1241 |
"step": 176
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.455012853470437,
|
| 1245 |
"grad_norm": 0.0,
|
| 1246 |
"learning_rate": 1.1945794819470805e-05,
|
| 1247 |
+
"loss": 1.6213,
|
| 1248 |
"step": 177
|
| 1249 |
},
|
| 1250 |
{
|
| 1251 |
"epoch": 0.45758354755784064,
|
| 1252 |
"grad_norm": 0.0,
|
| 1253 |
"learning_rate": 1.1863989569537165e-05,
|
| 1254 |
+
"loss": 1.5974,
|
| 1255 |
"step": 178
|
| 1256 |
},
|
| 1257 |
{
|
| 1258 |
"epoch": 0.4601542416452442,
|
| 1259 |
"grad_norm": 0.0,
|
| 1260 |
"learning_rate": 1.1782054882731377e-05,
|
| 1261 |
+
"loss": 1.5188,
|
| 1262 |
"step": 179
|
| 1263 |
},
|
| 1264 |
{
|
| 1265 |
"epoch": 0.46272493573264784,
|
| 1266 |
"grad_norm": 0.0,
|
| 1267 |
"learning_rate": 1.1699996448661242e-05,
|
| 1268 |
+
"loss": 1.5964,
|
| 1269 |
"step": 180
|
| 1270 |
},
|
| 1271 |
{
|
| 1272 |
"epoch": 0.4652956298200514,
|
| 1273 |
"grad_norm": 0.0,
|
| 1274 |
"learning_rate": 1.161781996552765e-05,
|
| 1275 |
+
"loss": 1.5681,
|
| 1276 |
"step": 181
|
| 1277 |
},
|
| 1278 |
{
|
| 1279 |
"epoch": 0.46786632390745503,
|
| 1280 |
"grad_norm": 0.0,
|
| 1281 |
"learning_rate": 1.1535531139728918e-05,
|
| 1282 |
+
"loss": 1.5938,
|
| 1283 |
"step": 182
|
| 1284 |
},
|
| 1285 |
{
|
| 1286 |
"epoch": 0.4704370179948586,
|
| 1287 |
"grad_norm": 0.0,
|
| 1288 |
"learning_rate": 1.1453135685464524e-05,
|
| 1289 |
+
"loss": 1.574,
|
| 1290 |
"step": 183
|
| 1291 |
},
|
| 1292 |
{
|
| 1293 |
"epoch": 0.4730077120822622,
|
| 1294 |
"grad_norm": 0.0,
|
| 1295 |
"learning_rate": 1.1370639324338313e-05,
|
| 1296 |
+
"loss": 1.5872,
|
| 1297 |
"step": 184
|
| 1298 |
},
|
| 1299 |
{
|
| 1300 |
"epoch": 0.4755784061696658,
|
| 1301 |
"grad_norm": 0.0,
|
| 1302 |
"learning_rate": 1.1288047784961166e-05,
|
| 1303 |
+
"loss": 1.5806,
|
| 1304 |
"step": 185
|
| 1305 |
},
|
| 1306 |
{
|
| 1307 |
"epoch": 0.4781491002570694,
|
| 1308 |
"grad_norm": 0.0,
|
| 1309 |
"learning_rate": 1.1205366802553231e-05,
|
| 1310 |
+
"loss": 1.5542,
|
| 1311 |
"step": 186
|
| 1312 |
},
|
| 1313 |
{
|
| 1314 |
"epoch": 0.480719794344473,
|
| 1315 |
"grad_norm": 0.0,
|
| 1316 |
"learning_rate": 1.1122602118545642e-05,
|
| 1317 |
+
"loss": 1.5723,
|
| 1318 |
"step": 187
|
| 1319 |
},
|
| 1320 |
{
|
| 1321 |
"epoch": 0.4832904884318766,
|
| 1322 |
"grad_norm": 0.0,
|
| 1323 |
"learning_rate": 1.1039759480181836e-05,
|
| 1324 |
+
"loss": 1.5645,
|
| 1325 |
"step": 188
|
| 1326 |
},
|
| 1327 |
{
|
| 1328 |
"epoch": 0.48586118251928023,
|
| 1329 |
"grad_norm": 0.0,
|
| 1330 |
"learning_rate": 1.0956844640118462e-05,
|
| 1331 |
+
"loss": 1.5884,
|
| 1332 |
"step": 189
|
| 1333 |
},
|
| 1334 |
{
|
| 1335 |
"epoch": 0.4884318766066838,
|
| 1336 |
"grad_norm": 0.0,
|
| 1337 |
"learning_rate": 1.0873863356025911e-05,
|
| 1338 |
+
"loss": 1.5559,
|
| 1339 |
"step": 190
|
| 1340 |
},
|
| 1341 |
{
|
| 1342 |
"epoch": 0.4910025706940874,
|
| 1343 |
"grad_norm": 0.0,
|
| 1344 |
"learning_rate": 1.0790821390188493e-05,
|
| 1345 |
+
"loss": 1.5623,
|
| 1346 |
"step": 191
|
| 1347 |
},
|
| 1348 |
{
|
| 1349 |
"epoch": 0.493573264781491,
|
| 1350 |
"grad_norm": 0.0,
|
| 1351 |
"learning_rate": 1.0707724509104318e-05,
|
| 1352 |
+
"loss": 1.5916,
|
| 1353 |
"step": 192
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.4961439588688946,
|
| 1357 |
"grad_norm": 0.0,
|
| 1358 |
"learning_rate": 1.062457848308484e-05,
|
| 1359 |
+
"loss": 1.5696,
|
| 1360 |
"step": 193
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.4987146529562982,
|
| 1364 |
"grad_norm": 0.0,
|
| 1365 |
"learning_rate": 1.0541389085854177e-05,
|
| 1366 |
+
"loss": 1.5913,
|
| 1367 |
"step": 194
|
| 1368 |
},
|
| 1369 |
{
|
| 1370 |
"epoch": 0.5012853470437018,
|
| 1371 |
"grad_norm": 0.0,
|
| 1372 |
"learning_rate": 1.0458162094148185e-05,
|
| 1373 |
+
"loss": 1.5439,
|
| 1374 |
"step": 195
|
| 1375 |
},
|
| 1376 |
{
|
| 1377 |
"epoch": 0.5038560411311054,
|
| 1378 |
"grad_norm": 0.0,
|
| 1379 |
"learning_rate": 1.0374903287313307e-05,
|
| 1380 |
+
"loss": 1.6013,
|
| 1381 |
"step": 196
|
| 1382 |
},
|
| 1383 |
{
|
| 1384 |
"epoch": 0.506426735218509,
|
| 1385 |
"grad_norm": 0.0,
|
| 1386 |
"learning_rate": 1.029161844690525e-05,
|
| 1387 |
+
"loss": 1.5813,
|
| 1388 |
"step": 197
|
| 1389 |
},
|
| 1390 |
{
|
| 1391 |
"epoch": 0.5089974293059126,
|
| 1392 |
"grad_norm": 0.0,
|
| 1393 |
"learning_rate": 1.0208313356287505e-05,
|
| 1394 |
+
"loss": 1.5757,
|
| 1395 |
"step": 198
|
| 1396 |
},
|
| 1397 |
{
|
| 1398 |
"epoch": 0.5115681233933161,
|
| 1399 |
"grad_norm": 0.0,
|
| 1400 |
"learning_rate": 1.0124993800229774e-05,
|
| 1401 |
+
"loss": 1.5508,
|
| 1402 |
"step": 199
|
| 1403 |
},
|
| 1404 |
{
|
| 1405 |
"epoch": 0.5141388174807198,
|
| 1406 |
"grad_norm": 0.0,
|
| 1407 |
"learning_rate": 1.004166556450623e-05,
|
| 1408 |
+
"loss": 1.5774,
|
| 1409 |
"step": 200
|
| 1410 |
},
|
| 1411 |
{
|
| 1412 |
"epoch": 0.5167095115681234,
|
| 1413 |
"grad_norm": 0.0,
|
| 1414 |
"learning_rate": 9.958334435493776e-06,
|
| 1415 |
+
"loss": 1.594,
|
| 1416 |
"step": 201
|
| 1417 |
},
|
| 1418 |
{
|
| 1419 |
"epoch": 0.519280205655527,
|
| 1420 |
"grad_norm": 0.0,
|
| 1421 |
"learning_rate": 9.87500619977023e-06,
|
| 1422 |
+
"loss": 1.5977,
|
| 1423 |
"step": 202
|
| 1424 |
},
|
| 1425 |
{
|
| 1426 |
"epoch": 0.5218508997429306,
|
| 1427 |
"grad_norm": 0.0,
|
| 1428 |
"learning_rate": 9.791686643712498e-06,
|
| 1429 |
+
"loss": 1.5938,
|
| 1430 |
"step": 203
|
| 1431 |
},
|
| 1432 |
{
|
| 1433 |
"epoch": 0.5244215938303342,
|
| 1434 |
"grad_norm": 0.0,
|
| 1435 |
"learning_rate": 9.708381553094754e-06,
|
| 1436 |
+
"loss": 1.5371,
|
| 1437 |
"step": 204
|
| 1438 |
},
|
| 1439 |
{
|
| 1440 |
"epoch": 0.5269922879177378,
|
| 1441 |
"grad_norm": 0.0,
|
| 1442 |
"learning_rate": 9.625096712686694e-06,
|
| 1443 |
+
"loss": 1.5315,
|
| 1444 |
"step": 205
|
| 1445 |
},
|
| 1446 |
{
|
| 1447 |
"epoch": 0.5295629820051414,
|
| 1448 |
"grad_norm": 0.0,
|
| 1449 |
"learning_rate": 9.541837905851817e-06,
|
| 1450 |
+
"loss": 1.5708,
|
| 1451 |
"step": 206
|
| 1452 |
},
|
| 1453 |
{
|
| 1454 |
"epoch": 0.532133676092545,
|
| 1455 |
"grad_norm": 0.0,
|
| 1456 |
"learning_rate": 9.458610914145826e-06,
|
| 1457 |
+
"loss": 1.5691,
|
| 1458 |
"step": 207
|
| 1459 |
},
|
| 1460 |
{
|
| 1461 |
"epoch": 0.5347043701799485,
|
| 1462 |
"grad_norm": 0.0,
|
| 1463 |
"learning_rate": 9.375421516915165e-06,
|
| 1464 |
+
"loss": 1.5881,
|
| 1465 |
"step": 208
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.5372750642673522,
|
| 1469 |
"grad_norm": 0.0,
|
| 1470 |
"learning_rate": 9.292275490895685e-06,
|
| 1471 |
+
"loss": 1.5732,
|
| 1472 |
"step": 209
|
| 1473 |
},
|
| 1474 |
{
|
| 1475 |
"epoch": 0.5398457583547558,
|
| 1476 |
"grad_norm": 0.0,
|
| 1477 |
"learning_rate": 9.209178609811509e-06,
|
| 1478 |
+
"loss": 1.5562,
|
| 1479 |
"step": 210
|
| 1480 |
},
|
| 1481 |
{
|
| 1482 |
"epoch": 0.5424164524421594,
|
| 1483 |
"grad_norm": 0.0,
|
| 1484 |
"learning_rate": 9.126136643974094e-06,
|
| 1485 |
+
"loss": 1.5603,
|
| 1486 |
"step": 211
|
| 1487 |
},
|
| 1488 |
{
|
| 1489 |
"epoch": 0.5449871465295629,
|
| 1490 |
"grad_norm": 0.0,
|
| 1491 |
"learning_rate": 9.043155359881538e-06,
|
| 1492 |
+
"loss": 1.5352,
|
| 1493 |
"step": 212
|
| 1494 |
},
|
| 1495 |
{
|
| 1496 |
"epoch": 0.5475578406169666,
|
| 1497 |
"grad_norm": 0.0,
|
| 1498 |
"learning_rate": 8.960240519818167e-06,
|
| 1499 |
+
"loss": 1.5647,
|
| 1500 |
"step": 213
|
| 1501 |
},
|
| 1502 |
{
|
| 1503 |
"epoch": 0.5501285347043702,
|
| 1504 |
"grad_norm": 0.0,
|
| 1505 |
"learning_rate": 8.877397881454358e-06,
|
| 1506 |
+
"loss": 1.5747,
|
| 1507 |
"step": 214
|
| 1508 |
},
|
| 1509 |
{
|
| 1510 |
"epoch": 0.5526992287917738,
|
| 1511 |
"grad_norm": 0.0,
|
| 1512 |
"learning_rate": 8.79463319744677e-06,
|
| 1513 |
+
"loss": 1.5586,
|
| 1514 |
"step": 215
|
| 1515 |
},
|
| 1516 |
{
|
| 1517 |
"epoch": 0.5552699228791774,
|
| 1518 |
"grad_norm": 0.0,
|
| 1519 |
"learning_rate": 8.711952215038837e-06,
|
| 1520 |
+
"loss": 1.5527,
|
| 1521 |
"step": 216
|
| 1522 |
},
|
| 1523 |
{
|
| 1524 |
"epoch": 0.5578406169665809,
|
| 1525 |
"grad_norm": 0.0,
|
| 1526 |
"learning_rate": 8.629360675661693e-06,
|
| 1527 |
+
"loss": 1.5374,
|
| 1528 |
"step": 217
|
| 1529 |
},
|
| 1530 |
{
|
| 1531 |
"epoch": 0.5604113110539846,
|
| 1532 |
"grad_norm": 0.0,
|
| 1533 |
"learning_rate": 8.546864314535478e-06,
|
| 1534 |
+
"loss": 1.5647,
|
| 1535 |
"step": 218
|
| 1536 |
},
|
| 1537 |
{
|
| 1538 |
"epoch": 0.5629820051413882,
|
| 1539 |
"grad_norm": 0.0,
|
| 1540 |
"learning_rate": 8.464468860271084e-06,
|
| 1541 |
+
"loss": 1.5356,
|
| 1542 |
"step": 219
|
| 1543 |
},
|
| 1544 |
{
|
| 1545 |
"epoch": 0.5655526992287918,
|
| 1546 |
"grad_norm": 0.0,
|
| 1547 |
"learning_rate": 8.382180034472353e-06,
|
| 1548 |
+
"loss": 1.5483,
|
| 1549 |
"step": 220
|
| 1550 |
},
|
| 1551 |
{
|
| 1552 |
"epoch": 0.5681233933161953,
|
| 1553 |
"grad_norm": 0.0,
|
| 1554 |
"learning_rate": 8.30000355133876e-06,
|
| 1555 |
+
"loss": 1.5386,
|
| 1556 |
"step": 221
|
| 1557 |
},
|
| 1558 |
{
|
| 1559 |
"epoch": 0.570694087403599,
|
| 1560 |
"grad_norm": 0.0,
|
| 1561 |
"learning_rate": 8.217945117268624e-06,
|
| 1562 |
+
"loss": 1.5552,
|
| 1563 |
"step": 222
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 0.5732647814910026,
|
| 1567 |
"grad_norm": 0.0,
|
| 1568 |
"learning_rate": 8.136010430462837e-06,
|
| 1569 |
+
"loss": 1.5635,
|
| 1570 |
"step": 223
|
| 1571 |
},
|
| 1572 |
{
|
| 1573 |
"epoch": 0.5758354755784062,
|
| 1574 |
"grad_norm": 0.0,
|
| 1575 |
"learning_rate": 8.0542051805292e-06,
|
| 1576 |
+
"loss": 1.5657,
|
| 1577 |
"step": 224
|
| 1578 |
},
|
| 1579 |
{
|
| 1580 |
"epoch": 0.5784061696658098,
|
| 1581 |
"grad_norm": 0.0,
|
| 1582 |
"learning_rate": 7.9725350480873e-06,
|
| 1583 |
+
"loss": 1.5386,
|
| 1584 |
"step": 225
|
| 1585 |
},
|
| 1586 |
{
|
| 1587 |
"epoch": 0.5809768637532133,
|
| 1588 |
"grad_norm": 0.0,
|
| 1589 |
"learning_rate": 7.89100570437408e-06,
|
| 1590 |
+
"loss": 1.6018,
|
| 1591 |
"step": 226
|
| 1592 |
},
|
| 1593 |
{
|
| 1594 |
"epoch": 0.583547557840617,
|
| 1595 |
"grad_norm": 0.0,
|
| 1596 |
"learning_rate": 7.809622810849986e-06,
|
| 1597 |
+
"loss": 1.5396,
|
| 1598 |
"step": 227
|
| 1599 |
},
|
| 1600 |
{
|
| 1601 |
"epoch": 0.5861182519280206,
|
| 1602 |
"grad_norm": 0.0,
|
| 1603 |
"learning_rate": 7.72839201880587e-06,
|
| 1604 |
+
"loss": 1.5474,
|
| 1605 |
"step": 228
|
| 1606 |
},
|
| 1607 |
{
|
| 1608 |
"epoch": 0.5886889460154242,
|
| 1609 |
"grad_norm": 0.0,
|
| 1610 |
"learning_rate": 7.647318968970528e-06,
|
| 1611 |
+
"loss": 1.5654,
|
| 1612 |
"step": 229
|
| 1613 |
},
|
| 1614 |
{
|
| 1615 |
"epoch": 0.5912596401028277,
|
| 1616 |
"grad_norm": 0.0,
|
| 1617 |
"learning_rate": 7.566409291119008e-06,
|
| 1618 |
+
"loss": 1.5732,
|
| 1619 |
"step": 230
|
| 1620 |
},
|
| 1621 |
{
|
| 1622 |
"epoch": 0.5938303341902313,
|
| 1623 |
"grad_norm": 0.0,
|
| 1624 |
"learning_rate": 7.485668603681706e-06,
|
| 1625 |
+
"loss": 1.5779,
|
| 1626 |
"step": 231
|
| 1627 |
},
|
| 1628 |
{
|
| 1629 |
"epoch": 0.596401028277635,
|
| 1630 |
"grad_norm": 0.0,
|
| 1631 |
"learning_rate": 7.405102513354166e-06,
|
| 1632 |
+
"loss": 1.5449,
|
| 1633 |
"step": 232
|
| 1634 |
},
|
| 1635 |
{
|
| 1636 |
"epoch": 0.5989717223650386,
|
| 1637 |
"grad_norm": 0.0,
|
| 1638 |
"learning_rate": 7.324716614707794e-06,
|
| 1639 |
+
"loss": 1.5408,
|
| 1640 |
"step": 233
|
| 1641 |
},
|
| 1642 |
{
|
| 1643 |
"epoch": 0.6015424164524421,
|
| 1644 |
"grad_norm": 0.0,
|
| 1645 |
"learning_rate": 7.2445164898013345e-06,
|
| 1646 |
+
"loss": 1.5403,
|
| 1647 |
"step": 234
|
| 1648 |
},
|
| 1649 |
{
|
| 1650 |
"epoch": 0.6041131105398457,
|
| 1651 |
"grad_norm": 0.0,
|
| 1652 |
"learning_rate": 7.1645077077932666e-06,
|
| 1653 |
+
"loss": 1.5159,
|
| 1654 |
"step": 235
|
| 1655 |
},
|
| 1656 |
{
|
| 1657 |
"epoch": 0.6066838046272494,
|
| 1658 |
"grad_norm": 0.0,
|
| 1659 |
"learning_rate": 7.084695824555074e-06,
|
| 1660 |
+
"loss": 1.5557,
|
| 1661 |
"step": 236
|
| 1662 |
},
|
| 1663 |
{
|
| 1664 |
"epoch": 0.609254498714653,
|
| 1665 |
"grad_norm": 0.0,
|
| 1666 |
"learning_rate": 7.005086382285426e-06,
|
| 1667 |
+
"loss": 1.5625,
|
| 1668 |
"step": 237
|
| 1669 |
},
|
| 1670 |
{
|
| 1671 |
"epoch": 0.6118251928020566,
|
| 1672 |
"grad_norm": 0.0,
|
| 1673 |
"learning_rate": 6.925684909125354e-06,
|
| 1674 |
+
"loss": 1.552,
|
| 1675 |
"step": 238
|
| 1676 |
},
|
| 1677 |
{
|
| 1678 |
"epoch": 0.6143958868894601,
|
| 1679 |
"grad_norm": 0.0,
|
| 1680 |
"learning_rate": 6.84649691877433e-06,
|
| 1681 |
+
"loss": 1.5488,
|
| 1682 |
"step": 239
|
| 1683 |
},
|
| 1684 |
{
|
| 1685 |
"epoch": 0.6169665809768637,
|
| 1686 |
"grad_norm": 0.0,
|
| 1687 |
"learning_rate": 6.767527910107437e-06,
|
| 1688 |
+
"loss": 1.5181,
|
| 1689 |
"step": 240
|
| 1690 |
},
|
| 1691 |
{
|
| 1692 |
"epoch": 0.6195372750642674,
|
| 1693 |
"grad_norm": 0.0,
|
| 1694 |
"learning_rate": 6.688783366793488e-06,
|
| 1695 |
+
"loss": 1.5403,
|
| 1696 |
"step": 241
|
| 1697 |
},
|
| 1698 |
{
|
| 1699 |
"epoch": 0.622107969151671,
|
| 1700 |
"grad_norm": 0.0,
|
| 1701 |
"learning_rate": 6.610268756914254e-06,
|
| 1702 |
+
"loss": 1.5662,
|
| 1703 |
"step": 242
|
| 1704 |
},
|
| 1705 |
{
|
| 1706 |
"epoch": 0.6246786632390745,
|
| 1707 |
"grad_norm": 0.0,
|
| 1708 |
"learning_rate": 6.5319895325847535e-06,
|
| 1709 |
+
"loss": 1.5222,
|
| 1710 |
"step": 243
|
| 1711 |
},
|
| 1712 |
{
|
| 1713 |
"epoch": 0.6272493573264781,
|
| 1714 |
"grad_norm": 0.0,
|
| 1715 |
"learning_rate": 6.453951129574644e-06,
|
| 1716 |
+
"loss": 1.5439,
|
| 1717 |
"step": 244
|
| 1718 |
},
|
| 1719 |
{
|
| 1720 |
"epoch": 0.6298200514138818,
|
| 1721 |
"grad_norm": 0.0,
|
| 1722 |
"learning_rate": 6.3761589669307745e-06,
|
| 1723 |
+
"loss": 1.5312,
|
| 1724 |
"step": 245
|
| 1725 |
},
|
| 1726 |
{
|
| 1727 |
"epoch": 0.6323907455012854,
|
| 1728 |
"grad_norm": 0.0,
|
| 1729 |
"learning_rate": 6.298618446600856e-06,
|
| 1730 |
+
"loss": 1.5383,
|
| 1731 |
"step": 246
|
| 1732 |
},
|
| 1733 |
{
|
| 1734 |
"epoch": 0.6349614395886889,
|
| 1735 |
"grad_norm": 0.0,
|
| 1736 |
"learning_rate": 6.221334953058389e-06,
|
| 1737 |
+
"loss": 1.5393,
|
| 1738 |
"step": 247
|
| 1739 |
},
|
| 1740 |
{
|
| 1741 |
"epoch": 0.6375321336760925,
|
| 1742 |
"grad_norm": 0.0,
|
| 1743 |
"learning_rate": 6.144313852928712e-06,
|
| 1744 |
+
"loss": 1.5247,
|
| 1745 |
"step": 248
|
| 1746 |
},
|
| 1747 |
{
|
| 1748 |
"epoch": 0.6401028277634961,
|
| 1749 |
"grad_norm": 0.0,
|
| 1750 |
"learning_rate": 6.067560494616374e-06,
|
| 1751 |
+
"loss": 1.5454,
|
| 1752 |
"step": 249
|
| 1753 |
},
|
| 1754 |
{
|
| 1755 |
"epoch": 0.6426735218508998,
|
| 1756 |
"grad_norm": 0.0,
|
| 1757 |
"learning_rate": 5.9910802079337285e-06,
|
| 1758 |
+
"loss": 1.5215,
|
| 1759 |
"step": 250
|
| 1760 |
},
|
| 1761 |
{
|
| 1762 |
"epoch": 0.6452442159383034,
|
| 1763 |
"grad_norm": 0.0,
|
| 1764 |
"learning_rate": 5.9148783037308154e-06,
|
| 1765 |
+
"loss": 1.5427,
|
| 1766 |
"step": 251
|
| 1767 |
},
|
| 1768 |
{
|
| 1769 |
"epoch": 0.6478149100257069,
|
| 1770 |
"grad_norm": 0.0,
|
| 1771 |
"learning_rate": 5.838960073526589e-06,
|
| 1772 |
+
"loss": 1.5427,
|
| 1773 |
"step": 252
|
| 1774 |
},
|
| 1775 |
{
|
| 1776 |
"epoch": 0.6503856041131105,
|
| 1777 |
"grad_norm": 0.0,
|
| 1778 |
"learning_rate": 5.763330789141457e-06,
|
| 1779 |
+
"loss": 1.5552,
|
| 1780 |
"step": 253
|
| 1781 |
},
|
| 1782 |
{
|
| 1783 |
"epoch": 0.6529562982005142,
|
| 1784 |
"grad_norm": 0.0,
|
| 1785 |
"learning_rate": 5.687995702331211e-06,
|
| 1786 |
+
"loss": 1.5388,
|
| 1787 |
"step": 254
|
| 1788 |
},
|
| 1789 |
{
|
| 1790 |
"epoch": 0.6555269922879178,
|
| 1791 |
"grad_norm": 0.0,
|
| 1792 |
"learning_rate": 5.612960044422335e-06,
|
| 1793 |
+
"loss": 1.5854,
|
| 1794 |
"step": 255
|
| 1795 |
},
|
| 1796 |
{
|
| 1797 |
"epoch": 0.6580976863753213,
|
| 1798 |
"grad_norm": 0.0,
|
| 1799 |
"learning_rate": 5.538229025948729e-06,
|
| 1800 |
+
"loss": 1.5588,
|
| 1801 |
"step": 256
|
| 1802 |
},
|
| 1803 |
{
|
| 1804 |
"epoch": 0.6606683804627249,
|
| 1805 |
"grad_norm": 0.0,
|
| 1806 |
"learning_rate": 5.463807836289921e-06,
|
| 1807 |
+
"loss": 1.5217,
|
| 1808 |
"step": 257
|
| 1809 |
},
|
| 1810 |
{
|
| 1811 |
"epoch": 0.6632390745501285,
|
| 1812 |
"grad_norm": 0.0,
|
| 1813 |
"learning_rate": 5.389701643310661e-06,
|
| 1814 |
+
"loss": 1.5066,
|
| 1815 |
"step": 258
|
| 1816 |
},
|
| 1817 |
{
|
| 1818 |
"epoch": 0.6658097686375322,
|
| 1819 |
"grad_norm": 0.0,
|
| 1820 |
"learning_rate": 5.3159155930021e-06,
|
| 1821 |
+
"loss": 1.5327,
|
| 1822 |
"step": 259
|
| 1823 |
},
|
| 1824 |
{
|
| 1825 |
"epoch": 0.6683804627249358,
|
| 1826 |
"grad_norm": 0.0,
|
| 1827 |
"learning_rate": 5.2424548091244334e-06,
|
| 1828 |
+
"loss": 1.5522,
|
| 1829 |
"step": 260
|
| 1830 |
},
|
| 1831 |
{
|
| 1832 |
"epoch": 0.6709511568123393,
|
| 1833 |
"grad_norm": 0.0,
|
| 1834 |
"learning_rate": 5.169324392851105e-06,
|
| 1835 |
+
"loss": 1.543,
|
| 1836 |
"step": 261
|
| 1837 |
},
|
| 1838 |
{
|
| 1839 |
"epoch": 0.6735218508997429,
|
| 1840 |
"grad_norm": 0.0,
|
| 1841 |
"learning_rate": 5.096529422414571e-06,
|
| 1842 |
+
"loss": 1.5483,
|
| 1843 |
"step": 262
|
| 1844 |
},
|
| 1845 |
{
|
| 1846 |
"epoch": 0.6760925449871465,
|
| 1847 |
"grad_norm": 0.0,
|
| 1848 |
"learning_rate": 5.0240749527536845e-06,
|
| 1849 |
+
"loss": 1.5234,
|
| 1850 |
"step": 263
|
| 1851 |
},
|
| 1852 |
{
|
| 1853 |
"epoch": 0.6786632390745502,
|
| 1854 |
"grad_norm": 0.0,
|
| 1855 |
"learning_rate": 4.951966015162652e-06,
|
| 1856 |
+
"loss": 1.5315,
|
| 1857 |
"step": 264
|
| 1858 |
},
|
| 1859 |
{
|
| 1860 |
"epoch": 0.6812339331619537,
|
| 1861 |
"grad_norm": 0.0,
|
| 1862 |
"learning_rate": 4.880207616941663e-06,
|
| 1863 |
+
"loss": 1.5193,
|
| 1864 |
"step": 265
|
| 1865 |
},
|
| 1866 |
{
|
| 1867 |
"epoch": 0.6838046272493573,
|
| 1868 |
"grad_norm": 0.0,
|
| 1869 |
"learning_rate": 4.8088047410492e-06,
|
| 1870 |
+
"loss": 1.5586,
|
| 1871 |
"step": 266
|
| 1872 |
},
|
| 1873 |
{
|
| 1874 |
"epoch": 0.6863753213367609,
|
| 1875 |
"grad_norm": 0.0,
|
| 1876 |
"learning_rate": 4.737762345755975e-06,
|
| 1877 |
+
"loss": 1.481,
|
| 1878 |
"step": 267
|
| 1879 |
},
|
| 1880 |
{
|
| 1881 |
"epoch": 0.6889460154241646,
|
| 1882 |
"grad_norm": 0.0,
|
| 1883 |
"learning_rate": 4.667085364300678e-06,
|
| 1884 |
+
"loss": 1.5869,
|
| 1885 |
"step": 268
|
| 1886 |
},
|
| 1887 |
{
|
| 1888 |
"epoch": 0.6915167095115681,
|
| 1889 |
"grad_norm": 0.0,
|
| 1890 |
"learning_rate": 4.596778704547359e-06,
|
| 1891 |
+
"loss": 1.5366,
|
| 1892 |
"step": 269
|
| 1893 |
},
|
| 1894 |
{
|
| 1895 |
"epoch": 0.6940874035989717,
|
| 1896 |
"grad_norm": 0.0,
|
| 1897 |
"learning_rate": 4.526847248644652e-06,
|
| 1898 |
+
"loss": 1.5007,
|
| 1899 |
"step": 270
|
| 1900 |
},
|
| 1901 |
{
|
| 1902 |
"epoch": 0.6966580976863753,
|
| 1903 |
"grad_norm": 0.0,
|
| 1904 |
"learning_rate": 4.457295852686746e-06,
|
| 1905 |
+
"loss": 1.5352,
|
| 1906 |
"step": 271
|
| 1907 |
},
|
| 1908 |
{
|
| 1909 |
"epoch": 0.699228791773779,
|
| 1910 |
"grad_norm": 0.0,
|
| 1911 |
"learning_rate": 4.388129346376177e-06,
|
| 1912 |
+
"loss": 1.5447,
|
| 1913 |
"step": 272
|
| 1914 |
},
|
| 1915 |
{
|
| 1916 |
"epoch": 0.7017994858611826,
|
| 1917 |
"grad_norm": 0.0,
|
| 1918 |
"learning_rate": 4.319352532688444e-06,
|
| 1919 |
+
"loss": 1.5701,
|
| 1920 |
"step": 273
|
| 1921 |
},
|
| 1922 |
{
|
| 1923 |
"epoch": 0.7043701799485861,
|
| 1924 |
"grad_norm": 0.0,
|
| 1925 |
"learning_rate": 4.250970187538484e-06,
|
| 1926 |
+
"loss": 1.5,
|
| 1927 |
"step": 274
|
| 1928 |
},
|
| 1929 |
{
|
| 1930 |
"epoch": 0.7069408740359897,
|
| 1931 |
"grad_norm": 0.0,
|
| 1932 |
"learning_rate": 4.182987059449056e-06,
|
| 1933 |
+
"loss": 1.5513,
|
| 1934 |
"step": 275
|
| 1935 |
},
|
| 1936 |
{
|
| 1937 |
"epoch": 0.7095115681233933,
|
| 1938 |
"grad_norm": 0.0,
|
| 1939 |
"learning_rate": 4.115407869220948e-06,
|
| 1940 |
+
"loss": 1.5007,
|
| 1941 |
"step": 276
|
| 1942 |
},
|
| 1943 |
{
|
| 1944 |
"epoch": 0.712082262210797,
|
| 1945 |
"grad_norm": 0.0,
|
| 1946 |
"learning_rate": 4.048237309605216e-06,
|
| 1947 |
+
"loss": 1.5398,
|
| 1948 |
"step": 277
|
| 1949 |
},
|
| 1950 |
{
|
| 1951 |
"epoch": 0.7146529562982005,
|
| 1952 |
"grad_norm": 0.0,
|
| 1953 |
"learning_rate": 3.981480044977284e-06,
|
| 1954 |
+
"loss": 1.5476,
|
| 1955 |
"step": 278
|
| 1956 |
},
|
| 1957 |
{
|
| 1958 |
"epoch": 0.7172236503856041,
|
| 1959 |
"grad_norm": 0.0,
|
| 1960 |
"learning_rate": 3.915140711013044e-06,
|
| 1961 |
+
"loss": 1.5015,
|
| 1962 |
"step": 279
|
| 1963 |
},
|
| 1964 |
{
|
| 1965 |
"epoch": 0.7197943444730077,
|
| 1966 |
"grad_norm": 0.0,
|
| 1967 |
"learning_rate": 3.849223914366981e-06,
|
| 1968 |
+
"loss": 1.5405,
|
| 1969 |
"step": 280
|
| 1970 |
},
|
| 1971 |
{
|
| 1972 |
"epoch": 0.7223650385604113,
|
| 1973 |
"grad_norm": 0.0,
|
| 1974 |
"learning_rate": 3.7837342323522454e-06,
|
| 1975 |
+
"loss": 1.5413,
|
| 1976 |
"step": 281
|
| 1977 |
},
|
| 1978 |
{
|
| 1979 |
"epoch": 0.7249357326478149,
|
| 1980 |
"grad_norm": 0.0,
|
| 1981 |
"learning_rate": 3.7186762126228227e-06,
|
| 1982 |
+
"loss": 1.5874,
|
| 1983 |
"step": 282
|
| 1984 |
},
|
| 1985 |
{
|
| 1986 |
"epoch": 0.7275064267352185,
|
| 1987 |
"grad_norm": 0.0,
|
| 1988 |
"learning_rate": 3.654054372857738e-06,
|
| 1989 |
+
"loss": 1.5122,
|
| 1990 |
"step": 283
|
| 1991 |
},
|
| 1992 |
{
|
| 1993 |
"epoch": 0.7300771208226221,
|
| 1994 |
"grad_norm": 0.0,
|
| 1995 |
"learning_rate": 3.5898732004473523e-06,
|
| 1996 |
+
"loss": 1.55,
|
| 1997 |
"step": 284
|
| 1998 |
},
|
| 1999 |
{
|
| 2000 |
"epoch": 0.7326478149100257,
|
| 2001 |
"grad_norm": 0.0,
|
| 2002 |
"learning_rate": 3.5261371521817247e-06,
|
| 2003 |
+
"loss": 1.5337,
|
| 2004 |
"step": 285
|
| 2005 |
},
|
| 2006 |
{
|
| 2007 |
"epoch": 0.7352185089974294,
|
| 2008 |
"grad_norm": 0.0,
|
| 2009 |
"learning_rate": 3.462850653941171e-06,
|
| 2010 |
+
"loss": 1.5159,
|
| 2011 |
"step": 286
|
| 2012 |
},
|
| 2013 |
{
|
| 2014 |
"epoch": 0.7377892030848329,
|
| 2015 |
"grad_norm": 0.0,
|
| 2016 |
"learning_rate": 3.4000181003889e-06,
|
| 2017 |
+
"loss": 1.5139,
|
| 2018 |
"step": 287
|
| 2019 |
},
|
| 2020 |
{
|
| 2021 |
"epoch": 0.7403598971722365,
|
| 2022 |
"grad_norm": 0.0,
|
| 2023 |
"learning_rate": 3.337643854665843e-06,
|
| 2024 |
+
"loss": 1.499,
|
| 2025 |
"step": 288
|
| 2026 |
},
|
| 2027 |
{
|
| 2028 |
"epoch": 0.7429305912596401,
|
| 2029 |
"grad_norm": 0.0,
|
| 2030 |
"learning_rate": 3.2757322480876996e-06,
|
| 2031 |
+
"loss": 1.5149,
|
| 2032 |
"step": 289
|
| 2033 |
},
|
| 2034 |
{
|
| 2035 |
"epoch": 0.7455012853470437,
|
| 2036 |
"grad_norm": 0.0,
|
| 2037 |
"learning_rate": 3.2142875798441376e-06,
|
| 2038 |
+
"loss": 1.5098,
|
| 2039 |
"step": 290
|
| 2040 |
},
|
| 2041 |
{
|
| 2042 |
"epoch": 0.7480719794344473,
|
| 2043 |
"grad_norm": 0.0,
|
| 2044 |
"learning_rate": 3.15331411670027e-06,
|
| 2045 |
+
"loss": 1.5217,
|
| 2046 |
"step": 291
|
| 2047 |
},
|
| 2048 |
{
|
| 2049 |
"epoch": 0.7506426735218509,
|
| 2050 |
"grad_norm": 0.0,
|
| 2051 |
"learning_rate": 3.092816092700366e-06,
|
| 2052 |
+
"loss": 1.5017,
|
| 2053 |
"step": 292
|
| 2054 |
},
|
| 2055 |
{
|
| 2056 |
"epoch": 0.7532133676092545,
|
| 2057 |
"grad_norm": 0.0,
|
| 2058 |
"learning_rate": 3.032797708873828e-06,
|
| 2059 |
+
"loss": 1.5398,
|
| 2060 |
"step": 293
|
| 2061 |
},
|
| 2062 |
{
|
| 2063 |
"epoch": 0.7557840616966581,
|
| 2064 |
"grad_norm": 0.0,
|
| 2065 |
"learning_rate": 2.97326313294349e-06,
|
| 2066 |
+
"loss": 1.4983,
|
| 2067 |
"step": 294
|
| 2068 |
},
|
| 2069 |
{
|
| 2070 |
"epoch": 0.7583547557840618,
|
| 2071 |
"grad_norm": 0.0,
|
| 2072 |
"learning_rate": 2.914216499036178e-06,
|
| 2073 |
+
"loss": 1.5271,
|
| 2074 |
"step": 295
|
| 2075 |
},
|
| 2076 |
{
|
| 2077 |
"epoch": 0.7609254498714653,
|
| 2078 |
"grad_norm": 0.0,
|
| 2079 |
"learning_rate": 2.855661907395655e-06,
|
| 2080 |
+
"loss": 1.5286,
|
| 2081 |
"step": 296
|
| 2082 |
},
|
| 2083 |
{
|
| 2084 |
"epoch": 0.7634961439588689,
|
| 2085 |
"grad_norm": 0.0,
|
| 2086 |
"learning_rate": 2.7976034240978834e-06,
|
| 2087 |
+
"loss": 1.4954,
|
| 2088 |
"step": 297
|
| 2089 |
},
|
| 2090 |
{
|
| 2091 |
"epoch": 0.7660668380462725,
|
| 2092 |
"grad_norm": 0.0,
|
| 2093 |
"learning_rate": 2.740045080768694e-06,
|
| 2094 |
+
"loss": 1.4653,
|
| 2095 |
"step": 298
|
| 2096 |
},
|
| 2097 |
{
|
| 2098 |
"epoch": 0.7686375321336761,
|
| 2099 |
"grad_norm": 0.0,
|
| 2100 |
"learning_rate": 2.6829908743037936e-06,
|
| 2101 |
+
"loss": 1.5271,
|
| 2102 |
"step": 299
|
| 2103 |
},
|
| 2104 |
{
|
| 2105 |
"epoch": 0.7712082262210797,
|
| 2106 |
"grad_norm": 0.0,
|
| 2107 |
"learning_rate": 2.626444766591253e-06,
|
| 2108 |
+
"loss": 1.48,
|
| 2109 |
"step": 300
|
| 2110 |
},
|
| 2111 |
{
|
| 2112 |
"epoch": 0.7737789203084833,
|
| 2113 |
"grad_norm": 0.0,
|
| 2114 |
"learning_rate": 2.570410684236365e-06,
|
| 2115 |
+
"loss": 1.5093,
|
| 2116 |
"step": 301
|
| 2117 |
},
|
| 2118 |
{
|
| 2119 |
"epoch": 0.7763496143958869,
|
| 2120 |
"grad_norm": 0.0,
|
| 2121 |
"learning_rate": 2.514892518288988e-06,
|
| 2122 |
+
"loss": 1.531,
|
| 2123 |
"step": 302
|
| 2124 |
},
|
| 2125 |
{
|
| 2126 |
"epoch": 0.7789203084832905,
|
| 2127 |
"grad_norm": 0.0,
|
| 2128 |
"learning_rate": 2.4598941239733555e-06,
|
| 2129 |
+
"loss": 1.4795,
|
| 2130 |
"step": 303
|
| 2131 |
},
|
| 2132 |
{
|
| 2133 |
"epoch": 0.781491002570694,
|
| 2134 |
"grad_norm": 0.0,
|
| 2135 |
"learning_rate": 2.4054193204203457e-06,
|
| 2136 |
+
"loss": 1.5056,
|
| 2137 |
"step": 304
|
| 2138 |
},
|
| 2139 |
{
|
| 2140 |
"epoch": 0.7840616966580977,
|
| 2141 |
"grad_norm": 0.0,
|
| 2142 |
"learning_rate": 2.3514718904022993e-06,
|
| 2143 |
+
"loss": 1.4841,
|
| 2144 |
"step": 305
|
| 2145 |
},
|
| 2146 |
{
|
| 2147 |
"epoch": 0.7866323907455013,
|
| 2148 |
"grad_norm": 0.0,
|
| 2149 |
"learning_rate": 2.2980555800703273e-06,
|
| 2150 |
+
"loss": 1.5337,
|
| 2151 |
"step": 306
|
| 2152 |
},
|
| 2153 |
{
|
| 2154 |
"epoch": 0.7892030848329049,
|
| 2155 |
"grad_norm": 0.0,
|
| 2156 |
"learning_rate": 2.2451740986941905e-06,
|
| 2157 |
+
"loss": 1.5212,
|
| 2158 |
"step": 307
|
| 2159 |
},
|
| 2160 |
{
|
| 2161 |
"epoch": 0.7917737789203085,
|
| 2162 |
"grad_norm": 0.0,
|
| 2163 |
"learning_rate": 2.1928311184046967e-06,
|
| 2164 |
+
"loss": 1.5308,
|
| 2165 |
"step": 308
|
| 2166 |
},
|
| 2167 |
{
|
| 2168 |
"epoch": 0.794344473007712,
|
| 2169 |
"grad_norm": 0.0,
|
| 2170 |
"learning_rate": 2.1410302739387424e-06,
|
| 2171 |
+
"loss": 1.5159,
|
| 2172 |
"step": 309
|
| 2173 |
},
|
| 2174 |
{
|
| 2175 |
"epoch": 0.7969151670951157,
|
| 2176 |
"grad_norm": 0.0,
|
| 2177 |
"learning_rate": 2.0897751623868833e-06,
|
| 2178 |
+
"loss": 1.5349,
|
| 2179 |
"step": 310
|
| 2180 |
},
|
| 2181 |
{
|
| 2182 |
"epoch": 0.7994858611825193,
|
| 2183 |
"grad_norm": 0.0,
|
| 2184 |
"learning_rate": 2.0390693429435626e-06,
|
| 2185 |
+
"loss": 1.5029,
|
| 2186 |
"step": 311
|
| 2187 |
},
|
| 2188 |
{
|
| 2189 |
"epoch": 0.8020565552699229,
|
| 2190 |
"grad_norm": 0.0,
|
| 2191 |
"learning_rate": 1.9889163366599607e-06,
|
| 2192 |
+
"loss": 1.519,
|
| 2193 |
"step": 312
|
| 2194 |
},
|
| 2195 |
{
|
| 2196 |
"epoch": 0.8046272493573264,
|
| 2197 |
"grad_norm": 0.0,
|
| 2198 |
"learning_rate": 1.939319626199483e-06,
|
| 2199 |
+
"loss": 1.5054,
|
| 2200 |
"step": 313
|
| 2201 |
},
|
| 2202 |
{
|
| 2203 |
"epoch": 0.8071979434447301,
|
| 2204 |
"grad_norm": 0.0,
|
| 2205 |
"learning_rate": 1.890282655595922e-06,
|
| 2206 |
+
"loss": 1.4736,
|
| 2207 |
"step": 314
|
| 2208 |
},
|
| 2209 |
{
|
| 2210 |
"epoch": 0.8097686375321337,
|
| 2211 |
"grad_norm": 0.0,
|
| 2212 |
"learning_rate": 1.8418088300143044e-06,
|
| 2213 |
+
"loss": 1.5242,
|
| 2214 |
"step": 315
|
| 2215 |
},
|
| 2216 |
{
|
| 2217 |
"epoch": 0.8123393316195373,
|
| 2218 |
"grad_norm": 0.0,
|
| 2219 |
"learning_rate": 1.7939015155144378e-06,
|
| 2220 |
+
"loss": 1.5208,
|
| 2221 |
"step": 316
|
| 2222 |
},
|
| 2223 |
{
|
| 2224 |
"epoch": 0.8149100257069408,
|
| 2225 |
"grad_norm": 0.0,
|
| 2226 |
"learning_rate": 1.7465640388171589e-06,
|
| 2227 |
+
"loss": 1.5332,
|
| 2228 |
"step": 317
|
| 2229 |
},
|
| 2230 |
{
|
| 2231 |
"epoch": 0.8174807197943444,
|
| 2232 |
"grad_norm": 0.0,
|
| 2233 |
"learning_rate": 1.6997996870733268e-06,
|
| 2234 |
+
"loss": 1.4978,
|
| 2235 |
"step": 318
|
| 2236 |
},
|
| 2237 |
{
|
| 2238 |
"epoch": 0.8200514138817481,
|
| 2239 |
"grad_norm": 0.0,
|
| 2240 |
"learning_rate": 1.6536117076355652e-06,
|
| 2241 |
+
"loss": 1.4961,
|
| 2242 |
"step": 319
|
| 2243 |
},
|
| 2244 |
{
|
| 2245 |
"epoch": 0.8226221079691517,
|
| 2246 |
"grad_norm": 0.0,
|
| 2247 |
"learning_rate": 1.6080033078327585e-06,
|
| 2248 |
+
"loss": 1.5559,
|
| 2249 |
"step": 320
|
| 2250 |
},
|
| 2251 |
{
|
| 2252 |
"epoch": 0.8251928020565553,
|
| 2253 |
"grad_norm": 0.0,
|
| 2254 |
"learning_rate": 1.5629776547473397e-06,
|
| 2255 |
+
"loss": 1.5435,
|
| 2256 |
"step": 321
|
| 2257 |
},
|
| 2258 |
{
|
| 2259 |
"epoch": 0.8277634961439588,
|
| 2260 |
"grad_norm": 0.0,
|
| 2261 |
"learning_rate": 1.5185378749953538e-06,
|
| 2262 |
+
"loss": 1.4744,
|
| 2263 |
"step": 322
|
| 2264 |
},
|
| 2265 |
{
|
| 2266 |
"epoch": 0.8303341902313625,
|
| 2267 |
"grad_norm": 0.0,
|
| 2268 |
"learning_rate": 1.4746870545093528e-06,
|
| 2269 |
+
"loss": 1.4885,
|
| 2270 |
"step": 323
|
| 2271 |
},
|
| 2272 |
{
|
| 2273 |
"epoch": 0.8329048843187661,
|
| 2274 |
"grad_norm": 0.0,
|
| 2275 |
"learning_rate": 1.4314282383241097e-06,
|
| 2276 |
+
"loss": 1.5088,
|
| 2277 |
"step": 324
|
| 2278 |
},
|
| 2279 |
{
|
| 2280 |
"epoch": 0.8354755784061697,
|
| 2281 |
"grad_norm": 0.0,
|
| 2282 |
"learning_rate": 1.388764430365147e-06,
|
| 2283 |
+
"loss": 1.4971,
|
| 2284 |
"step": 325
|
| 2285 |
},
|
| 2286 |
{
|
| 2287 |
"epoch": 0.8380462724935732,
|
| 2288 |
"grad_norm": 0.0,
|
| 2289 |
"learning_rate": 1.3466985932401743e-06,
|
| 2290 |
+
"loss": 1.5269,
|
| 2291 |
"step": 326
|
| 2292 |
},
|
| 2293 |
{
|
| 2294 |
"epoch": 0.8406169665809768,
|
| 2295 |
"grad_norm": 0.0,
|
| 2296 |
"learning_rate": 1.3052336480333372e-06,
|
| 2297 |
+
"loss": 1.5088,
|
| 2298 |
"step": 327
|
| 2299 |
},
|
| 2300 |
{
|
| 2301 |
"epoch": 0.8431876606683805,
|
| 2302 |
"grad_norm": 0.0,
|
| 2303 |
"learning_rate": 1.2643724741023845e-06,
|
| 2304 |
+
"loss": 1.5046,
|
| 2305 |
"step": 328
|
| 2306 |
},
|
| 2307 |
{
|
| 2308 |
"epoch": 0.8457583547557841,
|
| 2309 |
"grad_norm": 0.0,
|
| 2310 |
"learning_rate": 1.2241179088787192e-06,
|
| 2311 |
+
"loss": 1.5217,
|
| 2312 |
"step": 329
|
| 2313 |
},
|
| 2314 |
{
|
| 2315 |
"epoch": 0.8483290488431876,
|
| 2316 |
"grad_norm": 0.0,
|
| 2317 |
"learning_rate": 1.1844727476703776e-06,
|
| 2318 |
+
"loss": 1.4951,
|
| 2319 |
"step": 330
|
| 2320 |
},
|
| 2321 |
{
|
| 2322 |
"epoch": 0.8508997429305912,
|
| 2323 |
"grad_norm": 0.0,
|
| 2324 |
"learning_rate": 1.1454397434679022e-06,
|
| 2325 |
+
"loss": 1.4941,
|
| 2326 |
"step": 331
|
| 2327 |
},
|
| 2328 |
{
|
| 2329 |
"epoch": 0.8534704370179949,
|
| 2330 |
"grad_norm": 0.0,
|
| 2331 |
"learning_rate": 1.1070216067531825e-06,
|
| 2332 |
+
"loss": 1.5122,
|
| 2333 |
"step": 332
|
| 2334 |
},
|
| 2335 |
{
|
| 2336 |
"epoch": 0.8560411311053985,
|
| 2337 |
"grad_norm": 0.0,
|
| 2338 |
"learning_rate": 1.0692210053112451e-06,
|
| 2339 |
+
"loss": 1.5427,
|
| 2340 |
"step": 333
|
| 2341 |
},
|
| 2342 |
{
|
| 2343 |
"epoch": 0.8586118251928021,
|
| 2344 |
"grad_norm": 0.0,
|
| 2345 |
"learning_rate": 1.032040564044975e-06,
|
| 2346 |
+
"loss": 1.5278,
|
| 2347 |
"step": 334
|
| 2348 |
},
|
| 2349 |
{
|
| 2350 |
"epoch": 0.8611825192802056,
|
| 2351 |
"grad_norm": 0.0,
|
| 2352 |
"learning_rate": 9.954828647928727e-07,
|
| 2353 |
+
"loss": 1.4768,
|
| 2354 |
"step": 335
|
| 2355 |
},
|
| 2356 |
{
|
| 2357 |
"epoch": 0.8637532133676092,
|
| 2358 |
"grad_norm": 0.0,
|
| 2359 |
"learning_rate": 9.595504461497441e-07,
|
| 2360 |
+
"loss": 1.5066,
|
| 2361 |
"step": 336
|
| 2362 |
},
|
| 2363 |
{
|
| 2364 |
"epoch": 0.8663239074550129,
|
| 2365 |
"grad_norm": 0.0,
|
| 2366 |
"learning_rate": 9.242458032904311e-07,
|
| 2367 |
+
"loss": 1.4871,
|
| 2368 |
"step": 337
|
| 2369 |
},
|
| 2370 |
{
|
| 2371 |
"epoch": 0.8688946015424165,
|
| 2372 |
"grad_norm": 0.0,
|
| 2373 |
"learning_rate": 8.895713877965373e-07,
|
| 2374 |
+
"loss": 1.5212,
|
| 2375 |
"step": 338
|
| 2376 |
},
|
| 2377 |
{
|
| 2378 |
"epoch": 0.87146529562982,
|
| 2379 |
"grad_norm": 0.0,
|
| 2380 |
"learning_rate": 8.555296074861996e-07,
|
| 2381 |
+
"loss": 1.4919,
|
| 2382 |
"step": 339
|
| 2383 |
},
|
| 2384 |
{
|
| 2385 |
"epoch": 0.8740359897172236,
|
| 2386 |
"grad_norm": 0.0,
|
| 2387 |
"learning_rate": 8.22122826246875e-07,
|
| 2388 |
+
"loss": 1.5476,
|
| 2389 |
"step": 340
|
| 2390 |
},
|
| 2391 |
{
|
| 2392 |
"epoch": 0.8766066838046273,
|
| 2393 |
"grad_norm": 0.0,
|
| 2394 |
"learning_rate": 7.89353363871197e-07,
|
| 2395 |
+
"loss": 1.5142,
|
| 2396 |
"step": 341
|
| 2397 |
},
|
| 2398 |
{
|
| 2399 |
"epoch": 0.8791773778920309,
|
| 2400 |
"grad_norm": 0.0,
|
| 2401 |
"learning_rate": 7.572234958958846e-07,
|
| 2402 |
+
"loss": 1.5332,
|
| 2403 |
"step": 342
|
| 2404 |
},
|
| 2405 |
{
|
| 2406 |
"epoch": 0.8817480719794345,
|
| 2407 |
"grad_norm": 0.0,
|
| 2408 |
"learning_rate": 7.2573545344373e-07,
|
| 2409 |
+
"loss": 1.4924,
|
| 2410 |
"step": 343
|
| 2411 |
},
|
| 2412 |
{
|
| 2413 |
"epoch": 0.884318766066838,
|
| 2414 |
"grad_norm": 0.0,
|
| 2415 |
"learning_rate": 6.948914230686688e-07,
|
| 2416 |
+
"loss": 1.5181,
|
| 2417 |
"step": 344
|
| 2418 |
},
|
| 2419 |
{
|
| 2420 |
"epoch": 0.8868894601542416,
|
| 2421 |
"grad_norm": 0.0,
|
| 2422 |
"learning_rate": 6.646935466039373e-07,
|
| 2423 |
+
"loss": 1.5137,
|
| 2424 |
"step": 345
|
| 2425 |
},
|
| 2426 |
{
|
| 2427 |
"epoch": 0.8894601542416453,
|
| 2428 |
"grad_norm": 0.0,
|
| 2429 |
"learning_rate": 6.351439210133492e-07,
|
| 2430 |
+
"loss": 1.5056,
|
| 2431 |
"step": 346
|
| 2432 |
},
|
| 2433 |
{
|
| 2434 |
"epoch": 0.8920308483290489,
|
| 2435 |
"grad_norm": 0.0,
|
| 2436 |
"learning_rate": 6.062445982456777e-07,
|
| 2437 |
+
"loss": 1.4688,
|
| 2438 |
"step": 347
|
| 2439 |
},
|
| 2440 |
{
|
| 2441 |
"epoch": 0.8946015424164524,
|
| 2442 |
"grad_norm": 0.0,
|
| 2443 |
"learning_rate": 5.77997585092166e-07,
|
| 2444 |
+
"loss": 1.5146,
|
| 2445 |
"step": 348
|
| 2446 |
},
|
| 2447 |
{
|
| 2448 |
"epoch": 0.897172236503856,
|
| 2449 |
"grad_norm": 0.0,
|
| 2450 |
"learning_rate": 5.504048430471753e-07,
|
| 2451 |
+
"loss": 1.4695,
|
| 2452 |
"step": 349
|
| 2453 |
},
|
| 2454 |
{
|
| 2455 |
"epoch": 0.8997429305912596,
|
| 2456 |
"grad_norm": 0.0,
|
| 2457 |
"learning_rate": 5.234682881719766e-07,
|
| 2458 |
+
"loss": 1.5129,
|
| 2459 |
"step": 350
|
| 2460 |
},
|
| 2461 |
{
|
| 2462 |
"epoch": 0.9023136246786633,
|
| 2463 |
"grad_norm": 0.0,
|
| 2464 |
"learning_rate": 4.971897909616985e-07,
|
| 2465 |
+
"loss": 1.5061,
|
| 2466 |
"step": 351
|
| 2467 |
},
|
| 2468 |
{
|
| 2469 |
"epoch": 0.9048843187660668,
|
| 2470 |
"grad_norm": 0.0,
|
| 2471 |
"learning_rate": 4.715711762154362e-07,
|
| 2472 |
+
"loss": 1.4722,
|
| 2473 |
"step": 352
|
| 2474 |
},
|
| 2475 |
{
|
| 2476 |
"epoch": 0.9074550128534704,
|
| 2477 |
"grad_norm": 0.0,
|
| 2478 |
"learning_rate": 4.4661422290954495e-07,
|
| 2479 |
+
"loss": 1.5056,
|
| 2480 |
"step": 353
|
| 2481 |
},
|
| 2482 |
{
|
| 2483 |
"epoch": 0.910025706940874,
|
| 2484 |
"grad_norm": 0.0,
|
| 2485 |
"learning_rate": 4.2232066407409067e-07,
|
| 2486 |
+
"loss": 1.5017,
|
| 2487 |
"step": 354
|
| 2488 |
},
|
| 2489 |
{
|
| 2490 |
"epoch": 0.9125964010282777,
|
| 2491 |
"grad_norm": 0.0,
|
| 2492 |
"learning_rate": 3.986921866725202e-07,
|
| 2493 |
+
"loss": 1.5393,
|
| 2494 |
"step": 355
|
| 2495 |
},
|
| 2496 |
{
|
| 2497 |
"epoch": 0.9151670951156813,
|
| 2498 |
"grad_norm": 0.0,
|
| 2499 |
"learning_rate": 3.7573043148451673e-07,
|
| 2500 |
+
"loss": 1.5034,
|
| 2501 |
"step": 356
|
| 2502 |
},
|
| 2503 |
{
|
| 2504 |
"epoch": 0.9177377892030848,
|
| 2505 |
"grad_norm": 0.0,
|
| 2506 |
"learning_rate": 3.5343699299205003e-07,
|
| 2507 |
+
"loss": 1.5139,
|
| 2508 |
"step": 357
|
| 2509 |
},
|
| 2510 |
{
|
| 2511 |
"epoch": 0.9203084832904884,
|
| 2512 |
"grad_norm": 0.0,
|
| 2513 |
"learning_rate": 3.3181341926867283e-07,
|
| 2514 |
+
"loss": 1.4788,
|
| 2515 |
"step": 358
|
| 2516 |
},
|
| 2517 |
{
|
| 2518 |
"epoch": 0.922879177377892,
|
| 2519 |
"grad_norm": 0.0,
|
| 2520 |
"learning_rate": 3.1086121187200667e-07,
|
| 2521 |
+
"loss": 1.4746,
|
| 2522 |
"step": 359
|
| 2523 |
},
|
| 2524 |
{
|
| 2525 |
"epoch": 0.9254498714652957,
|
| 2526 |
"grad_norm": 0.0,
|
| 2527 |
"learning_rate": 2.905818257394799e-07,
|
| 2528 |
+
"loss": 1.5112,
|
| 2529 |
"step": 360
|
| 2530 |
},
|
| 2531 |
{
|
| 2532 |
"epoch": 0.9280205655526992,
|
| 2533 |
"grad_norm": 0.0,
|
| 2534 |
"learning_rate": 2.7097666908729283e-07,
|
| 2535 |
+
"loss": 1.5071,
|
| 2536 |
"step": 361
|
| 2537 |
},
|
| 2538 |
{
|
| 2539 |
"epoch": 0.9305912596401028,
|
| 2540 |
"grad_norm": 0.0,
|
| 2541 |
"learning_rate": 2.520471033126326e-07,
|
| 2542 |
+
"loss": 1.4773,
|
| 2543 |
"step": 362
|
| 2544 |
},
|
| 2545 |
{
|
| 2546 |
"epoch": 0.9331619537275064,
|
| 2547 |
"grad_norm": 0.0,
|
| 2548 |
"learning_rate": 2.3379444289913344e-07,
|
| 2549 |
+
"loss": 1.5146,
|
| 2550 |
"step": 363
|
| 2551 |
},
|
| 2552 |
{
|
| 2553 |
"epoch": 0.9357326478149101,
|
| 2554 |
"grad_norm": 0.0,
|
| 2555 |
"learning_rate": 2.1621995532559947e-07,
|
| 2556 |
+
"loss": 1.4978,
|
| 2557 |
"step": 364
|
| 2558 |
},
|
| 2559 |
{
|
| 2560 |
"epoch": 0.9383033419023136,
|
| 2561 |
"grad_norm": 0.0,
|
| 2562 |
"learning_rate": 1.9932486097799408e-07,
|
| 2563 |
+
"loss": 1.5183,
|
| 2564 |
"step": 365
|
| 2565 |
},
|
| 2566 |
{
|
| 2567 |
"epoch": 0.9408740359897172,
|
| 2568 |
"grad_norm": 0.0,
|
| 2569 |
"learning_rate": 1.8311033306468552e-07,
|
| 2570 |
+
"loss": 1.4761,
|
| 2571 |
"step": 366
|
| 2572 |
},
|
| 2573 |
{
|
| 2574 |
"epoch": 0.9434447300771208,
|
| 2575 |
"grad_norm": 0.0,
|
| 2576 |
"learning_rate": 1.6757749753498865e-07,
|
| 2577 |
+
"loss": 1.509,
|
| 2578 |
"step": 367
|
| 2579 |
},
|
| 2580 |
{
|
| 2581 |
"epoch": 0.9460154241645244,
|
| 2582 |
"grad_norm": 0.0,
|
| 2583 |
"learning_rate": 1.5272743300097316e-07,
|
| 2584 |
+
"loss": 1.5095,
|
| 2585 |
"step": 368
|
| 2586 |
},
|
| 2587 |
{
|
| 2588 |
"epoch": 0.9485861182519281,
|
| 2589 |
"grad_norm": 0.0,
|
| 2590 |
"learning_rate": 1.3856117066256225e-07,
|
| 2591 |
+
"loss": 1.5361,
|
| 2592 |
"step": 369
|
| 2593 |
},
|
| 2594 |
{
|
| 2595 |
"epoch": 0.9511568123393316,
|
| 2596 |
"grad_norm": 0.0,
|
| 2597 |
"learning_rate": 1.2507969423593225e-07,
|
| 2598 |
+
"loss": 1.5051,
|
| 2599 |
"step": 370
|
| 2600 |
},
|
| 2601 |
{
|
| 2602 |
"epoch": 0.9537275064267352,
|
| 2603 |
"grad_norm": 0.0,
|
| 2604 |
"learning_rate": 1.1228393988519381e-07,
|
| 2605 |
+
"loss": 1.5532,
|
| 2606 |
"step": 371
|
| 2607 |
},
|
| 2608 |
{
|
| 2609 |
"epoch": 0.9562982005141388,
|
| 2610 |
"grad_norm": 0.0,
|
| 2611 |
"learning_rate": 1.0017479615738957e-07,
|
| 2612 |
+
"loss": 1.553,
|
| 2613 |
"step": 372
|
| 2614 |
},
|
| 2615 |
{
|
| 2616 |
"epoch": 0.9588688946015425,
|
| 2617 |
"grad_norm": 0.0,
|
| 2618 |
"learning_rate": 8.875310392079118e-08,
|
| 2619 |
+
"loss": 1.5125,
|
| 2620 |
"step": 373
|
| 2621 |
},
|
| 2622 |
{
|
| 2623 |
"epoch": 0.961439588688946,
|
| 2624 |
"grad_norm": 0.0,
|
| 2625 |
"learning_rate": 7.801965630651165e-08,
|
| 2626 |
+
"loss": 1.4321,
|
| 2627 |
"step": 374
|
| 2628 |
},
|
| 2629 |
{
|
| 2630 |
"epoch": 0.9640102827763496,
|
| 2631 |
"grad_norm": 0.0,
|
| 2632 |
"learning_rate": 6.797519865342161e-08,
|
| 2633 |
+
"loss": 1.5005,
|
| 2634 |
"step": 375
|
| 2635 |
},
|
| 2636 |
{
|
| 2637 |
"epoch": 0.9665809768637532,
|
| 2638 |
"grad_norm": 0.0,
|
| 2639 |
"learning_rate": 5.862042845640403e-08,
|
| 2640 |
+
"loss": 1.4973,
|
| 2641 |
"step": 376
|
| 2642 |
},
|
| 2643 |
{
|
| 2644 |
"epoch": 0.9691516709511568,
|
| 2645 |
"grad_norm": 0.0,
|
| 2646 |
"learning_rate": 4.9955995317908514e-08,
|
| 2647 |
+
"loss": 1.5449,
|
| 2648 |
"step": 377
|
| 2649 |
},
|
| 2650 |
{
|
| 2651 |
"epoch": 0.9717223650385605,
|
| 2652 |
"grad_norm": 0.0,
|
| 2653 |
"learning_rate": 4.198250090284961e-08,
|
| 2654 |
+
"loss": 1.4795,
|
| 2655 |
"step": 378
|
| 2656 |
},
|
| 2657 |
{
|
| 2658 |
"epoch": 0.974293059125964,
|
| 2659 |
"grad_norm": 0.0,
|
| 2660 |
"learning_rate": 3.47004988968247e-08,
|
| 2661 |
+
"loss": 1.5508,
|
| 2662 |
"step": 379
|
| 2663 |
},
|
| 2664 |
{
|
| 2665 |
"epoch": 0.9768637532133676,
|
| 2666 |
"grad_norm": 0.0,
|
| 2667 |
"learning_rate": 2.8110494967664713e-08,
|
| 2668 |
+
"loss": 1.5095,
|
| 2669 |
"step": 380
|
| 2670 |
},
|
| 2671 |
{
|
| 2672 |
"epoch": 0.9794344473007712,
|
| 2673 |
"grad_norm": 0.0,
|
| 2674 |
"learning_rate": 2.221294673032004e-08,
|
| 2675 |
+
"loss": 1.5146,
|
| 2676 |
"step": 381
|
| 2677 |
},
|
| 2678 |
{
|
| 2679 |
"epoch": 0.9820051413881749,
|
| 2680 |
"grad_norm": 0.0,
|
| 2681 |
"learning_rate": 1.7008263715085904e-08,
|
| 2682 |
+
"loss": 1.5112,
|
| 2683 |
"step": 382
|
| 2684 |
},
|
| 2685 |
{
|
| 2686 |
"epoch": 0.9845758354755784,
|
| 2687 |
"grad_norm": 0.0,
|
| 2688 |
"learning_rate": 1.24968073391607e-08,
|
| 2689 |
+
"loss": 1.5144,
|
| 2690 |
"step": 383
|
| 2691 |
},
|
| 2692 |
{
|
| 2693 |
"epoch": 0.987146529562982,
|
| 2694 |
"grad_norm": 0.0,
|
| 2695 |
"learning_rate": 8.678890881552715e-09,
|
| 2696 |
+
"loss": 1.5459,
|
| 2697 |
"step": 384
|
| 2698 |
},
|
| 2699 |
{
|
| 2700 |
"epoch": 0.9897172236503856,
|
| 2701 |
"grad_norm": 0.0,
|
| 2702 |
"learning_rate": 5.554779461323101e-09,
|
| 2703 |
+
"loss": 1.4885,
|
| 2704 |
"step": 385
|
| 2705 |
},
|
| 2706 |
{
|
| 2707 |
"epoch": 0.9922879177377892,
|
| 2708 |
"grad_norm": 0.0,
|
| 2709 |
"learning_rate": 3.1246900191761463e-09,
|
| 2710 |
+
"loss": 1.4919,
|
| 2711 |
"step": 386
|
| 2712 |
},
|
| 2713 |
{
|
| 2714 |
"epoch": 0.9948586118251928,
|
| 2715 |
"grad_norm": 0.0,
|
| 2716 |
"learning_rate": 1.3887913023946652e-09,
|
| 2717 |
+
"loss": 1.5034,
|
| 2718 |
"step": 387
|
| 2719 |
},
|
| 2720 |
{
|
| 2721 |
"epoch": 0.9974293059125964,
|
| 2722 |
"grad_norm": 0.0,
|
| 2723 |
"learning_rate": 3.4720385312492223e-10,
|
| 2724 |
+
"loss": 1.4812,
|
| 2725 |
"step": 388
|
| 2726 |
},
|
| 2727 |
{
|
| 2728 |
"epoch": 1.0,
|
| 2729 |
"grad_norm": 0.0,
|
| 2730 |
"learning_rate": 0.0,
|
| 2731 |
+
"loss": 1.3696,
|
| 2732 |
"step": 389
|
| 2733 |
},
|
| 2734 |
{
|
| 2735 |
"epoch": 1.0,
|
| 2736 |
"step": 389,
|
| 2737 |
+
"total_flos": 1.4102482311698186e+18,
|
| 2738 |
+
"train_loss": 1.5950692380302056,
|
| 2739 |
+
"train_runtime": 5789.3639,
|
| 2740 |
+
"train_samples_per_second": 17.167,
|
| 2741 |
"train_steps_per_second": 0.067
|
| 2742 |
}
|
| 2743 |
],
|
|
|
|
| 2758 |
"attributes": {}
|
| 2759 |
}
|
| 2760 |
},
|
| 2761 |
+
"total_flos": 1.4102482311698186e+18,
|
| 2762 |
"train_batch_size": 8,
|
| 2763 |
"trial_name": null,
|
| 2764 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6776
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:922bf3c9651d9cec9642c0a40c9644a277ad37d1a89d17d2f00d7318b8038c35
|
| 3 |
size 6776
|