Commit ·
96d3218
1
Parent(s): a55b637
Upload 12 files
Browse files- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scaler.pt +1 -1
- scheduler.pt +1 -1
- trainer_state.json +24 -357
- training_args.bin +1 -1
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 941352733
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9439a91d77f577a769cc49680171045b9d1d75ad5bcb4e8509c4184ef9786f0
|
| 3 |
size 941352733
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 470689133
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04c42607ceaf9dfb9bfdde750cb68ffce0091526954ed3c098d4e53e78b15728
|
| 3 |
size 470689133
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:018ee511d461eca50aae3f96fd675ea399ab9a8fb2556ff297b2bb456934d1ef
|
| 3 |
size 14503
|
scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 559
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c1274f79c9e460ae6bf3e5f96099845a5976fc8e53216f53bd1006019f4d32d
|
| 3 |
size 559
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51175e113ec240e93413ad240ccb51d582b77522feac43f869dac2af5849fece
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,373 +1,40 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "safety_instruct_detect/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
-
"epoch":
|
| 12 |
-
"
|
| 13 |
-
"
|
| 14 |
-
"
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
"
|
| 18 |
-
"learning_rate": 1.965793254429774e-05,
|
| 19 |
-
"loss": 0.5267,
|
| 20 |
-
"step": 1000
|
| 21 |
-
},
|
| 22 |
-
{
|
| 23 |
-
"epoch": 0.05,
|
| 24 |
-
"learning_rate": 1.9486898816446604e-05,
|
| 25 |
-
"loss": 0.5018,
|
| 26 |
-
"step": 1500
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.07,
|
| 30 |
-
"learning_rate": 1.9315865088595472e-05,
|
| 31 |
-
"loss": 0.4908,
|
| 32 |
-
"step": 2000
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.09,
|
| 36 |
-
"learning_rate": 1.914483136074434e-05,
|
| 37 |
-
"loss": 0.4837,
|
| 38 |
-
"step": 2500
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.1,
|
| 42 |
-
"learning_rate": 1.8973797632893206e-05,
|
| 43 |
-
"loss": 0.4864,
|
| 44 |
-
"step": 3000
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.12,
|
| 48 |
-
"learning_rate": 1.8802763905042078e-05,
|
| 49 |
-
"loss": 0.4747,
|
| 50 |
-
"step": 3500
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.14,
|
| 54 |
-
"learning_rate": 1.8631730177190943e-05,
|
| 55 |
-
"loss": 0.4787,
|
| 56 |
-
"step": 4000
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.15,
|
| 60 |
-
"learning_rate": 1.846103851679551e-05,
|
| 61 |
-
"loss": 0.4779,
|
| 62 |
-
"step": 4500
|
| 63 |
-
},
|
| 64 |
-
{
|
| 65 |
-
"epoch": 0.17,
|
| 66 |
-
"learning_rate": 1.8290004788944384e-05,
|
| 67 |
-
"loss": 0.4713,
|
| 68 |
-
"step": 5000
|
| 69 |
-
},
|
| 70 |
-
{
|
| 71 |
-
"epoch": 0.19,
|
| 72 |
-
"learning_rate": 1.8119313128548952e-05,
|
| 73 |
-
"loss": 0.4745,
|
| 74 |
-
"step": 5500
|
| 75 |
-
},
|
| 76 |
-
{
|
| 77 |
-
"epoch": 0.21,
|
| 78 |
-
"learning_rate": 1.7948279400697817e-05,
|
| 79 |
-
"loss": 0.4712,
|
| 80 |
-
"step": 6000
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.22,
|
| 84 |
-
"learning_rate": 1.7777245672846686e-05,
|
| 85 |
-
"loss": 0.4757,
|
| 86 |
-
"step": 6500
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.24,
|
| 90 |
-
"learning_rate": 1.7606211944995554e-05,
|
| 91 |
-
"loss": 0.4725,
|
| 92 |
-
"step": 7000
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 0.26,
|
| 96 |
-
"learning_rate": 1.7435178217144423e-05,
|
| 97 |
-
"loss": 0.4649,
|
| 98 |
-
"step": 7500
|
| 99 |
-
},
|
| 100 |
-
{
|
| 101 |
-
"epoch": 0.27,
|
| 102 |
-
"learning_rate": 1.726414448929329e-05,
|
| 103 |
-
"loss": 0.4686,
|
| 104 |
-
"step": 8000
|
| 105 |
-
},
|
| 106 |
-
{
|
| 107 |
-
"epoch": 0.29,
|
| 108 |
-
"learning_rate": 1.7093110761442157e-05,
|
| 109 |
-
"loss": 0.4661,
|
| 110 |
-
"step": 8500
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 0.31,
|
| 114 |
-
"learning_rate": 1.6922077033591025e-05,
|
| 115 |
-
"loss": 0.4642,
|
| 116 |
-
"step": 9000
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 0.32,
|
| 120 |
-
"learning_rate": 1.6751385373195594e-05,
|
| 121 |
-
"loss": 0.4638,
|
| 122 |
-
"step": 9500
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.34,
|
| 126 |
-
"learning_rate": 1.6580693712800166e-05,
|
| 127 |
-
"loss": 0.4485,
|
| 128 |
-
"step": 10000
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.36,
|
| 132 |
-
"learning_rate": 1.6409659984949035e-05,
|
| 133 |
-
"loss": 0.4648,
|
| 134 |
-
"step": 10500
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 0.38,
|
| 138 |
-
"learning_rate": 1.62386262570979e-05,
|
| 139 |
-
"loss": 0.4565,
|
| 140 |
-
"step": 11000
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 0.39,
|
| 144 |
-
"learning_rate": 1.6067592529246768e-05,
|
| 145 |
-
"loss": 0.459,
|
| 146 |
-
"step": 11500
|
| 147 |
-
},
|
| 148 |
-
{
|
| 149 |
-
"epoch": 0.41,
|
| 150 |
-
"learning_rate": 1.5896558801395637e-05,
|
| 151 |
-
"loss": 0.4559,
|
| 152 |
-
"step": 12000
|
| 153 |
-
},
|
| 154 |
-
{
|
| 155 |
-
"epoch": 0.43,
|
| 156 |
-
"learning_rate": 1.5725525073544505e-05,
|
| 157 |
-
"loss": 0.45,
|
| 158 |
-
"step": 12500
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 0.44,
|
| 162 |
-
"learning_rate": 1.5554833413149074e-05,
|
| 163 |
-
"loss": 0.4533,
|
| 164 |
-
"step": 13000
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.46,
|
| 168 |
-
"learning_rate": 1.5383799685297943e-05,
|
| 169 |
-
"loss": 0.4578,
|
| 170 |
-
"step": 13500
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.48,
|
| 174 |
-
"learning_rate": 1.521276595744681e-05,
|
| 175 |
-
"loss": 0.4577,
|
| 176 |
-
"step": 14000
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 0.5,
|
| 180 |
-
"learning_rate": 1.5041732229595678e-05,
|
| 181 |
-
"loss": 0.4552,
|
| 182 |
-
"step": 14500
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 0.51,
|
| 186 |
-
"learning_rate": 1.4870698501744545e-05,
|
| 187 |
-
"loss": 0.4551,
|
| 188 |
-
"step": 15000
|
| 189 |
-
},
|
| 190 |
-
{
|
| 191 |
-
"epoch": 0.53,
|
| 192 |
-
"learning_rate": 1.4700006841349115e-05,
|
| 193 |
-
"loss": 0.4478,
|
| 194 |
-
"step": 15500
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 0.55,
|
| 198 |
-
"learning_rate": 1.4528973113497984e-05,
|
| 199 |
-
"loss": 0.4496,
|
| 200 |
-
"step": 16000
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 0.56,
|
| 204 |
-
"learning_rate": 1.435793938564685e-05,
|
| 205 |
-
"loss": 0.4499,
|
| 206 |
-
"step": 16500
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 0.58,
|
| 210 |
-
"learning_rate": 1.4186905657795717e-05,
|
| 211 |
-
"loss": 0.4403,
|
| 212 |
-
"step": 17000
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 0.6,
|
| 216 |
-
"learning_rate": 1.4015871929944587e-05,
|
| 217 |
-
"loss": 0.4492,
|
| 218 |
-
"step": 17500
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 0.62,
|
| 222 |
-
"learning_rate": 1.3844838202093454e-05,
|
| 223 |
-
"loss": 0.4447,
|
| 224 |
-
"step": 18000
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 0.63,
|
| 228 |
-
"learning_rate": 1.3673804474242323e-05,
|
| 229 |
-
"loss": 0.4442,
|
| 230 |
-
"step": 18500
|
| 231 |
-
},
|
| 232 |
-
{
|
| 233 |
-
"epoch": 0.65,
|
| 234 |
-
"learning_rate": 1.350277074639119e-05,
|
| 235 |
-
"loss": 0.4464,
|
| 236 |
-
"step": 19000
|
| 237 |
-
},
|
| 238 |
-
{
|
| 239 |
-
"epoch": 0.67,
|
| 240 |
-
"learning_rate": 1.333207908599576e-05,
|
| 241 |
-
"loss": 0.4445,
|
| 242 |
-
"step": 19500
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 0.68,
|
| 246 |
-
"learning_rate": 1.3161045358144627e-05,
|
| 247 |
-
"loss": 0.4458,
|
| 248 |
-
"step": 20000
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 0.7,
|
| 252 |
-
"learning_rate": 1.2990011630293495e-05,
|
| 253 |
-
"loss": 0.4532,
|
| 254 |
-
"step": 20500
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 0.72,
|
| 258 |
-
"learning_rate": 1.2818977902442362e-05,
|
| 259 |
-
"loss": 0.4497,
|
| 260 |
-
"step": 21000
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 0.74,
|
| 264 |
-
"learning_rate": 1.2648286242046933e-05,
|
| 265 |
-
"loss": 0.4424,
|
| 266 |
-
"step": 21500
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 0.75,
|
| 270 |
-
"learning_rate": 1.2477252514195801e-05,
|
| 271 |
-
"loss": 0.4441,
|
| 272 |
-
"step": 22000
|
| 273 |
-
},
|
| 274 |
-
{
|
| 275 |
-
"epoch": 0.77,
|
| 276 |
-
"learning_rate": 1.2306218786344668e-05,
|
| 277 |
-
"loss": 0.4388,
|
| 278 |
-
"step": 22500
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 0.79,
|
| 282 |
-
"learning_rate": 1.2135185058493535e-05,
|
| 283 |
-
"loss": 0.4494,
|
| 284 |
-
"step": 23000
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 0.8,
|
| 288 |
-
"learning_rate": 1.1964493398098105e-05,
|
| 289 |
-
"loss": 0.4491,
|
| 290 |
-
"step": 23500
|
| 291 |
-
},
|
| 292 |
-
{
|
| 293 |
-
"epoch": 0.82,
|
| 294 |
-
"learning_rate": 1.1793459670246974e-05,
|
| 295 |
-
"loss": 0.4487,
|
| 296 |
-
"step": 24000
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 0.84,
|
| 300 |
-
"learning_rate": 1.162242594239584e-05,
|
| 301 |
-
"loss": 0.4427,
|
| 302 |
-
"step": 24500
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 0.86,
|
| 306 |
-
"learning_rate": 1.145139221454471e-05,
|
| 307 |
-
"loss": 0.4408,
|
| 308 |
-
"step": 25000
|
| 309 |
-
},
|
| 310 |
-
{
|
| 311 |
-
"epoch": 0.87,
|
| 312 |
-
"learning_rate": 1.128070055414928e-05,
|
| 313 |
-
"loss": 0.4492,
|
| 314 |
-
"step": 25500
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 0.89,
|
| 318 |
-
"learning_rate": 1.1109666826298147e-05,
|
| 319 |
-
"loss": 0.4495,
|
| 320 |
-
"step": 26000
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 0.91,
|
| 324 |
-
"learning_rate": 1.0938633098447015e-05,
|
| 325 |
-
"loss": 0.4483,
|
| 326 |
-
"step": 26500
|
| 327 |
-
},
|
| 328 |
-
{
|
| 329 |
-
"epoch": 0.92,
|
| 330 |
-
"learning_rate": 1.0767599370595882e-05,
|
| 331 |
-
"loss": 0.4434,
|
| 332 |
-
"step": 27000
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 0.94,
|
| 336 |
-
"learning_rate": 1.0596565642744749e-05,
|
| 337 |
-
"loss": 0.4378,
|
| 338 |
-
"step": 27500
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 0.96,
|
| 342 |
-
"learning_rate": 1.042587398234932e-05,
|
| 343 |
-
"loss": 0.4519,
|
| 344 |
-
"step": 28000
|
| 345 |
-
},
|
| 346 |
-
{
|
| 347 |
-
"epoch": 0.97,
|
| 348 |
-
"learning_rate": 1.0254840254498188e-05,
|
| 349 |
-
"loss": 0.441,
|
| 350 |
-
"step": 28500
|
| 351 |
},
|
| 352 |
{
|
| 353 |
-
"epoch":
|
| 354 |
-
"learning_rate":
|
| 355 |
-
"loss": 0.
|
| 356 |
-
"step":
|
| 357 |
},
|
| 358 |
{
|
| 359 |
-
"epoch":
|
| 360 |
-
"eval_accuracy": 0.
|
| 361 |
-
"eval_loss": 0.
|
| 362 |
-
"eval_runtime":
|
| 363 |
-
"eval_samples_per_second":
|
| 364 |
-
"eval_steps_per_second":
|
| 365 |
-
"step":
|
| 366 |
}
|
| 367 |
],
|
| 368 |
-
"max_steps":
|
| 369 |
"num_train_epochs": 2,
|
| 370 |
-
"total_flos":
|
| 371 |
"trial_name": null,
|
| 372 |
"trial_params": null
|
| 373 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.09600851684808731,
|
| 3 |
+
"best_model_checkpoint": "safety_instruct_detect/checkpoint-686",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 686,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 1.0,
|
| 12 |
+
"eval_accuracy": 0.9537414965986395,
|
| 13 |
+
"eval_loss": 0.14035563170909882,
|
| 14 |
+
"eval_runtime": 35.4642,
|
| 15 |
+
"eval_samples_per_second": 145.076,
|
| 16 |
+
"eval_steps_per_second": 4.145,
|
| 17 |
+
"step": 343
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
},
|
| 19 |
{
|
| 20 |
+
"epoch": 1.46,
|
| 21 |
+
"learning_rate": 5.451895043731778e-06,
|
| 22 |
+
"loss": 0.2205,
|
| 23 |
+
"step": 500
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 2.0,
|
| 27 |
+
"eval_accuracy": 0.9706511175898931,
|
| 28 |
+
"eval_loss": 0.09600851684808731,
|
| 29 |
+
"eval_runtime": 35.524,
|
| 30 |
+
"eval_samples_per_second": 144.832,
|
| 31 |
+
"eval_steps_per_second": 4.138,
|
| 32 |
+
"step": 686
|
| 33 |
}
|
| 34 |
],
|
| 35 |
+
"max_steps": 686,
|
| 36 |
"num_train_epochs": 2,
|
| 37 |
+
"total_flos": 1291799489699544.0,
|
| 38 |
"trial_name": null,
|
| 39 |
"trial_params": null
|
| 40 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3311
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98d4deafadf31d0438ab604f19beb610e72db0a0498ab53564cf1a2752fccfaf
|
| 3 |
size 3311
|