Training in progress, epoch 0, 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 2610104820
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ede89cd7420b342554cf586111ef386bc4803fe1942c8c752e713c75eb639884
|
| 3 |
size 2610104820
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5210004271
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a83ac9d504c4f5a9d0b3ec6c9d0ab931281b9695216ad64ae558fdd4e9634d9
|
| 3 |
size 5210004271
|
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 sha256:9080aea5181d3066ab765d04bc9819f089e9674161d5e56c8bf2b7c839212160
|
| 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 sha256:3fd6739d9b468767c726a8685cd4457152a8323bb7a81cf6908a01dd282a18e8
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,342 +1,27 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
-
"best_model_checkpoint": "mgh6/HTH_prob/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 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": 0.9960454139558618,
|
| 13 |
-
"grad_norm": 0.
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss": 1.
|
| 16 |
"step": 244
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.9960454139558618,
|
| 20 |
-
"eval_loss": 1.
|
| 21 |
-
"eval_runtime": 25.
|
| 22 |
-
"eval_samples_per_second": 32.
|
| 23 |
-
"eval_steps_per_second": 16.
|
| 24 |
"step": 244
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 1.9960454139558617,
|
| 28 |
-
"grad_norm": 0.37684935331344604,
|
| 29 |
-
"learning_rate": 9.666666666666667e-05,
|
| 30 |
-
"loss": 1.0601,
|
| 31 |
-
"step": 488
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 1.9960454139558617,
|
| 35 |
-
"eval_loss": 1.0078188180923462,
|
| 36 |
-
"eval_runtime": 25.3488,
|
| 37 |
-
"eval_samples_per_second": 32.585,
|
| 38 |
-
"eval_steps_per_second": 16.293,
|
| 39 |
-
"step": 488
|
| 40 |
-
},
|
| 41 |
-
{
|
| 42 |
-
"epoch": 2.9960454139558617,
|
| 43 |
-
"grad_norm": 0.22615128755569458,
|
| 44 |
-
"learning_rate": 9.5e-05,
|
| 45 |
-
"loss": 0.9323,
|
| 46 |
-
"step": 732
|
| 47 |
-
},
|
| 48 |
-
{
|
| 49 |
-
"epoch": 2.9960454139558617,
|
| 50 |
-
"eval_loss": 0.9372721314430237,
|
| 51 |
-
"eval_runtime": 25.365,
|
| 52 |
-
"eval_samples_per_second": 32.565,
|
| 53 |
-
"eval_steps_per_second": 16.282,
|
| 54 |
-
"step": 732
|
| 55 |
-
},
|
| 56 |
-
{
|
| 57 |
-
"epoch": 3.9960454139558617,
|
| 58 |
-
"grad_norm": 0.15589652955532074,
|
| 59 |
-
"learning_rate": 9.333333333333334e-05,
|
| 60 |
-
"loss": 0.8778,
|
| 61 |
-
"step": 976
|
| 62 |
-
},
|
| 63 |
-
{
|
| 64 |
-
"epoch": 3.9960454139558617,
|
| 65 |
-
"eval_loss": 0.9012424349784851,
|
| 66 |
-
"eval_runtime": 25.3688,
|
| 67 |
-
"eval_samples_per_second": 32.56,
|
| 68 |
-
"eval_steps_per_second": 16.28,
|
| 69 |
-
"step": 976
|
| 70 |
-
},
|
| 71 |
-
{
|
| 72 |
-
"epoch": 4.996045413955862,
|
| 73 |
-
"grad_norm": 0.11001910269260406,
|
| 74 |
-
"learning_rate": 9.166666666666667e-05,
|
| 75 |
-
"loss": 0.854,
|
| 76 |
-
"step": 1220
|
| 77 |
-
},
|
| 78 |
-
{
|
| 79 |
-
"epoch": 4.996045413955862,
|
| 80 |
-
"eval_loss": 0.881626546382904,
|
| 81 |
-
"eval_runtime": 25.3578,
|
| 82 |
-
"eval_samples_per_second": 32.574,
|
| 83 |
-
"eval_steps_per_second": 16.287,
|
| 84 |
-
"step": 1220
|
| 85 |
-
},
|
| 86 |
-
{
|
| 87 |
-
"epoch": 5.996045413955862,
|
| 88 |
-
"grad_norm": 0.09984524548053741,
|
| 89 |
-
"learning_rate": 9e-05,
|
| 90 |
-
"loss": 0.8412,
|
| 91 |
-
"step": 1464
|
| 92 |
-
},
|
| 93 |
-
{
|
| 94 |
-
"epoch": 5.996045413955862,
|
| 95 |
-
"eval_loss": 0.8756476044654846,
|
| 96 |
-
"eval_runtime": 25.3658,
|
| 97 |
-
"eval_samples_per_second": 32.564,
|
| 98 |
-
"eval_steps_per_second": 16.282,
|
| 99 |
-
"step": 1464
|
| 100 |
-
},
|
| 101 |
-
{
|
| 102 |
-
"epoch": 6.996045413955862,
|
| 103 |
-
"grad_norm": 0.08100161701440811,
|
| 104 |
-
"learning_rate": 8.833333333333333e-05,
|
| 105 |
-
"loss": 0.8331,
|
| 106 |
-
"step": 1708
|
| 107 |
-
},
|
| 108 |
-
{
|
| 109 |
-
"epoch": 6.996045413955862,
|
| 110 |
-
"eval_loss": 0.8662193417549133,
|
| 111 |
-
"eval_runtime": 25.367,
|
| 112 |
-
"eval_samples_per_second": 32.562,
|
| 113 |
-
"eval_steps_per_second": 16.281,
|
| 114 |
-
"step": 1708
|
| 115 |
-
},
|
| 116 |
-
{
|
| 117 |
-
"epoch": 7.996045413955862,
|
| 118 |
-
"grad_norm": 0.07314834743738174,
|
| 119 |
-
"learning_rate": 8.666666666666667e-05,
|
| 120 |
-
"loss": 0.828,
|
| 121 |
-
"step": 1952
|
| 122 |
-
},
|
| 123 |
-
{
|
| 124 |
-
"epoch": 7.996045413955862,
|
| 125 |
-
"eval_loss": 0.8645371198654175,
|
| 126 |
-
"eval_runtime": 25.3405,
|
| 127 |
-
"eval_samples_per_second": 32.596,
|
| 128 |
-
"eval_steps_per_second": 16.298,
|
| 129 |
-
"step": 1952
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 8.996045413955862,
|
| 133 |
-
"grad_norm": 0.06718147546052933,
|
| 134 |
-
"learning_rate": 8.5e-05,
|
| 135 |
-
"loss": 0.8245,
|
| 136 |
-
"step": 2196
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 8.996045413955862,
|
| 140 |
-
"eval_loss": 0.8601691722869873,
|
| 141 |
-
"eval_runtime": 25.3375,
|
| 142 |
-
"eval_samples_per_second": 32.6,
|
| 143 |
-
"eval_steps_per_second": 16.3,
|
| 144 |
-
"step": 2196
|
| 145 |
-
},
|
| 146 |
-
{
|
| 147 |
-
"epoch": 9.996045413955862,
|
| 148 |
-
"grad_norm": 0.059008464217185974,
|
| 149 |
-
"learning_rate": 8.333333333333334e-05,
|
| 150 |
-
"loss": 0.822,
|
| 151 |
-
"step": 2440
|
| 152 |
-
},
|
| 153 |
-
{
|
| 154 |
-
"epoch": 9.996045413955862,
|
| 155 |
-
"eval_loss": 0.8585366010665894,
|
| 156 |
-
"eval_runtime": 25.3797,
|
| 157 |
-
"eval_samples_per_second": 32.546,
|
| 158 |
-
"eval_steps_per_second": 16.273,
|
| 159 |
-
"step": 2440
|
| 160 |
-
},
|
| 161 |
-
{
|
| 162 |
-
"epoch": 10.996045413955862,
|
| 163 |
-
"grad_norm": 0.05849480628967285,
|
| 164 |
-
"learning_rate": 8.166666666666667e-05,
|
| 165 |
-
"loss": 0.82,
|
| 166 |
-
"step": 2684
|
| 167 |
-
},
|
| 168 |
-
{
|
| 169 |
-
"epoch": 10.996045413955862,
|
| 170 |
-
"eval_loss": 0.8567091226577759,
|
| 171 |
-
"eval_runtime": 25.3556,
|
| 172 |
-
"eval_samples_per_second": 32.577,
|
| 173 |
-
"eval_steps_per_second": 16.288,
|
| 174 |
-
"step": 2684
|
| 175 |
-
},
|
| 176 |
-
{
|
| 177 |
-
"epoch": 11.996045413955862,
|
| 178 |
-
"grad_norm": 0.059530675411224365,
|
| 179 |
-
"learning_rate": 8e-05,
|
| 180 |
-
"loss": 0.8186,
|
| 181 |
-
"step": 2928
|
| 182 |
-
},
|
| 183 |
-
{
|
| 184 |
-
"epoch": 11.996045413955862,
|
| 185 |
-
"eval_loss": 0.8564208745956421,
|
| 186 |
-
"eval_runtime": 25.3424,
|
| 187 |
-
"eval_samples_per_second": 32.594,
|
| 188 |
-
"eval_steps_per_second": 16.297,
|
| 189 |
-
"step": 2928
|
| 190 |
-
},
|
| 191 |
-
{
|
| 192 |
-
"epoch": 12.996045413955862,
|
| 193 |
-
"grad_norm": 0.06416182219982147,
|
| 194 |
-
"learning_rate": 7.833333333333333e-05,
|
| 195 |
-
"loss": 0.8173,
|
| 196 |
-
"step": 3172
|
| 197 |
-
},
|
| 198 |
-
{
|
| 199 |
-
"epoch": 12.996045413955862,
|
| 200 |
-
"eval_loss": 0.8540862798690796,
|
| 201 |
-
"eval_runtime": 25.3529,
|
| 202 |
-
"eval_samples_per_second": 32.58,
|
| 203 |
-
"eval_steps_per_second": 16.29,
|
| 204 |
-
"step": 3172
|
| 205 |
-
},
|
| 206 |
-
{
|
| 207 |
-
"epoch": 13.996045413955862,
|
| 208 |
-
"grad_norm": 0.05353016406297684,
|
| 209 |
-
"learning_rate": 7.666666666666667e-05,
|
| 210 |
-
"loss": 0.8171,
|
| 211 |
-
"step": 3416
|
| 212 |
-
},
|
| 213 |
-
{
|
| 214 |
-
"epoch": 13.996045413955862,
|
| 215 |
-
"eval_loss": 0.8535267114639282,
|
| 216 |
-
"eval_runtime": 25.3125,
|
| 217 |
-
"eval_samples_per_second": 32.632,
|
| 218 |
-
"eval_steps_per_second": 16.316,
|
| 219 |
-
"step": 3416
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 14.996045413955862,
|
| 223 |
-
"grad_norm": 0.04900681599974632,
|
| 224 |
-
"learning_rate": 7.500000000000001e-05,
|
| 225 |
-
"loss": 0.8151,
|
| 226 |
-
"step": 3660
|
| 227 |
-
},
|
| 228 |
-
{
|
| 229 |
-
"epoch": 14.996045413955862,
|
| 230 |
-
"eval_loss": 0.8519299626350403,
|
| 231 |
-
"eval_runtime": 25.3532,
|
| 232 |
-
"eval_samples_per_second": 32.58,
|
| 233 |
-
"eval_steps_per_second": 16.29,
|
| 234 |
-
"step": 3660
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 15.996045413955862,
|
| 238 |
-
"grad_norm": 0.0492498017847538,
|
| 239 |
-
"learning_rate": 7.333333333333333e-05,
|
| 240 |
-
"loss": 0.8143,
|
| 241 |
-
"step": 3904
|
| 242 |
-
},
|
| 243 |
-
{
|
| 244 |
-
"epoch": 15.996045413955862,
|
| 245 |
-
"eval_loss": 0.8498228788375854,
|
| 246 |
-
"eval_runtime": 25.3097,
|
| 247 |
-
"eval_samples_per_second": 32.636,
|
| 248 |
-
"eval_steps_per_second": 16.318,
|
| 249 |
-
"step": 3904
|
| 250 |
-
},
|
| 251 |
-
{
|
| 252 |
-
"epoch": 16.996045413955862,
|
| 253 |
-
"grad_norm": 0.04555810987949371,
|
| 254 |
-
"learning_rate": 7.166666666666667e-05,
|
| 255 |
-
"loss": 0.8136,
|
| 256 |
-
"step": 4148
|
| 257 |
-
},
|
| 258 |
-
{
|
| 259 |
-
"epoch": 16.996045413955862,
|
| 260 |
-
"eval_loss": 0.8518642783164978,
|
| 261 |
-
"eval_runtime": 25.326,
|
| 262 |
-
"eval_samples_per_second": 32.615,
|
| 263 |
-
"eval_steps_per_second": 16.307,
|
| 264 |
-
"step": 4148
|
| 265 |
-
},
|
| 266 |
-
{
|
| 267 |
-
"epoch": 17.996045413955862,
|
| 268 |
-
"grad_norm": 0.03729957342147827,
|
| 269 |
-
"learning_rate": 7e-05,
|
| 270 |
-
"loss": 0.8134,
|
| 271 |
-
"step": 4392
|
| 272 |
-
},
|
| 273 |
-
{
|
| 274 |
-
"epoch": 17.996045413955862,
|
| 275 |
-
"eval_loss": 0.8507369756698608,
|
| 276 |
-
"eval_runtime": 25.3562,
|
| 277 |
-
"eval_samples_per_second": 32.576,
|
| 278 |
-
"eval_steps_per_second": 16.288,
|
| 279 |
-
"step": 4392
|
| 280 |
-
},
|
| 281 |
-
{
|
| 282 |
-
"epoch": 18.996045413955862,
|
| 283 |
-
"grad_norm": 0.037404902279376984,
|
| 284 |
-
"learning_rate": 6.833333333333333e-05,
|
| 285 |
-
"loss": 0.8124,
|
| 286 |
-
"step": 4636
|
| 287 |
-
},
|
| 288 |
-
{
|
| 289 |
-
"epoch": 18.996045413955862,
|
| 290 |
-
"eval_loss": 0.8495596647262573,
|
| 291 |
-
"eval_runtime": 25.3504,
|
| 292 |
-
"eval_samples_per_second": 32.583,
|
| 293 |
-
"eval_steps_per_second": 16.292,
|
| 294 |
-
"step": 4636
|
| 295 |
-
},
|
| 296 |
-
{
|
| 297 |
-
"epoch": 19.996045413955862,
|
| 298 |
-
"grad_norm": 0.045253317803144455,
|
| 299 |
-
"learning_rate": 6.666666666666667e-05,
|
| 300 |
-
"loss": 0.8117,
|
| 301 |
-
"step": 4880
|
| 302 |
-
},
|
| 303 |
-
{
|
| 304 |
-
"epoch": 19.996045413955862,
|
| 305 |
-
"eval_loss": 0.8522358536720276,
|
| 306 |
-
"eval_runtime": 25.356,
|
| 307 |
-
"eval_samples_per_second": 32.576,
|
| 308 |
-
"eval_steps_per_second": 16.288,
|
| 309 |
-
"step": 4880
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 20.996045413955862,
|
| 313 |
-
"grad_norm": 0.04286725074052811,
|
| 314 |
-
"learning_rate": 6.500000000000001e-05,
|
| 315 |
-
"loss": 0.8133,
|
| 316 |
-
"step": 5124
|
| 317 |
-
},
|
| 318 |
-
{
|
| 319 |
-
"epoch": 20.996045413955862,
|
| 320 |
-
"eval_loss": 0.8517967462539673,
|
| 321 |
-
"eval_runtime": 25.3545,
|
| 322 |
-
"eval_samples_per_second": 32.578,
|
| 323 |
-
"eval_steps_per_second": 16.289,
|
| 324 |
-
"step": 5124
|
| 325 |
-
},
|
| 326 |
-
{
|
| 327 |
-
"epoch": 21.996045413955862,
|
| 328 |
-
"grad_norm": 0.033235229551792145,
|
| 329 |
-
"learning_rate": 6.333333333333333e-05,
|
| 330 |
-
"loss": 0.8128,
|
| 331 |
-
"step": 5368
|
| 332 |
-
},
|
| 333 |
-
{
|
| 334 |
-
"epoch": 21.996045413955862,
|
| 335 |
-
"eval_loss": 0.8513291478157043,
|
| 336 |
-
"eval_runtime": 25.3035,
|
| 337 |
-
"eval_samples_per_second": 32.644,
|
| 338 |
-
"eval_steps_per_second": 16.322,
|
| 339 |
-
"step": 5368
|
| 340 |
}
|
| 341 |
],
|
| 342 |
"logging_steps": 500,
|
|
@@ -351,7 +36,7 @@
|
|
| 351 |
"early_stopping_threshold": 0.0
|
| 352 |
},
|
| 353 |
"attributes": {
|
| 354 |
-
"early_stopping_patience_counter":
|
| 355 |
}
|
| 356 |
},
|
| 357 |
"TrainerControl": {
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.3716533184051514,
|
| 3 |
+
"best_model_checkpoint": "mgh6/HTH_prob/checkpoint-244",
|
| 4 |
+
"epoch": 0.9960454139558618,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 244,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.9960454139558618,
|
| 13 |
+
"grad_norm": 0.04875887930393219,
|
| 14 |
+
"learning_rate": 0.0009833333333333332,
|
| 15 |
+
"loss": 1.3837,
|
| 16 |
"step": 244
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.9960454139558618,
|
| 20 |
+
"eval_loss": 1.3716533184051514,
|
| 21 |
+
"eval_runtime": 25.1667,
|
| 22 |
+
"eval_samples_per_second": 32.821,
|
| 23 |
+
"eval_steps_per_second": 16.411,
|
| 24 |
"step": 244
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
}
|
| 26 |
],
|
| 27 |
"logging_steps": 500,
|
|
|
|
| 36 |
"early_stopping_threshold": 0.0
|
| 37 |
},
|
| 38 |
"attributes": {
|
| 39 |
+
"early_stopping_patience_counter": 0
|
| 40 |
}
|
| 41 |
},
|
| 42 |
"TrainerControl": {
|
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 sha256:667cf2698d9700cfae15eb710eb29a7fe4af2b3f185b9513c6453b73cf79787c
|
| 3 |
size 5368
|