Training in progress, epoch 1, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66bf7ec1906085b9017dec52af7ee1d8cb6d38fa8ca3a526334a82ecc740dcbd
|
| 3 |
+
size 2708729576
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f48d93af9f829aae4cbfe51ee4fffab7189d58377b64586f31b892a561403e60
|
| 3 |
+
size 52499200
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15006
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:838e024e0a852529a2257dc1dc52b7019dd292f53abc55b9fa61835eb1065232
|
| 3 |
size 15006
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bca4dbe650e04bc8012dd3f1938dfb2a637329721abd75c3bd59d28a64007b54
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,1804 +1,180 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 10,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.07808687164470474,
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 9.921875000000001e-05,
|
| 15 |
-
"loss":
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.07808687164470474,
|
| 20 |
-
"eval_runtime": 19.
|
| 21 |
-
"eval_samples_per_second":
|
| 22 |
-
"eval_steps_per_second": 5.
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.15617374328940947,
|
| 27 |
-
"grad_norm":
|
| 28 |
"learning_rate": 9.84375e-05,
|
| 29 |
-
"loss":
|
| 30 |
"step": 20
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.15617374328940947,
|
| 34 |
-
"eval_runtime": 19.
|
| 35 |
-
"eval_samples_per_second":
|
| 36 |
-
"eval_steps_per_second": 5.
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.2342606149341142,
|
| 41 |
-
"grad_norm":
|
| 42 |
"learning_rate": 9.765625e-05,
|
| 43 |
-
"loss":
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.2342606149341142,
|
| 48 |
-
"eval_runtime": 19.
|
| 49 |
-
"eval_samples_per_second":
|
| 50 |
-
"eval_steps_per_second": 5.
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.31234748657881894,
|
| 55 |
-
"grad_norm":
|
| 56 |
"learning_rate": 9.687500000000001e-05,
|
| 57 |
-
"loss":
|
| 58 |
"step": 40
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.31234748657881894,
|
| 62 |
-
"eval_runtime": 19.
|
| 63 |
-
"eval_samples_per_second":
|
| 64 |
-
"eval_steps_per_second": 5.
|
| 65 |
"step": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.3904343582235237,
|
| 69 |
-
"grad_norm":
|
| 70 |
"learning_rate": 9.609375e-05,
|
| 71 |
-
"loss":
|
| 72 |
"step": 50
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.3904343582235237,
|
| 76 |
-
"eval_runtime": 19.
|
| 77 |
-
"eval_samples_per_second":
|
| 78 |
-
"eval_steps_per_second": 5.
|
| 79 |
"step": 50
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.4685212298682284,
|
| 83 |
-
"grad_norm":
|
| 84 |
"learning_rate": 9.53125e-05,
|
| 85 |
-
"loss":
|
| 86 |
"step": 60
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.4685212298682284,
|
| 90 |
-
"eval_runtime": 19.
|
| 91 |
-
"eval_samples_per_second":
|
| 92 |
-
"eval_steps_per_second": 5.
|
| 93 |
"step": 60
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.5466081015129332,
|
| 97 |
-
"grad_norm":
|
| 98 |
"learning_rate": 9.453125000000001e-05,
|
| 99 |
-
"loss":
|
| 100 |
"step": 70
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.5466081015129332,
|
| 104 |
-
"eval_runtime": 19.
|
| 105 |
-
"eval_samples_per_second":
|
| 106 |
-
"eval_steps_per_second": 5.
|
| 107 |
"step": 70
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.6246949731576379,
|
| 111 |
-
"grad_norm":
|
| 112 |
"learning_rate": 9.375e-05,
|
| 113 |
-
"loss":
|
| 114 |
"step": 80
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.6246949731576379,
|
| 118 |
-
"eval_runtime": 19.
|
| 119 |
-
"eval_samples_per_second":
|
| 120 |
-
"eval_steps_per_second": 5.
|
| 121 |
"step": 80
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.7027818448023426,
|
| 125 |
-
"grad_norm":
|
| 126 |
"learning_rate": 9.296875e-05,
|
| 127 |
-
"loss":
|
| 128 |
"step": 90
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.7027818448023426,
|
| 132 |
-
"eval_runtime": 19.
|
| 133 |
-
"eval_samples_per_second":
|
| 134 |
-
"eval_steps_per_second": 5.
|
| 135 |
"step": 90
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.7808687164470474,
|
| 139 |
-
"grad_norm":
|
| 140 |
"learning_rate": 9.21875e-05,
|
| 141 |
-
"loss":
|
| 142 |
"step": 100
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.7808687164470474,
|
| 146 |
-
"eval_runtime": 19.
|
| 147 |
-
"eval_samples_per_second":
|
| 148 |
-
"eval_steps_per_second": 5.
|
| 149 |
"step": 100
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.8589555880917521,
|
| 153 |
-
"grad_norm":
|
| 154 |
"learning_rate": 9.140625e-05,
|
| 155 |
-
"loss":
|
| 156 |
"step": 110
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.8589555880917521,
|
| 160 |
-
"eval_runtime": 19.
|
| 161 |
-
"eval_samples_per_second":
|
| 162 |
-
"eval_steps_per_second": 5.
|
| 163 |
"step": 110
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.9370424597364568,
|
| 167 |
-
"grad_norm":
|
| 168 |
"learning_rate": 9.062500000000001e-05,
|
| 169 |
-
"loss":
|
| 170 |
"step": 120
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.9370424597364568,
|
| 174 |
-
"eval_runtime": 19.
|
| 175 |
-
"eval_samples_per_second":
|
| 176 |
-
"eval_steps_per_second": 5.
|
| 177 |
"step": 120
|
| 178 |
-
},
|
| 179 |
-
{
|
| 180 |
-
"epoch": 1.0078086871644705,
|
| 181 |
-
"grad_norm": 688836.3125,
|
| 182 |
-
"learning_rate": 8.984375e-05,
|
| 183 |
-
"loss": 4461.798,
|
| 184 |
-
"step": 130
|
| 185 |
-
},
|
| 186 |
-
{
|
| 187 |
-
"epoch": 1.0078086871644705,
|
| 188 |
-
"eval_runtime": 19.4608,
|
| 189 |
-
"eval_samples_per_second": 22.198,
|
| 190 |
-
"eval_steps_per_second": 5.55,
|
| 191 |
-
"step": 130
|
| 192 |
-
},
|
| 193 |
-
{
|
| 194 |
-
"epoch": 1.0858955588091752,
|
| 195 |
-
"grad_norm": 564319.6875,
|
| 196 |
-
"learning_rate": 8.90625e-05,
|
| 197 |
-
"loss": 3695.1797,
|
| 198 |
-
"step": 140
|
| 199 |
-
},
|
| 200 |
-
{
|
| 201 |
-
"epoch": 1.0858955588091752,
|
| 202 |
-
"eval_runtime": 19.5702,
|
| 203 |
-
"eval_samples_per_second": 22.074,
|
| 204 |
-
"eval_steps_per_second": 5.519,
|
| 205 |
-
"step": 140
|
| 206 |
-
},
|
| 207 |
-
{
|
| 208 |
-
"epoch": 1.16398243045388,
|
| 209 |
-
"grad_norm": 297341.28125,
|
| 210 |
-
"learning_rate": 8.828125000000001e-05,
|
| 211 |
-
"loss": 3063.5557,
|
| 212 |
-
"step": 150
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 1.16398243045388,
|
| 216 |
-
"eval_runtime": 19.6107,
|
| 217 |
-
"eval_samples_per_second": 22.029,
|
| 218 |
-
"eval_steps_per_second": 5.507,
|
| 219 |
-
"step": 150
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 1.2420693020985847,
|
| 223 |
-
"grad_norm": 294460.375,
|
| 224 |
-
"learning_rate": 8.75e-05,
|
| 225 |
-
"loss": 3861.7359,
|
| 226 |
-
"step": 160
|
| 227 |
-
},
|
| 228 |
-
{
|
| 229 |
-
"epoch": 1.2420693020985847,
|
| 230 |
-
"eval_runtime": 19.6236,
|
| 231 |
-
"eval_samples_per_second": 22.014,
|
| 232 |
-
"eval_steps_per_second": 5.504,
|
| 233 |
-
"step": 160
|
| 234 |
-
},
|
| 235 |
-
{
|
| 236 |
-
"epoch": 1.3201561737432894,
|
| 237 |
-
"grad_norm": 446928.125,
|
| 238 |
-
"learning_rate": 8.671875e-05,
|
| 239 |
-
"loss": 3945.3711,
|
| 240 |
-
"step": 170
|
| 241 |
-
},
|
| 242 |
-
{
|
| 243 |
-
"epoch": 1.3201561737432894,
|
| 244 |
-
"eval_runtime": 19.6104,
|
| 245 |
-
"eval_samples_per_second": 22.029,
|
| 246 |
-
"eval_steps_per_second": 5.507,
|
| 247 |
-
"step": 170
|
| 248 |
-
},
|
| 249 |
-
{
|
| 250 |
-
"epoch": 1.3982430453879942,
|
| 251 |
-
"grad_norm": 477259.65625,
|
| 252 |
-
"learning_rate": 8.593750000000001e-05,
|
| 253 |
-
"loss": 3992.1383,
|
| 254 |
-
"step": 180
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 1.3982430453879942,
|
| 258 |
-
"eval_runtime": 19.6165,
|
| 259 |
-
"eval_samples_per_second": 22.022,
|
| 260 |
-
"eval_steps_per_second": 5.506,
|
| 261 |
-
"step": 180
|
| 262 |
-
},
|
| 263 |
-
{
|
| 264 |
-
"epoch": 1.476329917032699,
|
| 265 |
-
"grad_norm": 416177.8125,
|
| 266 |
-
"learning_rate": 8.515625e-05,
|
| 267 |
-
"loss": 3208.7566,
|
| 268 |
-
"step": 190
|
| 269 |
-
},
|
| 270 |
-
{
|
| 271 |
-
"epoch": 1.476329917032699,
|
| 272 |
-
"eval_runtime": 19.6029,
|
| 273 |
-
"eval_samples_per_second": 22.038,
|
| 274 |
-
"eval_steps_per_second": 5.509,
|
| 275 |
-
"step": 190
|
| 276 |
-
},
|
| 277 |
-
{
|
| 278 |
-
"epoch": 1.5544167886774036,
|
| 279 |
-
"grad_norm": 585219.5,
|
| 280 |
-
"learning_rate": 8.4375e-05,
|
| 281 |
-
"loss": 3544.4445,
|
| 282 |
-
"step": 200
|
| 283 |
-
},
|
| 284 |
-
{
|
| 285 |
-
"epoch": 1.5544167886774036,
|
| 286 |
-
"eval_runtime": 19.6024,
|
| 287 |
-
"eval_samples_per_second": 22.038,
|
| 288 |
-
"eval_steps_per_second": 5.51,
|
| 289 |
-
"step": 200
|
| 290 |
-
},
|
| 291 |
-
{
|
| 292 |
-
"epoch": 1.6325036603221084,
|
| 293 |
-
"grad_norm": 122144.359375,
|
| 294 |
-
"learning_rate": 8.359375000000001e-05,
|
| 295 |
-
"loss": 3495.2234,
|
| 296 |
-
"step": 210
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 1.6325036603221084,
|
| 300 |
-
"eval_runtime": 19.6241,
|
| 301 |
-
"eval_samples_per_second": 22.014,
|
| 302 |
-
"eval_steps_per_second": 5.503,
|
| 303 |
-
"step": 210
|
| 304 |
-
},
|
| 305 |
-
{
|
| 306 |
-
"epoch": 1.710590531966813,
|
| 307 |
-
"grad_norm": 517473.9375,
|
| 308 |
-
"learning_rate": 8.28125e-05,
|
| 309 |
-
"loss": 3106.2523,
|
| 310 |
-
"step": 220
|
| 311 |
-
},
|
| 312 |
-
{
|
| 313 |
-
"epoch": 1.710590531966813,
|
| 314 |
-
"eval_runtime": 19.6219,
|
| 315 |
-
"eval_samples_per_second": 22.016,
|
| 316 |
-
"eval_steps_per_second": 5.504,
|
| 317 |
-
"step": 220
|
| 318 |
-
},
|
| 319 |
-
{
|
| 320 |
-
"epoch": 1.7886774036115178,
|
| 321 |
-
"grad_norm": 279783.34375,
|
| 322 |
-
"learning_rate": 8.203125e-05,
|
| 323 |
-
"loss": 2969.224,
|
| 324 |
-
"step": 230
|
| 325 |
-
},
|
| 326 |
-
{
|
| 327 |
-
"epoch": 1.7886774036115178,
|
| 328 |
-
"eval_runtime": 19.6274,
|
| 329 |
-
"eval_samples_per_second": 22.01,
|
| 330 |
-
"eval_steps_per_second": 5.503,
|
| 331 |
-
"step": 230
|
| 332 |
-
},
|
| 333 |
-
{
|
| 334 |
-
"epoch": 1.8667642752562226,
|
| 335 |
-
"grad_norm": 500301.21875,
|
| 336 |
-
"learning_rate": 8.125000000000001e-05,
|
| 337 |
-
"loss": 2546.8709,
|
| 338 |
-
"step": 240
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 1.8667642752562226,
|
| 342 |
-
"eval_runtime": 19.6224,
|
| 343 |
-
"eval_samples_per_second": 22.016,
|
| 344 |
-
"eval_steps_per_second": 5.504,
|
| 345 |
-
"step": 240
|
| 346 |
-
},
|
| 347 |
-
{
|
| 348 |
-
"epoch": 1.9448511469009273,
|
| 349 |
-
"grad_norm": 420029.9375,
|
| 350 |
-
"learning_rate": 8.046875e-05,
|
| 351 |
-
"loss": 3124.2377,
|
| 352 |
-
"step": 250
|
| 353 |
-
},
|
| 354 |
-
{
|
| 355 |
-
"epoch": 1.9448511469009273,
|
| 356 |
-
"eval_runtime": 19.62,
|
| 357 |
-
"eval_samples_per_second": 22.018,
|
| 358 |
-
"eval_steps_per_second": 5.505,
|
| 359 |
-
"step": 250
|
| 360 |
-
},
|
| 361 |
-
{
|
| 362 |
-
"epoch": 2.015617374328941,
|
| 363 |
-
"grad_norm": 287152.59375,
|
| 364 |
-
"learning_rate": 7.96875e-05,
|
| 365 |
-
"loss": 2647.9721,
|
| 366 |
-
"step": 260
|
| 367 |
-
},
|
| 368 |
-
{
|
| 369 |
-
"epoch": 2.015617374328941,
|
| 370 |
-
"eval_runtime": 19.5051,
|
| 371 |
-
"eval_samples_per_second": 22.148,
|
| 372 |
-
"eval_steps_per_second": 5.537,
|
| 373 |
-
"step": 260
|
| 374 |
-
},
|
| 375 |
-
{
|
| 376 |
-
"epoch": 2.0937042459736457,
|
| 377 |
-
"grad_norm": 227855.6875,
|
| 378 |
-
"learning_rate": 7.890625000000001e-05,
|
| 379 |
-
"loss": 2429.9967,
|
| 380 |
-
"step": 270
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 2.0937042459736457,
|
| 384 |
-
"eval_runtime": 19.5753,
|
| 385 |
-
"eval_samples_per_second": 22.069,
|
| 386 |
-
"eval_steps_per_second": 5.517,
|
| 387 |
-
"step": 270
|
| 388 |
-
},
|
| 389 |
-
{
|
| 390 |
-
"epoch": 2.1717911176183504,
|
| 391 |
-
"grad_norm": 180075.734375,
|
| 392 |
-
"learning_rate": 7.8125e-05,
|
| 393 |
-
"loss": 2353.5262,
|
| 394 |
-
"step": 280
|
| 395 |
-
},
|
| 396 |
-
{
|
| 397 |
-
"epoch": 2.1717911176183504,
|
| 398 |
-
"eval_runtime": 19.6192,
|
| 399 |
-
"eval_samples_per_second": 22.019,
|
| 400 |
-
"eval_steps_per_second": 5.505,
|
| 401 |
-
"step": 280
|
| 402 |
-
},
|
| 403 |
-
{
|
| 404 |
-
"epoch": 2.249877989263055,
|
| 405 |
-
"grad_norm": 374717.40625,
|
| 406 |
-
"learning_rate": 7.734375e-05,
|
| 407 |
-
"loss": 2223.9002,
|
| 408 |
-
"step": 290
|
| 409 |
-
},
|
| 410 |
-
{
|
| 411 |
-
"epoch": 2.249877989263055,
|
| 412 |
-
"eval_runtime": 19.6086,
|
| 413 |
-
"eval_samples_per_second": 22.031,
|
| 414 |
-
"eval_steps_per_second": 5.508,
|
| 415 |
-
"step": 290
|
| 416 |
-
},
|
| 417 |
-
{
|
| 418 |
-
"epoch": 2.32796486090776,
|
| 419 |
-
"grad_norm": 302520.9375,
|
| 420 |
-
"learning_rate": 7.65625e-05,
|
| 421 |
-
"loss": 1899.2279,
|
| 422 |
-
"step": 300
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 2.32796486090776,
|
| 426 |
-
"eval_runtime": 19.6274,
|
| 427 |
-
"eval_samples_per_second": 22.01,
|
| 428 |
-
"eval_steps_per_second": 5.503,
|
| 429 |
-
"step": 300
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 2.4060517325524646,
|
| 433 |
-
"grad_norm": 178822.03125,
|
| 434 |
-
"learning_rate": 7.578125e-05,
|
| 435 |
-
"loss": 1682.3375,
|
| 436 |
-
"step": 310
|
| 437 |
-
},
|
| 438 |
-
{
|
| 439 |
-
"epoch": 2.4060517325524646,
|
| 440 |
-
"eval_runtime": 19.6223,
|
| 441 |
-
"eval_samples_per_second": 22.016,
|
| 442 |
-
"eval_steps_per_second": 5.504,
|
| 443 |
-
"step": 310
|
| 444 |
-
},
|
| 445 |
-
{
|
| 446 |
-
"epoch": 2.4841386041971694,
|
| 447 |
-
"grad_norm": 495346.4375,
|
| 448 |
-
"learning_rate": 7.500000000000001e-05,
|
| 449 |
-
"loss": 2398.752,
|
| 450 |
-
"step": 320
|
| 451 |
-
},
|
| 452 |
-
{
|
| 453 |
-
"epoch": 2.4841386041971694,
|
| 454 |
-
"eval_runtime": 19.5967,
|
| 455 |
-
"eval_samples_per_second": 22.044,
|
| 456 |
-
"eval_steps_per_second": 5.511,
|
| 457 |
-
"step": 320
|
| 458 |
-
},
|
| 459 |
-
{
|
| 460 |
-
"epoch": 2.562225475841874,
|
| 461 |
-
"grad_norm": 389064.15625,
|
| 462 |
-
"learning_rate": 7.421875e-05,
|
| 463 |
-
"loss": 3152.5979,
|
| 464 |
-
"step": 330
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 2.562225475841874,
|
| 468 |
-
"eval_runtime": 19.613,
|
| 469 |
-
"eval_samples_per_second": 22.026,
|
| 470 |
-
"eval_steps_per_second": 5.507,
|
| 471 |
-
"step": 330
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 2.640312347486579,
|
| 475 |
-
"grad_norm": 364833.375,
|
| 476 |
-
"learning_rate": 7.34375e-05,
|
| 477 |
-
"loss": 2780.8895,
|
| 478 |
-
"step": 340
|
| 479 |
-
},
|
| 480 |
-
{
|
| 481 |
-
"epoch": 2.640312347486579,
|
| 482 |
-
"eval_runtime": 19.6058,
|
| 483 |
-
"eval_samples_per_second": 22.034,
|
| 484 |
-
"eval_steps_per_second": 5.509,
|
| 485 |
-
"step": 340
|
| 486 |
-
},
|
| 487 |
-
{
|
| 488 |
-
"epoch": 2.7183992191312836,
|
| 489 |
-
"grad_norm": 250624.234375,
|
| 490 |
-
"learning_rate": 7.265625000000001e-05,
|
| 491 |
-
"loss": 2240.6416,
|
| 492 |
-
"step": 350
|
| 493 |
-
},
|
| 494 |
-
{
|
| 495 |
-
"epoch": 2.7183992191312836,
|
| 496 |
-
"eval_runtime": 19.6306,
|
| 497 |
-
"eval_samples_per_second": 22.006,
|
| 498 |
-
"eval_steps_per_second": 5.502,
|
| 499 |
-
"step": 350
|
| 500 |
-
},
|
| 501 |
-
{
|
| 502 |
-
"epoch": 2.7964860907759883,
|
| 503 |
-
"grad_norm": 355084.1875,
|
| 504 |
-
"learning_rate": 7.1875e-05,
|
| 505 |
-
"loss": 1683.0152,
|
| 506 |
-
"step": 360
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 2.7964860907759883,
|
| 510 |
-
"eval_runtime": 19.6274,
|
| 511 |
-
"eval_samples_per_second": 22.01,
|
| 512 |
-
"eval_steps_per_second": 5.503,
|
| 513 |
-
"step": 360
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 2.874572962420693,
|
| 517 |
-
"grad_norm": 408760.53125,
|
| 518 |
-
"learning_rate": 7.109375e-05,
|
| 519 |
-
"loss": 1970.6135,
|
| 520 |
-
"step": 370
|
| 521 |
-
},
|
| 522 |
-
{
|
| 523 |
-
"epoch": 2.874572962420693,
|
| 524 |
-
"eval_runtime": 19.6037,
|
| 525 |
-
"eval_samples_per_second": 22.037,
|
| 526 |
-
"eval_steps_per_second": 5.509,
|
| 527 |
-
"step": 370
|
| 528 |
-
},
|
| 529 |
-
{
|
| 530 |
-
"epoch": 2.952659834065398,
|
| 531 |
-
"grad_norm": 253820.046875,
|
| 532 |
-
"learning_rate": 7.031250000000001e-05,
|
| 533 |
-
"loss": 2003.0234,
|
| 534 |
-
"step": 380
|
| 535 |
-
},
|
| 536 |
-
{
|
| 537 |
-
"epoch": 2.952659834065398,
|
| 538 |
-
"eval_runtime": 19.6328,
|
| 539 |
-
"eval_samples_per_second": 22.004,
|
| 540 |
-
"eval_steps_per_second": 5.501,
|
| 541 |
-
"step": 380
|
| 542 |
-
},
|
| 543 |
-
{
|
| 544 |
-
"epoch": 3.0234260614934114,
|
| 545 |
-
"grad_norm": 322866.875,
|
| 546 |
-
"learning_rate": 6.953125e-05,
|
| 547 |
-
"loss": 1676.3783,
|
| 548 |
-
"step": 390
|
| 549 |
-
},
|
| 550 |
-
{
|
| 551 |
-
"epoch": 3.0234260614934114,
|
| 552 |
-
"eval_runtime": 19.5185,
|
| 553 |
-
"eval_samples_per_second": 22.133,
|
| 554 |
-
"eval_steps_per_second": 5.533,
|
| 555 |
-
"step": 390
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 3.101512933138116,
|
| 559 |
-
"grad_norm": 238511.0625,
|
| 560 |
-
"learning_rate": 6.875e-05,
|
| 561 |
-
"loss": 1455.4051,
|
| 562 |
-
"step": 400
|
| 563 |
-
},
|
| 564 |
-
{
|
| 565 |
-
"epoch": 3.101512933138116,
|
| 566 |
-
"eval_runtime": 19.6242,
|
| 567 |
-
"eval_samples_per_second": 22.014,
|
| 568 |
-
"eval_steps_per_second": 5.503,
|
| 569 |
-
"step": 400
|
| 570 |
-
},
|
| 571 |
-
{
|
| 572 |
-
"epoch": 3.179599804782821,
|
| 573 |
-
"grad_norm": 91918.703125,
|
| 574 |
-
"learning_rate": 6.796875000000001e-05,
|
| 575 |
-
"loss": 989.8731,
|
| 576 |
-
"step": 410
|
| 577 |
-
},
|
| 578 |
-
{
|
| 579 |
-
"epoch": 3.179599804782821,
|
| 580 |
-
"eval_runtime": 19.636,
|
| 581 |
-
"eval_samples_per_second": 22.0,
|
| 582 |
-
"eval_steps_per_second": 5.5,
|
| 583 |
-
"step": 410
|
| 584 |
-
},
|
| 585 |
-
{
|
| 586 |
-
"epoch": 3.2576866764275256,
|
| 587 |
-
"grad_norm": 362047.0,
|
| 588 |
-
"learning_rate": 6.71875e-05,
|
| 589 |
-
"loss": 1274.4269,
|
| 590 |
-
"step": 420
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 3.2576866764275256,
|
| 594 |
-
"eval_runtime": 19.6293,
|
| 595 |
-
"eval_samples_per_second": 22.008,
|
| 596 |
-
"eval_steps_per_second": 5.502,
|
| 597 |
-
"step": 420
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 3.3357735480722304,
|
| 601 |
-
"grad_norm": 230756.203125,
|
| 602 |
-
"learning_rate": 6.640625e-05,
|
| 603 |
-
"loss": 1462.9975,
|
| 604 |
-
"step": 430
|
| 605 |
-
},
|
| 606 |
-
{
|
| 607 |
-
"epoch": 3.3357735480722304,
|
| 608 |
-
"eval_runtime": 19.6142,
|
| 609 |
-
"eval_samples_per_second": 22.025,
|
| 610 |
-
"eval_steps_per_second": 5.506,
|
| 611 |
-
"step": 430
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 3.413860419716935,
|
| 615 |
-
"grad_norm": 551364.8125,
|
| 616 |
-
"learning_rate": 6.562500000000001e-05,
|
| 617 |
-
"loss": 1467.0002,
|
| 618 |
-
"step": 440
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 3.413860419716935,
|
| 622 |
-
"eval_runtime": 19.6339,
|
| 623 |
-
"eval_samples_per_second": 22.003,
|
| 624 |
-
"eval_steps_per_second": 5.501,
|
| 625 |
-
"step": 440
|
| 626 |
-
},
|
| 627 |
-
{
|
| 628 |
-
"epoch": 3.49194729136164,
|
| 629 |
-
"grad_norm": 238589.6875,
|
| 630 |
-
"learning_rate": 6.484375e-05,
|
| 631 |
-
"loss": 1181.1227,
|
| 632 |
-
"step": 450
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 3.49194729136164,
|
| 636 |
-
"eval_runtime": 19.6121,
|
| 637 |
-
"eval_samples_per_second": 22.027,
|
| 638 |
-
"eval_steps_per_second": 5.507,
|
| 639 |
-
"step": 450
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 3.5700341630063446,
|
| 643 |
-
"grad_norm": 200804.40625,
|
| 644 |
-
"learning_rate": 6.40625e-05,
|
| 645 |
-
"loss": 1450.5544,
|
| 646 |
-
"step": 460
|
| 647 |
-
},
|
| 648 |
-
{
|
| 649 |
-
"epoch": 3.5700341630063446,
|
| 650 |
-
"eval_runtime": 19.6183,
|
| 651 |
-
"eval_samples_per_second": 22.02,
|
| 652 |
-
"eval_steps_per_second": 5.505,
|
| 653 |
-
"step": 460
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 3.6481210346510493,
|
| 657 |
-
"grad_norm": 148718.125,
|
| 658 |
-
"learning_rate": 6.328125e-05,
|
| 659 |
-
"loss": 1220.6021,
|
| 660 |
-
"step": 470
|
| 661 |
-
},
|
| 662 |
-
{
|
| 663 |
-
"epoch": 3.6481210346510493,
|
| 664 |
-
"eval_runtime": 19.5968,
|
| 665 |
-
"eval_samples_per_second": 22.044,
|
| 666 |
-
"eval_steps_per_second": 5.511,
|
| 667 |
-
"step": 470
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 3.726207906295754,
|
| 671 |
-
"grad_norm": 154837.71875,
|
| 672 |
-
"learning_rate": 6.25e-05,
|
| 673 |
-
"loss": 1027.0058,
|
| 674 |
-
"step": 480
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 3.726207906295754,
|
| 678 |
-
"eval_runtime": 19.6342,
|
| 679 |
-
"eval_samples_per_second": 22.002,
|
| 680 |
-
"eval_steps_per_second": 5.501,
|
| 681 |
-
"step": 480
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 3.804294777940459,
|
| 685 |
-
"grad_norm": 388905.5625,
|
| 686 |
-
"learning_rate": 6.171875e-05,
|
| 687 |
-
"loss": 1102.8801,
|
| 688 |
-
"step": 490
|
| 689 |
-
},
|
| 690 |
-
{
|
| 691 |
-
"epoch": 3.804294777940459,
|
| 692 |
-
"eval_runtime": 19.6156,
|
| 693 |
-
"eval_samples_per_second": 22.023,
|
| 694 |
-
"eval_steps_per_second": 5.506,
|
| 695 |
-
"step": 490
|
| 696 |
-
},
|
| 697 |
-
{
|
| 698 |
-
"epoch": 3.8823816495851635,
|
| 699 |
-
"grad_norm": 251027.203125,
|
| 700 |
-
"learning_rate": 6.0937500000000004e-05,
|
| 701 |
-
"loss": 955.4972,
|
| 702 |
-
"step": 500
|
| 703 |
-
},
|
| 704 |
-
{
|
| 705 |
-
"epoch": 3.8823816495851635,
|
| 706 |
-
"eval_runtime": 19.6156,
|
| 707 |
-
"eval_samples_per_second": 22.023,
|
| 708 |
-
"eval_steps_per_second": 5.506,
|
| 709 |
-
"step": 500
|
| 710 |
-
},
|
| 711 |
-
{
|
| 712 |
-
"epoch": 3.9604685212298683,
|
| 713 |
-
"grad_norm": 446547.46875,
|
| 714 |
-
"learning_rate": 6.015625e-05,
|
| 715 |
-
"loss": 1092.9813,
|
| 716 |
-
"step": 510
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 3.9604685212298683,
|
| 720 |
-
"eval_runtime": 19.6311,
|
| 721 |
-
"eval_samples_per_second": 22.006,
|
| 722 |
-
"eval_steps_per_second": 5.501,
|
| 723 |
-
"step": 510
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 4.031234748657882,
|
| 727 |
-
"grad_norm": 255615.1875,
|
| 728 |
-
"learning_rate": 5.9375e-05,
|
| 729 |
-
"loss": 900.3142,
|
| 730 |
-
"step": 520
|
| 731 |
-
},
|
| 732 |
-
{
|
| 733 |
-
"epoch": 4.031234748657882,
|
| 734 |
-
"eval_runtime": 19.5145,
|
| 735 |
-
"eval_samples_per_second": 22.137,
|
| 736 |
-
"eval_steps_per_second": 5.534,
|
| 737 |
-
"step": 520
|
| 738 |
-
},
|
| 739 |
-
{
|
| 740 |
-
"epoch": 4.109321620302587,
|
| 741 |
-
"grad_norm": 318956.46875,
|
| 742 |
-
"learning_rate": 5.8593750000000005e-05,
|
| 743 |
-
"loss": 998.6581,
|
| 744 |
-
"step": 530
|
| 745 |
-
},
|
| 746 |
-
{
|
| 747 |
-
"epoch": 4.109321620302587,
|
| 748 |
-
"eval_runtime": 19.5902,
|
| 749 |
-
"eval_samples_per_second": 22.052,
|
| 750 |
-
"eval_steps_per_second": 5.513,
|
| 751 |
-
"step": 530
|
| 752 |
-
},
|
| 753 |
-
{
|
| 754 |
-
"epoch": 4.187408491947291,
|
| 755 |
-
"grad_norm": 291643.28125,
|
| 756 |
-
"learning_rate": 5.78125e-05,
|
| 757 |
-
"loss": 1226.5887,
|
| 758 |
-
"step": 540
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 4.187408491947291,
|
| 762 |
-
"eval_runtime": 19.6079,
|
| 763 |
-
"eval_samples_per_second": 22.032,
|
| 764 |
-
"eval_steps_per_second": 5.508,
|
| 765 |
-
"step": 540
|
| 766 |
-
},
|
| 767 |
-
{
|
| 768 |
-
"epoch": 4.265495363591996,
|
| 769 |
-
"grad_norm": 283195.4375,
|
| 770 |
-
"learning_rate": 5.703125e-05,
|
| 771 |
-
"loss": 1072.7449,
|
| 772 |
-
"step": 550
|
| 773 |
-
},
|
| 774 |
-
{
|
| 775 |
-
"epoch": 4.265495363591996,
|
| 776 |
-
"eval_runtime": 19.6116,
|
| 777 |
-
"eval_samples_per_second": 22.028,
|
| 778 |
-
"eval_steps_per_second": 5.507,
|
| 779 |
-
"step": 550
|
| 780 |
-
},
|
| 781 |
-
{
|
| 782 |
-
"epoch": 4.343582235236701,
|
| 783 |
-
"grad_norm": 335369.625,
|
| 784 |
-
"learning_rate": 5.6250000000000005e-05,
|
| 785 |
-
"loss": 801.423,
|
| 786 |
-
"step": 560
|
| 787 |
-
},
|
| 788 |
-
{
|
| 789 |
-
"epoch": 4.343582235236701,
|
| 790 |
-
"eval_runtime": 19.6203,
|
| 791 |
-
"eval_samples_per_second": 22.018,
|
| 792 |
-
"eval_steps_per_second": 5.505,
|
| 793 |
-
"step": 560
|
| 794 |
-
},
|
| 795 |
-
{
|
| 796 |
-
"epoch": 4.421669106881406,
|
| 797 |
-
"grad_norm": 363015.75,
|
| 798 |
-
"learning_rate": 5.546875e-05,
|
| 799 |
-
"loss": 1544.8841,
|
| 800 |
-
"step": 570
|
| 801 |
-
},
|
| 802 |
-
{
|
| 803 |
-
"epoch": 4.421669106881406,
|
| 804 |
-
"eval_runtime": 19.633,
|
| 805 |
-
"eval_samples_per_second": 22.004,
|
| 806 |
-
"eval_steps_per_second": 5.501,
|
| 807 |
-
"step": 570
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 4.49975597852611,
|
| 811 |
-
"grad_norm": 314545.34375,
|
| 812 |
-
"learning_rate": 5.46875e-05,
|
| 813 |
-
"loss": 1318.7243,
|
| 814 |
-
"step": 580
|
| 815 |
-
},
|
| 816 |
-
{
|
| 817 |
-
"epoch": 4.49975597852611,
|
| 818 |
-
"eval_runtime": 19.6216,
|
| 819 |
-
"eval_samples_per_second": 22.017,
|
| 820 |
-
"eval_steps_per_second": 5.504,
|
| 821 |
-
"step": 580
|
| 822 |
-
},
|
| 823 |
-
{
|
| 824 |
-
"epoch": 4.577842850170815,
|
| 825 |
-
"grad_norm": 436113.09375,
|
| 826 |
-
"learning_rate": 5.3906250000000006e-05,
|
| 827 |
-
"loss": 1256.4427,
|
| 828 |
-
"step": 590
|
| 829 |
-
},
|
| 830 |
-
{
|
| 831 |
-
"epoch": 4.577842850170815,
|
| 832 |
-
"eval_runtime": 19.6143,
|
| 833 |
-
"eval_samples_per_second": 22.025,
|
| 834 |
-
"eval_steps_per_second": 5.506,
|
| 835 |
-
"step": 590
|
| 836 |
-
},
|
| 837 |
-
{
|
| 838 |
-
"epoch": 4.65592972181552,
|
| 839 |
-
"grad_norm": 307997.78125,
|
| 840 |
-
"learning_rate": 5.3125000000000004e-05,
|
| 841 |
-
"loss": 1325.2429,
|
| 842 |
-
"step": 600
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 4.65592972181552,
|
| 846 |
-
"eval_runtime": 19.6167,
|
| 847 |
-
"eval_samples_per_second": 22.022,
|
| 848 |
-
"eval_steps_per_second": 5.506,
|
| 849 |
-
"step": 600
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 4.7340165934602245,
|
| 853 |
-
"grad_norm": 383733.25,
|
| 854 |
-
"learning_rate": 5.234375e-05,
|
| 855 |
-
"loss": 1247.956,
|
| 856 |
-
"step": 610
|
| 857 |
-
},
|
| 858 |
-
{
|
| 859 |
-
"epoch": 4.7340165934602245,
|
| 860 |
-
"eval_runtime": 19.6214,
|
| 861 |
-
"eval_samples_per_second": 22.017,
|
| 862 |
-
"eval_steps_per_second": 5.504,
|
| 863 |
-
"step": 610
|
| 864 |
-
},
|
| 865 |
-
{
|
| 866 |
-
"epoch": 4.812103465104929,
|
| 867 |
-
"grad_norm": 434974.0,
|
| 868 |
-
"learning_rate": 5.15625e-05,
|
| 869 |
-
"loss": 1198.3052,
|
| 870 |
-
"step": 620
|
| 871 |
-
},
|
| 872 |
-
{
|
| 873 |
-
"epoch": 4.812103465104929,
|
| 874 |
-
"eval_runtime": 19.6132,
|
| 875 |
-
"eval_samples_per_second": 22.026,
|
| 876 |
-
"eval_steps_per_second": 5.506,
|
| 877 |
-
"step": 620
|
| 878 |
-
},
|
| 879 |
-
{
|
| 880 |
-
"epoch": 4.890190336749634,
|
| 881 |
-
"grad_norm": 182716.515625,
|
| 882 |
-
"learning_rate": 5.0781250000000004e-05,
|
| 883 |
-
"loss": 853.0629,
|
| 884 |
-
"step": 630
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 4.890190336749634,
|
| 888 |
-
"eval_runtime": 19.618,
|
| 889 |
-
"eval_samples_per_second": 22.021,
|
| 890 |
-
"eval_steps_per_second": 5.505,
|
| 891 |
-
"step": 630
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 4.968277208394339,
|
| 895 |
-
"grad_norm": 364455.15625,
|
| 896 |
-
"learning_rate": 5e-05,
|
| 897 |
-
"loss": 798.3182,
|
| 898 |
-
"step": 640
|
| 899 |
-
},
|
| 900 |
-
{
|
| 901 |
-
"epoch": 4.968277208394339,
|
| 902 |
-
"eval_runtime": 19.6019,
|
| 903 |
-
"eval_samples_per_second": 22.039,
|
| 904 |
-
"eval_steps_per_second": 5.51,
|
| 905 |
-
"step": 640
|
| 906 |
-
},
|
| 907 |
-
{
|
| 908 |
-
"epoch": 5.039043435822352,
|
| 909 |
-
"grad_norm": 185726.75,
|
| 910 |
-
"learning_rate": 4.921875e-05,
|
| 911 |
-
"loss": 753.334,
|
| 912 |
-
"step": 650
|
| 913 |
-
},
|
| 914 |
-
{
|
| 915 |
-
"epoch": 5.039043435822352,
|
| 916 |
-
"eval_runtime": 19.544,
|
| 917 |
-
"eval_samples_per_second": 22.104,
|
| 918 |
-
"eval_steps_per_second": 5.526,
|
| 919 |
-
"step": 650
|
| 920 |
-
},
|
| 921 |
-
{
|
| 922 |
-
"epoch": 5.1171303074670575,
|
| 923 |
-
"grad_norm": 219226.359375,
|
| 924 |
-
"learning_rate": 4.8437500000000005e-05,
|
| 925 |
-
"loss": 1082.2985,
|
| 926 |
-
"step": 660
|
| 927 |
-
},
|
| 928 |
-
{
|
| 929 |
-
"epoch": 5.1171303074670575,
|
| 930 |
-
"eval_runtime": 19.5817,
|
| 931 |
-
"eval_samples_per_second": 22.061,
|
| 932 |
-
"eval_steps_per_second": 5.515,
|
| 933 |
-
"step": 660
|
| 934 |
-
},
|
| 935 |
-
{
|
| 936 |
-
"epoch": 5.195217179111761,
|
| 937 |
-
"grad_norm": 207455.5625,
|
| 938 |
-
"learning_rate": 4.765625e-05,
|
| 939 |
-
"loss": 879.2348,
|
| 940 |
-
"step": 670
|
| 941 |
-
},
|
| 942 |
-
{
|
| 943 |
-
"epoch": 5.195217179111761,
|
| 944 |
-
"eval_runtime": 19.5896,
|
| 945 |
-
"eval_samples_per_second": 22.053,
|
| 946 |
-
"eval_steps_per_second": 5.513,
|
| 947 |
-
"step": 670
|
| 948 |
-
},
|
| 949 |
-
{
|
| 950 |
-
"epoch": 5.273304050756467,
|
| 951 |
-
"grad_norm": 379599.875,
|
| 952 |
-
"learning_rate": 4.6875e-05,
|
| 953 |
-
"loss": 896.7317,
|
| 954 |
-
"step": 680
|
| 955 |
-
},
|
| 956 |
-
{
|
| 957 |
-
"epoch": 5.273304050756467,
|
| 958 |
-
"eval_runtime": 19.5928,
|
| 959 |
-
"eval_samples_per_second": 22.049,
|
| 960 |
-
"eval_steps_per_second": 5.512,
|
| 961 |
-
"step": 680
|
| 962 |
-
},
|
| 963 |
-
{
|
| 964 |
-
"epoch": 5.351390922401171,
|
| 965 |
-
"grad_norm": 303185.625,
|
| 966 |
-
"learning_rate": 4.609375e-05,
|
| 967 |
-
"loss": 754.6702,
|
| 968 |
-
"step": 690
|
| 969 |
-
},
|
| 970 |
-
{
|
| 971 |
-
"epoch": 5.351390922401171,
|
| 972 |
-
"eval_runtime": 19.5979,
|
| 973 |
-
"eval_samples_per_second": 22.043,
|
| 974 |
-
"eval_steps_per_second": 5.511,
|
| 975 |
-
"step": 690
|
| 976 |
-
},
|
| 977 |
-
{
|
| 978 |
-
"epoch": 5.4294777940458765,
|
| 979 |
-
"grad_norm": 405251.96875,
|
| 980 |
-
"learning_rate": 4.5312500000000004e-05,
|
| 981 |
-
"loss": 715.7025,
|
| 982 |
-
"step": 700
|
| 983 |
-
},
|
| 984 |
-
{
|
| 985 |
-
"epoch": 5.4294777940458765,
|
| 986 |
-
"eval_runtime": 19.6118,
|
| 987 |
-
"eval_samples_per_second": 22.028,
|
| 988 |
-
"eval_steps_per_second": 5.507,
|
| 989 |
-
"step": 700
|
| 990 |
-
},
|
| 991 |
-
{
|
| 992 |
-
"epoch": 5.50756466569058,
|
| 993 |
-
"grad_norm": 204986.03125,
|
| 994 |
-
"learning_rate": 4.453125e-05,
|
| 995 |
-
"loss": 728.391,
|
| 996 |
-
"step": 710
|
| 997 |
-
},
|
| 998 |
-
{
|
| 999 |
-
"epoch": 5.50756466569058,
|
| 1000 |
-
"eval_runtime": 19.6098,
|
| 1001 |
-
"eval_samples_per_second": 22.03,
|
| 1002 |
-
"eval_steps_per_second": 5.507,
|
| 1003 |
-
"step": 710
|
| 1004 |
-
},
|
| 1005 |
-
{
|
| 1006 |
-
"epoch": 5.585651537335286,
|
| 1007 |
-
"grad_norm": 321820.75,
|
| 1008 |
-
"learning_rate": 4.375e-05,
|
| 1009 |
-
"loss": 699.9287,
|
| 1010 |
-
"step": 720
|
| 1011 |
-
},
|
| 1012 |
-
{
|
| 1013 |
-
"epoch": 5.585651537335286,
|
| 1014 |
-
"eval_runtime": 19.618,
|
| 1015 |
-
"eval_samples_per_second": 22.021,
|
| 1016 |
-
"eval_steps_per_second": 5.505,
|
| 1017 |
-
"step": 720
|
| 1018 |
-
},
|
| 1019 |
-
{
|
| 1020 |
-
"epoch": 5.66373840897999,
|
| 1021 |
-
"grad_norm": 168577.65625,
|
| 1022 |
-
"learning_rate": 4.2968750000000004e-05,
|
| 1023 |
-
"loss": 586.6525,
|
| 1024 |
-
"step": 730
|
| 1025 |
-
},
|
| 1026 |
-
{
|
| 1027 |
-
"epoch": 5.66373840897999,
|
| 1028 |
-
"eval_runtime": 19.6206,
|
| 1029 |
-
"eval_samples_per_second": 22.018,
|
| 1030 |
-
"eval_steps_per_second": 5.504,
|
| 1031 |
-
"step": 730
|
| 1032 |
-
},
|
| 1033 |
-
{
|
| 1034 |
-
"epoch": 5.741825280624695,
|
| 1035 |
-
"grad_norm": 244243.9375,
|
| 1036 |
-
"learning_rate": 4.21875e-05,
|
| 1037 |
-
"loss": 625.2781,
|
| 1038 |
-
"step": 740
|
| 1039 |
-
},
|
| 1040 |
-
{
|
| 1041 |
-
"epoch": 5.741825280624695,
|
| 1042 |
-
"eval_runtime": 19.6218,
|
| 1043 |
-
"eval_samples_per_second": 22.016,
|
| 1044 |
-
"eval_steps_per_second": 5.504,
|
| 1045 |
-
"step": 740
|
| 1046 |
-
},
|
| 1047 |
-
{
|
| 1048 |
-
"epoch": 5.819912152269399,
|
| 1049 |
-
"grad_norm": 192660.640625,
|
| 1050 |
-
"learning_rate": 4.140625e-05,
|
| 1051 |
-
"loss": 719.4099,
|
| 1052 |
-
"step": 750
|
| 1053 |
-
},
|
| 1054 |
-
{
|
| 1055 |
-
"epoch": 5.819912152269399,
|
| 1056 |
-
"eval_runtime": 19.6162,
|
| 1057 |
-
"eval_samples_per_second": 22.023,
|
| 1058 |
-
"eval_steps_per_second": 5.506,
|
| 1059 |
-
"step": 750
|
| 1060 |
-
},
|
| 1061 |
-
{
|
| 1062 |
-
"epoch": 5.897999023914105,
|
| 1063 |
-
"grad_norm": 146700.25,
|
| 1064 |
-
"learning_rate": 4.0625000000000005e-05,
|
| 1065 |
-
"loss": 577.4713,
|
| 1066 |
-
"step": 760
|
| 1067 |
-
},
|
| 1068 |
-
{
|
| 1069 |
-
"epoch": 5.897999023914105,
|
| 1070 |
-
"eval_runtime": 19.6213,
|
| 1071 |
-
"eval_samples_per_second": 22.017,
|
| 1072 |
-
"eval_steps_per_second": 5.504,
|
| 1073 |
-
"step": 760
|
| 1074 |
-
},
|
| 1075 |
-
{
|
| 1076 |
-
"epoch": 5.976085895558809,
|
| 1077 |
-
"grad_norm": 209883.296875,
|
| 1078 |
-
"learning_rate": 3.984375e-05,
|
| 1079 |
-
"loss": 637.9117,
|
| 1080 |
-
"step": 770
|
| 1081 |
-
},
|
| 1082 |
-
{
|
| 1083 |
-
"epoch": 5.976085895558809,
|
| 1084 |
-
"eval_runtime": 19.6267,
|
| 1085 |
-
"eval_samples_per_second": 22.011,
|
| 1086 |
-
"eval_steps_per_second": 5.503,
|
| 1087 |
-
"step": 770
|
| 1088 |
-
},
|
| 1089 |
-
{
|
| 1090 |
-
"epoch": 6.046852122986823,
|
| 1091 |
-
"grad_norm": 468637.6875,
|
| 1092 |
-
"learning_rate": 3.90625e-05,
|
| 1093 |
-
"loss": 621.1816,
|
| 1094 |
-
"step": 780
|
| 1095 |
-
},
|
| 1096 |
-
{
|
| 1097 |
-
"epoch": 6.046852122986823,
|
| 1098 |
-
"eval_runtime": 19.6074,
|
| 1099 |
-
"eval_samples_per_second": 22.032,
|
| 1100 |
-
"eval_steps_per_second": 5.508,
|
| 1101 |
-
"step": 780
|
| 1102 |
-
},
|
| 1103 |
-
{
|
| 1104 |
-
"epoch": 6.124938994631528,
|
| 1105 |
-
"grad_norm": 374385.40625,
|
| 1106 |
-
"learning_rate": 3.828125e-05,
|
| 1107 |
-
"loss": 914.6565,
|
| 1108 |
-
"step": 790
|
| 1109 |
-
},
|
| 1110 |
-
{
|
| 1111 |
-
"epoch": 6.124938994631528,
|
| 1112 |
-
"eval_runtime": 19.626,
|
| 1113 |
-
"eval_samples_per_second": 22.012,
|
| 1114 |
-
"eval_steps_per_second": 5.503,
|
| 1115 |
-
"step": 790
|
| 1116 |
-
},
|
| 1117 |
-
{
|
| 1118 |
-
"epoch": 6.203025866276232,
|
| 1119 |
-
"grad_norm": 283100.3125,
|
| 1120 |
-
"learning_rate": 3.7500000000000003e-05,
|
| 1121 |
-
"loss": 830.0896,
|
| 1122 |
-
"step": 800
|
| 1123 |
-
},
|
| 1124 |
-
{
|
| 1125 |
-
"epoch": 6.203025866276232,
|
| 1126 |
-
"eval_runtime": 19.6195,
|
| 1127 |
-
"eval_samples_per_second": 22.019,
|
| 1128 |
-
"eval_steps_per_second": 5.505,
|
| 1129 |
-
"step": 800
|
| 1130 |
-
},
|
| 1131 |
-
{
|
| 1132 |
-
"epoch": 6.281112737920937,
|
| 1133 |
-
"grad_norm": 186444.921875,
|
| 1134 |
-
"learning_rate": 3.671875e-05,
|
| 1135 |
-
"loss": 822.068,
|
| 1136 |
-
"step": 810
|
| 1137 |
-
},
|
| 1138 |
-
{
|
| 1139 |
-
"epoch": 6.281112737920937,
|
| 1140 |
-
"eval_runtime": 19.6377,
|
| 1141 |
-
"eval_samples_per_second": 21.998,
|
| 1142 |
-
"eval_steps_per_second": 5.5,
|
| 1143 |
-
"step": 810
|
| 1144 |
-
},
|
| 1145 |
-
{
|
| 1146 |
-
"epoch": 6.359199609565642,
|
| 1147 |
-
"grad_norm": 431370.5,
|
| 1148 |
-
"learning_rate": 3.59375e-05,
|
| 1149 |
-
"loss": 690.5968,
|
| 1150 |
-
"step": 820
|
| 1151 |
-
},
|
| 1152 |
-
{
|
| 1153 |
-
"epoch": 6.359199609565642,
|
| 1154 |
-
"eval_runtime": 19.6147,
|
| 1155 |
-
"eval_samples_per_second": 22.024,
|
| 1156 |
-
"eval_steps_per_second": 5.506,
|
| 1157 |
-
"step": 820
|
| 1158 |
-
},
|
| 1159 |
-
{
|
| 1160 |
-
"epoch": 6.4372864812103465,
|
| 1161 |
-
"grad_norm": 435885.03125,
|
| 1162 |
-
"learning_rate": 3.5156250000000004e-05,
|
| 1163 |
-
"loss": 1016.4219,
|
| 1164 |
-
"step": 830
|
| 1165 |
-
},
|
| 1166 |
-
{
|
| 1167 |
-
"epoch": 6.4372864812103465,
|
| 1168 |
-
"eval_runtime": 19.6089,
|
| 1169 |
-
"eval_samples_per_second": 22.031,
|
| 1170 |
-
"eval_steps_per_second": 5.508,
|
| 1171 |
-
"step": 830
|
| 1172 |
-
},
|
| 1173 |
-
{
|
| 1174 |
-
"epoch": 6.515373352855051,
|
| 1175 |
-
"grad_norm": 473700.9375,
|
| 1176 |
-
"learning_rate": 3.4375e-05,
|
| 1177 |
-
"loss": 806.1664,
|
| 1178 |
-
"step": 840
|
| 1179 |
-
},
|
| 1180 |
-
{
|
| 1181 |
-
"epoch": 6.515373352855051,
|
| 1182 |
-
"eval_runtime": 19.6159,
|
| 1183 |
-
"eval_samples_per_second": 22.023,
|
| 1184 |
-
"eval_steps_per_second": 5.506,
|
| 1185 |
-
"step": 840
|
| 1186 |
-
},
|
| 1187 |
-
{
|
| 1188 |
-
"epoch": 6.593460224499756,
|
| 1189 |
-
"grad_norm": 398720.625,
|
| 1190 |
-
"learning_rate": 3.359375e-05,
|
| 1191 |
-
"loss": 668.8824,
|
| 1192 |
-
"step": 850
|
| 1193 |
-
},
|
| 1194 |
-
{
|
| 1195 |
-
"epoch": 6.593460224499756,
|
| 1196 |
-
"eval_runtime": 19.6281,
|
| 1197 |
-
"eval_samples_per_second": 22.009,
|
| 1198 |
-
"eval_steps_per_second": 5.502,
|
| 1199 |
-
"step": 850
|
| 1200 |
-
},
|
| 1201 |
-
{
|
| 1202 |
-
"epoch": 6.671547096144461,
|
| 1203 |
-
"grad_norm": 305606.28125,
|
| 1204 |
-
"learning_rate": 3.2812500000000005e-05,
|
| 1205 |
-
"loss": 522.7901,
|
| 1206 |
-
"step": 860
|
| 1207 |
-
},
|
| 1208 |
-
{
|
| 1209 |
-
"epoch": 6.671547096144461,
|
| 1210 |
-
"eval_runtime": 19.6214,
|
| 1211 |
-
"eval_samples_per_second": 22.017,
|
| 1212 |
-
"eval_steps_per_second": 5.504,
|
| 1213 |
-
"step": 860
|
| 1214 |
-
},
|
| 1215 |
-
{
|
| 1216 |
-
"epoch": 6.7496339677891655,
|
| 1217 |
-
"grad_norm": 263199.21875,
|
| 1218 |
-
"learning_rate": 3.203125e-05,
|
| 1219 |
-
"loss": 559.4416,
|
| 1220 |
-
"step": 870
|
| 1221 |
-
},
|
| 1222 |
-
{
|
| 1223 |
-
"epoch": 6.7496339677891655,
|
| 1224 |
-
"eval_runtime": 19.6359,
|
| 1225 |
-
"eval_samples_per_second": 22.0,
|
| 1226 |
-
"eval_steps_per_second": 5.5,
|
| 1227 |
-
"step": 870
|
| 1228 |
-
},
|
| 1229 |
-
{
|
| 1230 |
-
"epoch": 6.82772083943387,
|
| 1231 |
-
"grad_norm": 203541.921875,
|
| 1232 |
-
"learning_rate": 3.125e-05,
|
| 1233 |
-
"loss": 602.3201,
|
| 1234 |
-
"step": 880
|
| 1235 |
-
},
|
| 1236 |
-
{
|
| 1237 |
-
"epoch": 6.82772083943387,
|
| 1238 |
-
"eval_runtime": 19.6292,
|
| 1239 |
-
"eval_samples_per_second": 22.008,
|
| 1240 |
-
"eval_steps_per_second": 5.502,
|
| 1241 |
-
"step": 880
|
| 1242 |
-
},
|
| 1243 |
-
{
|
| 1244 |
-
"epoch": 6.905807711078575,
|
| 1245 |
-
"grad_norm": 109014.71875,
|
| 1246 |
-
"learning_rate": 3.0468750000000002e-05,
|
| 1247 |
-
"loss": 527.6462,
|
| 1248 |
-
"step": 890
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 6.905807711078575,
|
| 1252 |
-
"eval_runtime": 19.615,
|
| 1253 |
-
"eval_samples_per_second": 22.024,
|
| 1254 |
-
"eval_steps_per_second": 5.506,
|
| 1255 |
-
"step": 890
|
| 1256 |
-
},
|
| 1257 |
-
{
|
| 1258 |
-
"epoch": 6.98389458272328,
|
| 1259 |
-
"grad_norm": 173269.515625,
|
| 1260 |
-
"learning_rate": 2.96875e-05,
|
| 1261 |
-
"loss": 509.9485,
|
| 1262 |
-
"step": 900
|
| 1263 |
-
},
|
| 1264 |
-
{
|
| 1265 |
-
"epoch": 6.98389458272328,
|
| 1266 |
-
"eval_runtime": 19.6287,
|
| 1267 |
-
"eval_samples_per_second": 22.009,
|
| 1268 |
-
"eval_steps_per_second": 5.502,
|
| 1269 |
-
"step": 900
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 7.054660810151294,
|
| 1273 |
-
"grad_norm": 89940.78125,
|
| 1274 |
-
"learning_rate": 2.890625e-05,
|
| 1275 |
-
"loss": 514.7005,
|
| 1276 |
-
"step": 910
|
| 1277 |
-
},
|
| 1278 |
-
{
|
| 1279 |
-
"epoch": 7.054660810151294,
|
| 1280 |
-
"eval_runtime": 19.6378,
|
| 1281 |
-
"eval_samples_per_second": 21.998,
|
| 1282 |
-
"eval_steps_per_second": 5.5,
|
| 1283 |
-
"step": 910
|
| 1284 |
-
},
|
| 1285 |
-
{
|
| 1286 |
-
"epoch": 7.132747681795998,
|
| 1287 |
-
"grad_norm": 64957.203125,
|
| 1288 |
-
"learning_rate": 2.8125000000000003e-05,
|
| 1289 |
-
"loss": 516.4828,
|
| 1290 |
-
"step": 920
|
| 1291 |
-
},
|
| 1292 |
-
{
|
| 1293 |
-
"epoch": 7.132747681795998,
|
| 1294 |
-
"eval_runtime": 19.6159,
|
| 1295 |
-
"eval_samples_per_second": 22.023,
|
| 1296 |
-
"eval_steps_per_second": 5.506,
|
| 1297 |
-
"step": 920
|
| 1298 |
-
},
|
| 1299 |
-
{
|
| 1300 |
-
"epoch": 7.210834553440703,
|
| 1301 |
-
"grad_norm": 191568.875,
|
| 1302 |
-
"learning_rate": 2.734375e-05,
|
| 1303 |
-
"loss": 469.4625,
|
| 1304 |
-
"step": 930
|
| 1305 |
-
},
|
| 1306 |
-
{
|
| 1307 |
-
"epoch": 7.210834553440703,
|
| 1308 |
-
"eval_runtime": 19.6149,
|
| 1309 |
-
"eval_samples_per_second": 22.024,
|
| 1310 |
-
"eval_steps_per_second": 5.506,
|
| 1311 |
-
"step": 930
|
| 1312 |
-
},
|
| 1313 |
-
{
|
| 1314 |
-
"epoch": 7.288921425085407,
|
| 1315 |
-
"grad_norm": 180467.84375,
|
| 1316 |
-
"learning_rate": 2.6562500000000002e-05,
|
| 1317 |
-
"loss": 627.1263,
|
| 1318 |
-
"step": 940
|
| 1319 |
-
},
|
| 1320 |
-
{
|
| 1321 |
-
"epoch": 7.288921425085407,
|
| 1322 |
-
"eval_runtime": 19.6273,
|
| 1323 |
-
"eval_samples_per_second": 22.01,
|
| 1324 |
-
"eval_steps_per_second": 5.503,
|
| 1325 |
-
"step": 940
|
| 1326 |
-
},
|
| 1327 |
-
{
|
| 1328 |
-
"epoch": 7.367008296730113,
|
| 1329 |
-
"grad_norm": 175262.3125,
|
| 1330 |
-
"learning_rate": 2.578125e-05,
|
| 1331 |
-
"loss": 497.1456,
|
| 1332 |
-
"step": 950
|
| 1333 |
-
},
|
| 1334 |
-
{
|
| 1335 |
-
"epoch": 7.367008296730113,
|
| 1336 |
-
"eval_runtime": 19.6298,
|
| 1337 |
-
"eval_samples_per_second": 22.007,
|
| 1338 |
-
"eval_steps_per_second": 5.502,
|
| 1339 |
-
"step": 950
|
| 1340 |
-
},
|
| 1341 |
-
{
|
| 1342 |
-
"epoch": 7.4450951683748166,
|
| 1343 |
-
"grad_norm": 194304.703125,
|
| 1344 |
-
"learning_rate": 2.5e-05,
|
| 1345 |
-
"loss": 651.0766,
|
| 1346 |
-
"step": 960
|
| 1347 |
-
},
|
| 1348 |
-
{
|
| 1349 |
-
"epoch": 7.4450951683748166,
|
| 1350 |
-
"eval_runtime": 19.6413,
|
| 1351 |
-
"eval_samples_per_second": 21.994,
|
| 1352 |
-
"eval_steps_per_second": 5.499,
|
| 1353 |
-
"step": 960
|
| 1354 |
-
},
|
| 1355 |
-
{
|
| 1356 |
-
"epoch": 7.523182040019522,
|
| 1357 |
-
"grad_norm": 221815.171875,
|
| 1358 |
-
"learning_rate": 2.4218750000000003e-05,
|
| 1359 |
-
"loss": 419.7869,
|
| 1360 |
-
"step": 970
|
| 1361 |
-
},
|
| 1362 |
-
{
|
| 1363 |
-
"epoch": 7.523182040019522,
|
| 1364 |
-
"eval_runtime": 19.6255,
|
| 1365 |
-
"eval_samples_per_second": 22.012,
|
| 1366 |
-
"eval_steps_per_second": 5.503,
|
| 1367 |
-
"step": 970
|
| 1368 |
-
},
|
| 1369 |
-
{
|
| 1370 |
-
"epoch": 7.601268911664226,
|
| 1371 |
-
"grad_norm": 330153.84375,
|
| 1372 |
-
"learning_rate": 2.34375e-05,
|
| 1373 |
-
"loss": 606.2052,
|
| 1374 |
-
"step": 980
|
| 1375 |
-
},
|
| 1376 |
-
{
|
| 1377 |
-
"epoch": 7.601268911664226,
|
| 1378 |
-
"eval_runtime": 19.6246,
|
| 1379 |
-
"eval_samples_per_second": 22.013,
|
| 1380 |
-
"eval_steps_per_second": 5.503,
|
| 1381 |
-
"step": 980
|
| 1382 |
-
},
|
| 1383 |
-
{
|
| 1384 |
-
"epoch": 7.679355783308932,
|
| 1385 |
-
"grad_norm": 134138.4375,
|
| 1386 |
-
"learning_rate": 2.2656250000000002e-05,
|
| 1387 |
-
"loss": 636.721,
|
| 1388 |
-
"step": 990
|
| 1389 |
-
},
|
| 1390 |
-
{
|
| 1391 |
-
"epoch": 7.679355783308932,
|
| 1392 |
-
"eval_runtime": 19.622,
|
| 1393 |
-
"eval_samples_per_second": 22.016,
|
| 1394 |
-
"eval_steps_per_second": 5.504,
|
| 1395 |
-
"step": 990
|
| 1396 |
-
},
|
| 1397 |
-
{
|
| 1398 |
-
"epoch": 7.7574426549536355,
|
| 1399 |
-
"grad_norm": 86497.265625,
|
| 1400 |
-
"learning_rate": 2.1875e-05,
|
| 1401 |
-
"loss": 565.6015,
|
| 1402 |
-
"step": 1000
|
| 1403 |
-
},
|
| 1404 |
-
{
|
| 1405 |
-
"epoch": 7.7574426549536355,
|
| 1406 |
-
"eval_runtime": 19.6167,
|
| 1407 |
-
"eval_samples_per_second": 22.022,
|
| 1408 |
-
"eval_steps_per_second": 5.506,
|
| 1409 |
-
"step": 1000
|
| 1410 |
-
},
|
| 1411 |
-
{
|
| 1412 |
-
"epoch": 7.835529526598341,
|
| 1413 |
-
"grad_norm": 256321.390625,
|
| 1414 |
-
"learning_rate": 2.109375e-05,
|
| 1415 |
-
"loss": 383.8543,
|
| 1416 |
-
"step": 1010
|
| 1417 |
-
},
|
| 1418 |
-
{
|
| 1419 |
-
"epoch": 7.835529526598341,
|
| 1420 |
-
"eval_runtime": 19.6312,
|
| 1421 |
-
"eval_samples_per_second": 22.006,
|
| 1422 |
-
"eval_steps_per_second": 5.501,
|
| 1423 |
-
"step": 1010
|
| 1424 |
-
},
|
| 1425 |
-
{
|
| 1426 |
-
"epoch": 7.913616398243045,
|
| 1427 |
-
"grad_norm": 300360.125,
|
| 1428 |
-
"learning_rate": 2.0312500000000002e-05,
|
| 1429 |
-
"loss": 386.0633,
|
| 1430 |
-
"step": 1020
|
| 1431 |
-
},
|
| 1432 |
-
{
|
| 1433 |
-
"epoch": 7.913616398243045,
|
| 1434 |
-
"eval_runtime": 19.6333,
|
| 1435 |
-
"eval_samples_per_second": 22.003,
|
| 1436 |
-
"eval_steps_per_second": 5.501,
|
| 1437 |
-
"step": 1020
|
| 1438 |
-
},
|
| 1439 |
-
{
|
| 1440 |
-
"epoch": 7.991703269887751,
|
| 1441 |
-
"grad_norm": 65186.7578125,
|
| 1442 |
-
"learning_rate": 1.953125e-05,
|
| 1443 |
-
"loss": 379.998,
|
| 1444 |
-
"step": 1030
|
| 1445 |
-
},
|
| 1446 |
-
{
|
| 1447 |
-
"epoch": 7.991703269887751,
|
| 1448 |
-
"eval_runtime": 19.6079,
|
| 1449 |
-
"eval_samples_per_second": 22.032,
|
| 1450 |
-
"eval_steps_per_second": 5.508,
|
| 1451 |
-
"step": 1030
|
| 1452 |
-
},
|
| 1453 |
-
{
|
| 1454 |
-
"epoch": 8.062469497315764,
|
| 1455 |
-
"grad_norm": 306364.5,
|
| 1456 |
-
"learning_rate": 1.8750000000000002e-05,
|
| 1457 |
-
"loss": 458.2917,
|
| 1458 |
-
"step": 1040
|
| 1459 |
-
},
|
| 1460 |
-
{
|
| 1461 |
-
"epoch": 8.062469497315764,
|
| 1462 |
-
"eval_runtime": 19.626,
|
| 1463 |
-
"eval_samples_per_second": 22.012,
|
| 1464 |
-
"eval_steps_per_second": 5.503,
|
| 1465 |
-
"step": 1040
|
| 1466 |
-
},
|
| 1467 |
-
{
|
| 1468 |
-
"epoch": 8.140556368960468,
|
| 1469 |
-
"grad_norm": 287573.96875,
|
| 1470 |
-
"learning_rate": 1.796875e-05,
|
| 1471 |
-
"loss": 459.69,
|
| 1472 |
-
"step": 1050
|
| 1473 |
-
},
|
| 1474 |
-
{
|
| 1475 |
-
"epoch": 8.140556368960468,
|
| 1476 |
-
"eval_runtime": 19.64,
|
| 1477 |
-
"eval_samples_per_second": 21.996,
|
| 1478 |
-
"eval_steps_per_second": 5.499,
|
| 1479 |
-
"step": 1050
|
| 1480 |
-
},
|
| 1481 |
-
{
|
| 1482 |
-
"epoch": 8.218643240605173,
|
| 1483 |
-
"grad_norm": 87142.5625,
|
| 1484 |
-
"learning_rate": 1.71875e-05,
|
| 1485 |
-
"loss": 380.1467,
|
| 1486 |
-
"step": 1060
|
| 1487 |
-
},
|
| 1488 |
-
{
|
| 1489 |
-
"epoch": 8.218643240605173,
|
| 1490 |
-
"eval_runtime": 19.6334,
|
| 1491 |
-
"eval_samples_per_second": 22.003,
|
| 1492 |
-
"eval_steps_per_second": 5.501,
|
| 1493 |
-
"step": 1060
|
| 1494 |
-
},
|
| 1495 |
-
{
|
| 1496 |
-
"epoch": 8.296730112249879,
|
| 1497 |
-
"grad_norm": 301649.96875,
|
| 1498 |
-
"learning_rate": 1.6406250000000002e-05,
|
| 1499 |
-
"loss": 384.5057,
|
| 1500 |
-
"step": 1070
|
| 1501 |
-
},
|
| 1502 |
-
{
|
| 1503 |
-
"epoch": 8.296730112249879,
|
| 1504 |
-
"eval_runtime": 19.6212,
|
| 1505 |
-
"eval_samples_per_second": 22.017,
|
| 1506 |
-
"eval_steps_per_second": 5.504,
|
| 1507 |
-
"step": 1070
|
| 1508 |
-
},
|
| 1509 |
-
{
|
| 1510 |
-
"epoch": 8.374816983894583,
|
| 1511 |
-
"grad_norm": 236951.1875,
|
| 1512 |
-
"learning_rate": 1.5625e-05,
|
| 1513 |
-
"loss": 374.7868,
|
| 1514 |
-
"step": 1080
|
| 1515 |
-
},
|
| 1516 |
-
{
|
| 1517 |
-
"epoch": 8.374816983894583,
|
| 1518 |
-
"eval_runtime": 19.6187,
|
| 1519 |
-
"eval_samples_per_second": 22.02,
|
| 1520 |
-
"eval_steps_per_second": 5.505,
|
| 1521 |
-
"step": 1080
|
| 1522 |
-
},
|
| 1523 |
-
{
|
| 1524 |
-
"epoch": 8.452903855539287,
|
| 1525 |
-
"grad_norm": 76360.2734375,
|
| 1526 |
-
"learning_rate": 1.484375e-05,
|
| 1527 |
-
"loss": 312.2099,
|
| 1528 |
-
"step": 1090
|
| 1529 |
-
},
|
| 1530 |
-
{
|
| 1531 |
-
"epoch": 8.452903855539287,
|
| 1532 |
-
"eval_runtime": 19.6209,
|
| 1533 |
-
"eval_samples_per_second": 22.017,
|
| 1534 |
-
"eval_steps_per_second": 5.504,
|
| 1535 |
-
"step": 1090
|
| 1536 |
-
},
|
| 1537 |
-
{
|
| 1538 |
-
"epoch": 8.530990727183992,
|
| 1539 |
-
"grad_norm": 76876.0859375,
|
| 1540 |
-
"learning_rate": 1.4062500000000001e-05,
|
| 1541 |
-
"loss": 303.1329,
|
| 1542 |
-
"step": 1100
|
| 1543 |
-
},
|
| 1544 |
-
{
|
| 1545 |
-
"epoch": 8.530990727183992,
|
| 1546 |
-
"eval_runtime": 19.6126,
|
| 1547 |
-
"eval_samples_per_second": 22.027,
|
| 1548 |
-
"eval_steps_per_second": 5.507,
|
| 1549 |
-
"step": 1100
|
| 1550 |
-
},
|
| 1551 |
-
{
|
| 1552 |
-
"epoch": 8.609077598828698,
|
| 1553 |
-
"grad_norm": 80524.609375,
|
| 1554 |
-
"learning_rate": 1.3281250000000001e-05,
|
| 1555 |
-
"loss": 336.2521,
|
| 1556 |
-
"step": 1110
|
| 1557 |
-
},
|
| 1558 |
-
{
|
| 1559 |
-
"epoch": 8.609077598828698,
|
| 1560 |
-
"eval_runtime": 19.5907,
|
| 1561 |
-
"eval_samples_per_second": 22.051,
|
| 1562 |
-
"eval_steps_per_second": 5.513,
|
| 1563 |
-
"step": 1110
|
| 1564 |
-
},
|
| 1565 |
-
{
|
| 1566 |
-
"epoch": 8.687164470473402,
|
| 1567 |
-
"grad_norm": 114438.3828125,
|
| 1568 |
-
"learning_rate": 1.25e-05,
|
| 1569 |
-
"loss": 342.3281,
|
| 1570 |
-
"step": 1120
|
| 1571 |
-
},
|
| 1572 |
-
{
|
| 1573 |
-
"epoch": 8.687164470473402,
|
| 1574 |
-
"eval_runtime": 19.614,
|
| 1575 |
-
"eval_samples_per_second": 22.025,
|
| 1576 |
-
"eval_steps_per_second": 5.506,
|
| 1577 |
-
"step": 1120
|
| 1578 |
-
},
|
| 1579 |
-
{
|
| 1580 |
-
"epoch": 8.765251342118106,
|
| 1581 |
-
"grad_norm": 281197.375,
|
| 1582 |
-
"learning_rate": 1.171875e-05,
|
| 1583 |
-
"loss": 358.492,
|
| 1584 |
-
"step": 1130
|
| 1585 |
-
},
|
| 1586 |
-
{
|
| 1587 |
-
"epoch": 8.765251342118106,
|
| 1588 |
-
"eval_runtime": 19.5986,
|
| 1589 |
-
"eval_samples_per_second": 22.042,
|
| 1590 |
-
"eval_steps_per_second": 5.511,
|
| 1591 |
-
"step": 1130
|
| 1592 |
-
},
|
| 1593 |
-
{
|
| 1594 |
-
"epoch": 8.843338213762811,
|
| 1595 |
-
"grad_norm": 39132.8515625,
|
| 1596 |
-
"learning_rate": 1.09375e-05,
|
| 1597 |
-
"loss": 297.0417,
|
| 1598 |
-
"step": 1140
|
| 1599 |
-
},
|
| 1600 |
-
{
|
| 1601 |
-
"epoch": 8.843338213762811,
|
| 1602 |
-
"eval_runtime": 19.6013,
|
| 1603 |
-
"eval_samples_per_second": 22.039,
|
| 1604 |
-
"eval_steps_per_second": 5.51,
|
| 1605 |
-
"step": 1140
|
| 1606 |
-
},
|
| 1607 |
-
{
|
| 1608 |
-
"epoch": 8.921425085407517,
|
| 1609 |
-
"grad_norm": 270795.65625,
|
| 1610 |
-
"learning_rate": 1.0156250000000001e-05,
|
| 1611 |
-
"loss": 306.2402,
|
| 1612 |
-
"step": 1150
|
| 1613 |
-
},
|
| 1614 |
-
{
|
| 1615 |
-
"epoch": 8.921425085407517,
|
| 1616 |
-
"eval_runtime": 19.6264,
|
| 1617 |
-
"eval_samples_per_second": 22.011,
|
| 1618 |
-
"eval_steps_per_second": 5.503,
|
| 1619 |
-
"step": 1150
|
| 1620 |
-
},
|
| 1621 |
-
{
|
| 1622 |
-
"epoch": 8.99951195705222,
|
| 1623 |
-
"grad_norm": 124614.0390625,
|
| 1624 |
-
"learning_rate": 9.375000000000001e-06,
|
| 1625 |
-
"loss": 354.61,
|
| 1626 |
-
"step": 1160
|
| 1627 |
-
},
|
| 1628 |
-
{
|
| 1629 |
-
"epoch": 8.99951195705222,
|
| 1630 |
-
"eval_runtime": 19.6258,
|
| 1631 |
-
"eval_samples_per_second": 22.012,
|
| 1632 |
-
"eval_steps_per_second": 5.503,
|
| 1633 |
-
"step": 1160
|
| 1634 |
-
},
|
| 1635 |
-
{
|
| 1636 |
-
"epoch": 9.070278184480234,
|
| 1637 |
-
"grad_norm": 194401.953125,
|
| 1638 |
-
"learning_rate": 8.59375e-06,
|
| 1639 |
-
"loss": 306.9392,
|
| 1640 |
-
"step": 1170
|
| 1641 |
-
},
|
| 1642 |
-
{
|
| 1643 |
-
"epoch": 9.070278184480234,
|
| 1644 |
-
"eval_runtime": 19.6327,
|
| 1645 |
-
"eval_samples_per_second": 22.004,
|
| 1646 |
-
"eval_steps_per_second": 5.501,
|
| 1647 |
-
"step": 1170
|
| 1648 |
-
},
|
| 1649 |
-
{
|
| 1650 |
-
"epoch": 9.14836505612494,
|
| 1651 |
-
"grad_norm": 134935.84375,
|
| 1652 |
-
"learning_rate": 7.8125e-06,
|
| 1653 |
-
"loss": 309.7358,
|
| 1654 |
-
"step": 1180
|
| 1655 |
-
},
|
| 1656 |
-
{
|
| 1657 |
-
"epoch": 9.14836505612494,
|
| 1658 |
-
"eval_runtime": 19.625,
|
| 1659 |
-
"eval_samples_per_second": 22.013,
|
| 1660 |
-
"eval_steps_per_second": 5.503,
|
| 1661 |
-
"step": 1180
|
| 1662 |
-
},
|
| 1663 |
-
{
|
| 1664 |
-
"epoch": 9.226451927769643,
|
| 1665 |
-
"grad_norm": 169914.25,
|
| 1666 |
-
"learning_rate": 7.031250000000001e-06,
|
| 1667 |
-
"loss": 316.8677,
|
| 1668 |
-
"step": 1190
|
| 1669 |
-
},
|
| 1670 |
-
{
|
| 1671 |
-
"epoch": 9.226451927769643,
|
| 1672 |
-
"eval_runtime": 19.6279,
|
| 1673 |
-
"eval_samples_per_second": 22.01,
|
| 1674 |
-
"eval_steps_per_second": 5.502,
|
| 1675 |
-
"step": 1190
|
| 1676 |
-
},
|
| 1677 |
-
{
|
| 1678 |
-
"epoch": 9.304538799414349,
|
| 1679 |
-
"grad_norm": 145255.390625,
|
| 1680 |
-
"learning_rate": 6.25e-06,
|
| 1681 |
-
"loss": 281.0021,
|
| 1682 |
-
"step": 1200
|
| 1683 |
-
},
|
| 1684 |
-
{
|
| 1685 |
-
"epoch": 9.304538799414349,
|
| 1686 |
-
"eval_runtime": 19.6322,
|
| 1687 |
-
"eval_samples_per_second": 22.005,
|
| 1688 |
-
"eval_steps_per_second": 5.501,
|
| 1689 |
-
"step": 1200
|
| 1690 |
-
},
|
| 1691 |
-
{
|
| 1692 |
-
"epoch": 9.382625671059053,
|
| 1693 |
-
"grad_norm": 77945.0625,
|
| 1694 |
-
"learning_rate": 5.46875e-06,
|
| 1695 |
-
"loss": 225.1448,
|
| 1696 |
-
"step": 1210
|
| 1697 |
-
},
|
| 1698 |
-
{
|
| 1699 |
-
"epoch": 9.382625671059053,
|
| 1700 |
-
"eval_runtime": 19.6173,
|
| 1701 |
-
"eval_samples_per_second": 22.021,
|
| 1702 |
-
"eval_steps_per_second": 5.505,
|
| 1703 |
-
"step": 1210
|
| 1704 |
-
},
|
| 1705 |
-
{
|
| 1706 |
-
"epoch": 9.460712542703758,
|
| 1707 |
-
"grad_norm": 106036.7421875,
|
| 1708 |
-
"learning_rate": 4.6875000000000004e-06,
|
| 1709 |
-
"loss": 260.6584,
|
| 1710 |
-
"step": 1220
|
| 1711 |
-
},
|
| 1712 |
-
{
|
| 1713 |
-
"epoch": 9.460712542703758,
|
| 1714 |
-
"eval_runtime": 19.6147,
|
| 1715 |
-
"eval_samples_per_second": 22.024,
|
| 1716 |
-
"eval_steps_per_second": 5.506,
|
| 1717 |
-
"step": 1220
|
| 1718 |
-
},
|
| 1719 |
-
{
|
| 1720 |
-
"epoch": 9.538799414348462,
|
| 1721 |
-
"grad_norm": 103023.9296875,
|
| 1722 |
-
"learning_rate": 3.90625e-06,
|
| 1723 |
-
"loss": 246.4592,
|
| 1724 |
-
"step": 1230
|
| 1725 |
-
},
|
| 1726 |
-
{
|
| 1727 |
-
"epoch": 9.538799414348462,
|
| 1728 |
-
"eval_runtime": 19.6287,
|
| 1729 |
-
"eval_samples_per_second": 22.009,
|
| 1730 |
-
"eval_steps_per_second": 5.502,
|
| 1731 |
-
"step": 1230
|
| 1732 |
-
},
|
| 1733 |
-
{
|
| 1734 |
-
"epoch": 9.616886285993168,
|
| 1735 |
-
"grad_norm": 69943.5,
|
| 1736 |
-
"learning_rate": 3.125e-06,
|
| 1737 |
-
"loss": 254.6815,
|
| 1738 |
-
"step": 1240
|
| 1739 |
-
},
|
| 1740 |
-
{
|
| 1741 |
-
"epoch": 9.616886285993168,
|
| 1742 |
-
"eval_runtime": 19.6092,
|
| 1743 |
-
"eval_samples_per_second": 22.031,
|
| 1744 |
-
"eval_steps_per_second": 5.508,
|
| 1745 |
-
"step": 1240
|
| 1746 |
-
},
|
| 1747 |
-
{
|
| 1748 |
-
"epoch": 9.694973157637872,
|
| 1749 |
-
"grad_norm": 157203.640625,
|
| 1750 |
-
"learning_rate": 2.3437500000000002e-06,
|
| 1751 |
-
"loss": 247.872,
|
| 1752 |
-
"step": 1250
|
| 1753 |
-
},
|
| 1754 |
-
{
|
| 1755 |
-
"epoch": 9.694973157637872,
|
| 1756 |
-
"eval_runtime": 19.6301,
|
| 1757 |
-
"eval_samples_per_second": 22.007,
|
| 1758 |
-
"eval_steps_per_second": 5.502,
|
| 1759 |
-
"step": 1250
|
| 1760 |
-
},
|
| 1761 |
-
{
|
| 1762 |
-
"epoch": 9.773060029282577,
|
| 1763 |
-
"grad_norm": 40032.71484375,
|
| 1764 |
-
"learning_rate": 1.5625e-06,
|
| 1765 |
-
"loss": 239.199,
|
| 1766 |
-
"step": 1260
|
| 1767 |
-
},
|
| 1768 |
-
{
|
| 1769 |
-
"epoch": 9.773060029282577,
|
| 1770 |
-
"eval_runtime": 19.6402,
|
| 1771 |
-
"eval_samples_per_second": 21.996,
|
| 1772 |
-
"eval_steps_per_second": 5.499,
|
| 1773 |
-
"step": 1260
|
| 1774 |
-
},
|
| 1775 |
-
{
|
| 1776 |
-
"epoch": 9.851146900927281,
|
| 1777 |
-
"grad_norm": 77447.0546875,
|
| 1778 |
-
"learning_rate": 7.8125e-07,
|
| 1779 |
-
"loss": 222.1013,
|
| 1780 |
-
"step": 1270
|
| 1781 |
-
},
|
| 1782 |
-
{
|
| 1783 |
-
"epoch": 9.851146900927281,
|
| 1784 |
-
"eval_runtime": 19.623,
|
| 1785 |
-
"eval_samples_per_second": 22.015,
|
| 1786 |
-
"eval_steps_per_second": 5.504,
|
| 1787 |
-
"step": 1270
|
| 1788 |
-
},
|
| 1789 |
-
{
|
| 1790 |
-
"epoch": 9.929233772571987,
|
| 1791 |
-
"grad_norm": 100750.3046875,
|
| 1792 |
-
"learning_rate": 0.0,
|
| 1793 |
-
"loss": 234.0448,
|
| 1794 |
-
"step": 1280
|
| 1795 |
-
},
|
| 1796 |
-
{
|
| 1797 |
-
"epoch": 9.929233772571987,
|
| 1798 |
-
"eval_runtime": 19.621,
|
| 1799 |
-
"eval_samples_per_second": 22.017,
|
| 1800 |
-
"eval_steps_per_second": 5.504,
|
| 1801 |
-
"step": 1280
|
| 1802 |
}
|
| 1803 |
],
|
| 1804 |
"logging_steps": 10,
|
|
@@ -1813,7 +189,7 @@
|
|
| 1813 |
"should_evaluate": false,
|
| 1814 |
"should_log": false,
|
| 1815 |
"should_save": true,
|
| 1816 |
-
"should_training_stop":
|
| 1817 |
},
|
| 1818 |
"attributes": {}
|
| 1819 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 10,
|
| 6 |
+
"global_step": 129,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.07808687164470474,
|
| 13 |
+
"grad_norm": 562671.25,
|
| 14 |
"learning_rate": 9.921875000000001e-05,
|
| 15 |
+
"loss": 41909.175,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.07808687164470474,
|
| 20 |
+
"eval_runtime": 19.6603,
|
| 21 |
+
"eval_samples_per_second": 21.973,
|
| 22 |
+
"eval_steps_per_second": 5.493,
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 0.15617374328940947,
|
| 27 |
+
"grad_norm": 344896.03125,
|
| 28 |
"learning_rate": 9.84375e-05,
|
| 29 |
+
"loss": 13890.8641,
|
| 30 |
"step": 20
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 0.15617374328940947,
|
| 34 |
+
"eval_runtime": 19.7074,
|
| 35 |
+
"eval_samples_per_second": 21.921,
|
| 36 |
+
"eval_steps_per_second": 5.48,
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 0.2342606149341142,
|
| 41 |
+
"grad_norm": 515524.21875,
|
| 42 |
"learning_rate": 9.765625e-05,
|
| 43 |
+
"loss": 11401.4898,
|
| 44 |
"step": 30
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 0.2342606149341142,
|
| 48 |
+
"eval_runtime": 19.7344,
|
| 49 |
+
"eval_samples_per_second": 21.891,
|
| 50 |
+
"eval_steps_per_second": 5.473,
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 0.31234748657881894,
|
| 55 |
+
"grad_norm": 388208.53125,
|
| 56 |
"learning_rate": 9.687500000000001e-05,
|
| 57 |
+
"loss": 6596.7703,
|
| 58 |
"step": 40
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 0.31234748657881894,
|
| 62 |
+
"eval_runtime": 19.7515,
|
| 63 |
+
"eval_samples_per_second": 21.872,
|
| 64 |
+
"eval_steps_per_second": 5.468,
|
| 65 |
"step": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 0.3904343582235237,
|
| 69 |
+
"grad_norm": 236512.34375,
|
| 70 |
"learning_rate": 9.609375e-05,
|
| 71 |
+
"loss": 3515.6215,
|
| 72 |
"step": 50
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"epoch": 0.3904343582235237,
|
| 76 |
+
"eval_runtime": 19.738,
|
| 77 |
+
"eval_samples_per_second": 21.887,
|
| 78 |
+
"eval_steps_per_second": 5.472,
|
| 79 |
"step": 50
|
| 80 |
},
|
| 81 |
{
|
| 82 |
"epoch": 0.4685212298682284,
|
| 83 |
+
"grad_norm": 694649.5625,
|
| 84 |
"learning_rate": 9.53125e-05,
|
| 85 |
+
"loss": 4275.273,
|
| 86 |
"step": 60
|
| 87 |
},
|
| 88 |
{
|
| 89 |
"epoch": 0.4685212298682284,
|
| 90 |
+
"eval_runtime": 19.7487,
|
| 91 |
+
"eval_samples_per_second": 21.875,
|
| 92 |
+
"eval_steps_per_second": 5.469,
|
| 93 |
"step": 60
|
| 94 |
},
|
| 95 |
{
|
| 96 |
"epoch": 0.5466081015129332,
|
| 97 |
+
"grad_norm": 375526.53125,
|
| 98 |
"learning_rate": 9.453125000000001e-05,
|
| 99 |
+
"loss": 5022.9281,
|
| 100 |
"step": 70
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.5466081015129332,
|
| 104 |
+
"eval_runtime": 19.7307,
|
| 105 |
+
"eval_samples_per_second": 21.895,
|
| 106 |
+
"eval_steps_per_second": 5.474,
|
| 107 |
"step": 70
|
| 108 |
},
|
| 109 |
{
|
| 110 |
"epoch": 0.6246949731576379,
|
| 111 |
+
"grad_norm": 603116.25,
|
| 112 |
"learning_rate": 9.375e-05,
|
| 113 |
+
"loss": 3630.4668,
|
| 114 |
"step": 80
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 0.6246949731576379,
|
| 118 |
+
"eval_runtime": 19.7405,
|
| 119 |
+
"eval_samples_per_second": 21.884,
|
| 120 |
+
"eval_steps_per_second": 5.471,
|
| 121 |
"step": 80
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.7027818448023426,
|
| 125 |
+
"grad_norm": 440137.71875,
|
| 126 |
"learning_rate": 9.296875e-05,
|
| 127 |
+
"loss": 3183.9186,
|
| 128 |
"step": 90
|
| 129 |
},
|
| 130 |
{
|
| 131 |
"epoch": 0.7027818448023426,
|
| 132 |
+
"eval_runtime": 19.7964,
|
| 133 |
+
"eval_samples_per_second": 21.822,
|
| 134 |
+
"eval_steps_per_second": 5.456,
|
| 135 |
"step": 90
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"epoch": 0.7808687164470474,
|
| 139 |
+
"grad_norm": 574069.4375,
|
| 140 |
"learning_rate": 9.21875e-05,
|
| 141 |
+
"loss": 3536.15,
|
| 142 |
"step": 100
|
| 143 |
},
|
| 144 |
{
|
| 145 |
"epoch": 0.7808687164470474,
|
| 146 |
+
"eval_runtime": 19.7759,
|
| 147 |
+
"eval_samples_per_second": 21.845,
|
| 148 |
+
"eval_steps_per_second": 5.461,
|
| 149 |
"step": 100
|
| 150 |
},
|
| 151 |
{
|
| 152 |
"epoch": 0.8589555880917521,
|
| 153 |
+
"grad_norm": 639365.375,
|
| 154 |
"learning_rate": 9.140625e-05,
|
| 155 |
+
"loss": 4252.8562,
|
| 156 |
"step": 110
|
| 157 |
},
|
| 158 |
{
|
| 159 |
"epoch": 0.8589555880917521,
|
| 160 |
+
"eval_runtime": 19.7724,
|
| 161 |
+
"eval_samples_per_second": 21.849,
|
| 162 |
+
"eval_steps_per_second": 5.462,
|
| 163 |
"step": 110
|
| 164 |
},
|
| 165 |
{
|
| 166 |
"epoch": 0.9370424597364568,
|
| 167 |
+
"grad_norm": 398296.34375,
|
| 168 |
"learning_rate": 9.062500000000001e-05,
|
| 169 |
+
"loss": 3483.168,
|
| 170 |
"step": 120
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 0.9370424597364568,
|
| 174 |
+
"eval_runtime": 19.7473,
|
| 175 |
+
"eval_samples_per_second": 21.876,
|
| 176 |
+
"eval_steps_per_second": 5.469,
|
| 177 |
"step": 120
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 178 |
}
|
| 179 |
],
|
| 180 |
"logging_steps": 10,
|
|
|
|
| 189 |
"should_evaluate": false,
|
| 190 |
"should_log": false,
|
| 191 |
"should_save": true,
|
| 192 |
+
"should_training_stop": false
|
| 193 |
},
|
| 194 |
"attributes": {}
|
| 195 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5368
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00e1815a75e87f99681c46c6d470e12a833e128941d3a562bbcb63c47c459934
|
| 3 |
size 5368
|