Commit
•
5d2725a
1
Parent(s):
7cf79db
Training in progress, step 20, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
|
@@ -20,9 +20,9 @@
|
|
| 20 |
"rank_pattern": {},
|
| 21 |
"revision": null,
|
| 22 |
"target_modules": [
|
| 23 |
-
"word_embeddings",
|
| 24 |
"dense_h_to_4h",
|
| 25 |
-
"query_key_valuelm_head"
|
|
|
|
| 26 |
],
|
| 27 |
"task_type": "CAUSAL_LM",
|
| 28 |
"use_dora": false,
|
|
|
|
| 20 |
"rank_pattern": {},
|
| 21 |
"revision": null,
|
| 22 |
"target_modules": [
|
|
|
|
| 23 |
"dense_h_to_4h",
|
| 24 |
+
"query_key_valuelm_head",
|
| 25 |
+
"word_embeddings"
|
| 26 |
],
|
| 27 |
"task_type": "CAUSAL_LM",
|
| 28 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13982248
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b1631171f6439b6a14b2e153dcf540167f86737bb2c711c2b140dea1fe25d6b1
|
| 3 |
size 13982248
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 7062522
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7260ad99737e832312585651bb9f4fa31542e715dde742931df691aee36b0dc5
|
| 3 |
size 7062522
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9eb03dd380d1de2cd7fa7a5f5e88bd2d0511a2b1ef619fa3362c6bbf618b480e
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d0351811b292a925d9e90fecda402ef74b5027ecf82c420e29489a6a985de561
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,3310 +1,27 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 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": 0.0,
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss": 4.
|
| 16 |
"step": 20
|
| 17 |
-
},
|
| 18 |
-
{
|
| 19 |
-
"epoch": 0.0,
|
| 20 |
-
"grad_norm": 39.39016342163086,
|
| 21 |
-
"learning_rate": 2e-06,
|
| 22 |
-
"loss": 4.4077,
|
| 23 |
-
"step": 40
|
| 24 |
-
},
|
| 25 |
-
{
|
| 26 |
-
"epoch": 0.0,
|
| 27 |
-
"grad_norm": 54.24020767211914,
|
| 28 |
-
"learning_rate": 3e-06,
|
| 29 |
-
"loss": 4.4807,
|
| 30 |
-
"step": 60
|
| 31 |
-
},
|
| 32 |
-
{
|
| 33 |
-
"epoch": 0.01,
|
| 34 |
-
"grad_norm": 30.161609649658203,
|
| 35 |
-
"learning_rate": 4e-06,
|
| 36 |
-
"loss": 4.5756,
|
| 37 |
-
"step": 80
|
| 38 |
-
},
|
| 39 |
-
{
|
| 40 |
-
"epoch": 0.01,
|
| 41 |
-
"grad_norm": 40.131675720214844,
|
| 42 |
-
"learning_rate": 4.9999999999999996e-06,
|
| 43 |
-
"loss": 4.4352,
|
| 44 |
-
"step": 100
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.01,
|
| 48 |
-
"grad_norm": 52.3621940612793,
|
| 49 |
-
"learning_rate": 6e-06,
|
| 50 |
-
"loss": 4.5096,
|
| 51 |
-
"step": 120
|
| 52 |
-
},
|
| 53 |
-
{
|
| 54 |
-
"epoch": 0.01,
|
| 55 |
-
"grad_norm": 49.86561584472656,
|
| 56 |
-
"learning_rate": 7e-06,
|
| 57 |
-
"loss": 4.493,
|
| 58 |
-
"step": 140
|
| 59 |
-
},
|
| 60 |
-
{
|
| 61 |
-
"epoch": 0.01,
|
| 62 |
-
"grad_norm": 20.034923553466797,
|
| 63 |
-
"learning_rate": 8e-06,
|
| 64 |
-
"loss": 4.4088,
|
| 65 |
-
"step": 160
|
| 66 |
-
},
|
| 67 |
-
{
|
| 68 |
-
"epoch": 0.01,
|
| 69 |
-
"grad_norm": 50.790679931640625,
|
| 70 |
-
"learning_rate": 9e-06,
|
| 71 |
-
"loss": 4.4901,
|
| 72 |
-
"step": 180
|
| 73 |
-
},
|
| 74 |
-
{
|
| 75 |
-
"epoch": 0.02,
|
| 76 |
-
"grad_norm": 48.5693473815918,
|
| 77 |
-
"learning_rate": 9.999999999999999e-06,
|
| 78 |
-
"loss": 4.3628,
|
| 79 |
-
"step": 200
|
| 80 |
-
},
|
| 81 |
-
{
|
| 82 |
-
"epoch": 0.02,
|
| 83 |
-
"grad_norm": 37.95353698730469,
|
| 84 |
-
"learning_rate": 1.1e-05,
|
| 85 |
-
"loss": 4.3298,
|
| 86 |
-
"step": 220
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.02,
|
| 90 |
-
"grad_norm": 35.7153434753418,
|
| 91 |
-
"learning_rate": 1.2e-05,
|
| 92 |
-
"loss": 4.2839,
|
| 93 |
-
"step": 240
|
| 94 |
-
},
|
| 95 |
-
{
|
| 96 |
-
"epoch": 0.02,
|
| 97 |
-
"grad_norm": 91.47773742675781,
|
| 98 |
-
"learning_rate": 1.3000000000000001e-05,
|
| 99 |
-
"loss": 4.1238,
|
| 100 |
-
"step": 260
|
| 101 |
-
},
|
| 102 |
-
{
|
| 103 |
-
"epoch": 0.02,
|
| 104 |
-
"grad_norm": 23.16193389892578,
|
| 105 |
-
"learning_rate": 1.4e-05,
|
| 106 |
-
"loss": 4.1245,
|
| 107 |
-
"step": 280
|
| 108 |
-
},
|
| 109 |
-
{
|
| 110 |
-
"epoch": 0.02,
|
| 111 |
-
"grad_norm": 28.304485321044922,
|
| 112 |
-
"learning_rate": 1.5e-05,
|
| 113 |
-
"loss": 4.2198,
|
| 114 |
-
"step": 300
|
| 115 |
-
},
|
| 116 |
-
{
|
| 117 |
-
"epoch": 0.03,
|
| 118 |
-
"grad_norm": 34.03230285644531,
|
| 119 |
-
"learning_rate": 1.6e-05,
|
| 120 |
-
"loss": 4.0958,
|
| 121 |
-
"step": 320
|
| 122 |
-
},
|
| 123 |
-
{
|
| 124 |
-
"epoch": 0.03,
|
| 125 |
-
"grad_norm": 29.786975860595703,
|
| 126 |
-
"learning_rate": 1.7e-05,
|
| 127 |
-
"loss": 4.024,
|
| 128 |
-
"step": 340
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.03,
|
| 132 |
-
"grad_norm": 33.04754638671875,
|
| 133 |
-
"learning_rate": 1.8e-05,
|
| 134 |
-
"loss": 4.0832,
|
| 135 |
-
"step": 360
|
| 136 |
-
},
|
| 137 |
-
{
|
| 138 |
-
"epoch": 0.03,
|
| 139 |
-
"grad_norm": 28.68460464477539,
|
| 140 |
-
"learning_rate": 1.9e-05,
|
| 141 |
-
"loss": 3.9827,
|
| 142 |
-
"step": 380
|
| 143 |
-
},
|
| 144 |
-
{
|
| 145 |
-
"epoch": 0.03,
|
| 146 |
-
"grad_norm": 26.463253021240234,
|
| 147 |
-
"learning_rate": 1.9999999999999998e-05,
|
| 148 |
-
"loss": 3.9454,
|
| 149 |
-
"step": 400
|
| 150 |
-
},
|
| 151 |
-
{
|
| 152 |
-
"epoch": 0.03,
|
| 153 |
-
"grad_norm": 19.407127380371094,
|
| 154 |
-
"learning_rate": 2.1e-05,
|
| 155 |
-
"loss": 4.0119,
|
| 156 |
-
"step": 420
|
| 157 |
-
},
|
| 158 |
-
{
|
| 159 |
-
"epoch": 0.04,
|
| 160 |
-
"grad_norm": 26.383380889892578,
|
| 161 |
-
"learning_rate": 2.2e-05,
|
| 162 |
-
"loss": 3.9554,
|
| 163 |
-
"step": 440
|
| 164 |
-
},
|
| 165 |
-
{
|
| 166 |
-
"epoch": 0.04,
|
| 167 |
-
"grad_norm": 33.225223541259766,
|
| 168 |
-
"learning_rate": 2.3000000000000003e-05,
|
| 169 |
-
"loss": 3.8172,
|
| 170 |
-
"step": 460
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.04,
|
| 174 |
-
"grad_norm": 26.000978469848633,
|
| 175 |
-
"learning_rate": 2.4e-05,
|
| 176 |
-
"loss": 3.8934,
|
| 177 |
-
"step": 480
|
| 178 |
-
},
|
| 179 |
-
{
|
| 180 |
-
"epoch": 0.04,
|
| 181 |
-
"grad_norm": 28.714366912841797,
|
| 182 |
-
"learning_rate": 2.5e-05,
|
| 183 |
-
"loss": 3.9194,
|
| 184 |
-
"step": 500
|
| 185 |
-
},
|
| 186 |
-
{
|
| 187 |
-
"epoch": 0.04,
|
| 188 |
-
"grad_norm": 28.721248626708984,
|
| 189 |
-
"learning_rate": 2.6000000000000002e-05,
|
| 190 |
-
"loss": 3.8144,
|
| 191 |
-
"step": 520
|
| 192 |
-
},
|
| 193 |
-
{
|
| 194 |
-
"epoch": 0.04,
|
| 195 |
-
"grad_norm": 24.934555053710938,
|
| 196 |
-
"learning_rate": 2.7000000000000002e-05,
|
| 197 |
-
"loss": 3.9166,
|
| 198 |
-
"step": 540
|
| 199 |
-
},
|
| 200 |
-
{
|
| 201 |
-
"epoch": 0.04,
|
| 202 |
-
"grad_norm": 23.113840103149414,
|
| 203 |
-
"learning_rate": 2.8e-05,
|
| 204 |
-
"loss": 3.8248,
|
| 205 |
-
"step": 560
|
| 206 |
-
},
|
| 207 |
-
{
|
| 208 |
-
"epoch": 0.05,
|
| 209 |
-
"grad_norm": 21.58758544921875,
|
| 210 |
-
"learning_rate": 2.9e-05,
|
| 211 |
-
"loss": 3.7538,
|
| 212 |
-
"step": 580
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 0.05,
|
| 216 |
-
"grad_norm": 22.339618682861328,
|
| 217 |
-
"learning_rate": 3e-05,
|
| 218 |
-
"loss": 3.726,
|
| 219 |
-
"step": 600
|
| 220 |
-
},
|
| 221 |
-
{
|
| 222 |
-
"epoch": 0.05,
|
| 223 |
-
"grad_norm": 49.25693893432617,
|
| 224 |
-
"learning_rate": 2.999992132854894e-05,
|
| 225 |
-
"loss": 3.8692,
|
| 226 |
-
"step": 620
|
| 227 |
-
},
|
| 228 |
-
{
|
| 229 |
-
"epoch": 0.05,
|
| 230 |
-
"grad_norm": 45.1494026184082,
|
| 231 |
-
"learning_rate": 2.999968531502098e-05,
|
| 232 |
-
"loss": 3.7374,
|
| 233 |
-
"step": 640
|
| 234 |
-
},
|
| 235 |
-
{
|
| 236 |
-
"epoch": 0.05,
|
| 237 |
-
"grad_norm": 72.25853729248047,
|
| 238 |
-
"learning_rate": 2.99992919618918e-05,
|
| 239 |
-
"loss": 3.7735,
|
| 240 |
-
"step": 660
|
| 241 |
-
},
|
| 242 |
-
{
|
| 243 |
-
"epoch": 0.05,
|
| 244 |
-
"grad_norm": 39.445220947265625,
|
| 245 |
-
"learning_rate": 2.999874127328748e-05,
|
| 246 |
-
"loss": 3.759,
|
| 247 |
-
"step": 680
|
| 248 |
-
},
|
| 249 |
-
{
|
| 250 |
-
"epoch": 0.06,
|
| 251 |
-
"grad_norm": 21.18370246887207,
|
| 252 |
-
"learning_rate": 2.9998033254984483e-05,
|
| 253 |
-
"loss": 3.7841,
|
| 254 |
-
"step": 700
|
| 255 |
-
},
|
| 256 |
-
{
|
| 257 |
-
"epoch": 0.06,
|
| 258 |
-
"grad_norm": 24.310373306274414,
|
| 259 |
-
"learning_rate": 2.999716791440959e-05,
|
| 260 |
-
"loss": 3.679,
|
| 261 |
-
"step": 720
|
| 262 |
-
},
|
| 263 |
-
{
|
| 264 |
-
"epoch": 0.06,
|
| 265 |
-
"grad_norm": 36.432350158691406,
|
| 266 |
-
"learning_rate": 2.9996145260639812e-05,
|
| 267 |
-
"loss": 3.6796,
|
| 268 |
-
"step": 740
|
| 269 |
-
},
|
| 270 |
-
{
|
| 271 |
-
"epoch": 0.06,
|
| 272 |
-
"grad_norm": 32.12275314331055,
|
| 273 |
-
"learning_rate": 2.9994965304402304e-05,
|
| 274 |
-
"loss": 3.7613,
|
| 275 |
-
"step": 760
|
| 276 |
-
},
|
| 277 |
-
{
|
| 278 |
-
"epoch": 0.06,
|
| 279 |
-
"grad_norm": 38.32442855834961,
|
| 280 |
-
"learning_rate": 2.999362805807425e-05,
|
| 281 |
-
"loss": 3.7586,
|
| 282 |
-
"step": 780
|
| 283 |
-
},
|
| 284 |
-
{
|
| 285 |
-
"epoch": 0.06,
|
| 286 |
-
"grad_norm": 30.289432525634766,
|
| 287 |
-
"learning_rate": 2.9992133535682725e-05,
|
| 288 |
-
"loss": 3.6919,
|
| 289 |
-
"step": 800
|
| 290 |
-
},
|
| 291 |
-
{
|
| 292 |
-
"epoch": 0.07,
|
| 293 |
-
"grad_norm": 32.69138717651367,
|
| 294 |
-
"learning_rate": 2.9990481752904566e-05,
|
| 295 |
-
"loss": 3.6855,
|
| 296 |
-
"step": 820
|
| 297 |
-
},
|
| 298 |
-
{
|
| 299 |
-
"epoch": 0.07,
|
| 300 |
-
"grad_norm": 46.554874420166016,
|
| 301 |
-
"learning_rate": 2.9988672727066197e-05,
|
| 302 |
-
"loss": 3.7201,
|
| 303 |
-
"step": 840
|
| 304 |
-
},
|
| 305 |
-
{
|
| 306 |
-
"epoch": 0.07,
|
| 307 |
-
"grad_norm": 28.671123504638672,
|
| 308 |
-
"learning_rate": 2.9986706477143436e-05,
|
| 309 |
-
"loss": 3.6594,
|
| 310 |
-
"step": 860
|
| 311 |
-
},
|
| 312 |
-
{
|
| 313 |
-
"epoch": 0.07,
|
| 314 |
-
"grad_norm": 49.44480895996094,
|
| 315 |
-
"learning_rate": 2.9984583023761318e-05,
|
| 316 |
-
"loss": 3.7271,
|
| 317 |
-
"step": 880
|
| 318 |
-
},
|
| 319 |
-
{
|
| 320 |
-
"epoch": 0.07,
|
| 321 |
-
"grad_norm": 26.61457061767578,
|
| 322 |
-
"learning_rate": 2.998230238919386e-05,
|
| 323 |
-
"loss": 3.7376,
|
| 324 |
-
"step": 900
|
| 325 |
-
},
|
| 326 |
-
{
|
| 327 |
-
"epoch": 0.07,
|
| 328 |
-
"grad_norm": 27.453275680541992,
|
| 329 |
-
"learning_rate": 2.9979864597363846e-05,
|
| 330 |
-
"loss": 3.6716,
|
| 331 |
-
"step": 920
|
| 332 |
-
},
|
| 333 |
-
{
|
| 334 |
-
"epoch": 0.08,
|
| 335 |
-
"grad_norm": 22.791175842285156,
|
| 336 |
-
"learning_rate": 2.9977269673842554e-05,
|
| 337 |
-
"loss": 3.6172,
|
| 338 |
-
"step": 940
|
| 339 |
-
},
|
| 340 |
-
{
|
| 341 |
-
"epoch": 0.08,
|
| 342 |
-
"grad_norm": 58.2718620300293,
|
| 343 |
-
"learning_rate": 2.997451764584951e-05,
|
| 344 |
-
"loss": 3.7494,
|
| 345 |
-
"step": 960
|
| 346 |
-
},
|
| 347 |
-
{
|
| 348 |
-
"epoch": 0.08,
|
| 349 |
-
"grad_norm": 33.610286712646484,
|
| 350 |
-
"learning_rate": 2.9971608542252175e-05,
|
| 351 |
-
"loss": 3.7077,
|
| 352 |
-
"step": 980
|
| 353 |
-
},
|
| 354 |
-
{
|
| 355 |
-
"epoch": 0.08,
|
| 356 |
-
"grad_norm": 25.48147201538086,
|
| 357 |
-
"learning_rate": 2.9968542393565674e-05,
|
| 358 |
-
"loss": 3.6721,
|
| 359 |
-
"step": 1000
|
| 360 |
-
},
|
| 361 |
-
{
|
| 362 |
-
"epoch": 0.08,
|
| 363 |
-
"grad_norm": 27.07135581970215,
|
| 364 |
-
"learning_rate": 2.996531923195246e-05,
|
| 365 |
-
"loss": 3.7106,
|
| 366 |
-
"step": 1020
|
| 367 |
-
},
|
| 368 |
-
{
|
| 369 |
-
"epoch": 0.08,
|
| 370 |
-
"grad_norm": 37.24673843383789,
|
| 371 |
-
"learning_rate": 2.996193909122197e-05,
|
| 372 |
-
"loss": 3.7447,
|
| 373 |
-
"step": 1040
|
| 374 |
-
},
|
| 375 |
-
{
|
| 376 |
-
"epoch": 0.08,
|
| 377 |
-
"grad_norm": 26.41890525817871,
|
| 378 |
-
"learning_rate": 2.995840200683028e-05,
|
| 379 |
-
"loss": 3.5839,
|
| 380 |
-
"step": 1060
|
| 381 |
-
},
|
| 382 |
-
{
|
| 383 |
-
"epoch": 0.09,
|
| 384 |
-
"grad_norm": 32.88002014160156,
|
| 385 |
-
"learning_rate": 2.995470801587973e-05,
|
| 386 |
-
"loss": 3.6606,
|
| 387 |
-
"step": 1080
|
| 388 |
-
},
|
| 389 |
-
{
|
| 390 |
-
"epoch": 0.09,
|
| 391 |
-
"grad_norm": 32.0895881652832,
|
| 392 |
-
"learning_rate": 2.9950857157118544e-05,
|
| 393 |
-
"loss": 3.677,
|
| 394 |
-
"step": 1100
|
| 395 |
-
},
|
| 396 |
-
{
|
| 397 |
-
"epoch": 0.09,
|
| 398 |
-
"grad_norm": 37.726783752441406,
|
| 399 |
-
"learning_rate": 2.9946849470940395e-05,
|
| 400 |
-
"loss": 3.5546,
|
| 401 |
-
"step": 1120
|
| 402 |
-
},
|
| 403 |
-
{
|
| 404 |
-
"epoch": 0.09,
|
| 405 |
-
"grad_norm": 56.246299743652344,
|
| 406 |
-
"learning_rate": 2.9942684999384034e-05,
|
| 407 |
-
"loss": 3.6391,
|
| 408 |
-
"step": 1140
|
| 409 |
-
},
|
| 410 |
-
{
|
| 411 |
-
"epoch": 0.09,
|
| 412 |
-
"grad_norm": 35.675662994384766,
|
| 413 |
-
"learning_rate": 2.993836378613278e-05,
|
| 414 |
-
"loss": 3.5918,
|
| 415 |
-
"step": 1160
|
| 416 |
-
},
|
| 417 |
-
{
|
| 418 |
-
"epoch": 0.09,
|
| 419 |
-
"grad_norm": 26.685134887695312,
|
| 420 |
-
"learning_rate": 2.993388587651412e-05,
|
| 421 |
-
"loss": 3.6331,
|
| 422 |
-
"step": 1180
|
| 423 |
-
},
|
| 424 |
-
{
|
| 425 |
-
"epoch": 0.1,
|
| 426 |
-
"grad_norm": 27.400333404541016,
|
| 427 |
-
"learning_rate": 2.992925131749921e-05,
|
| 428 |
-
"loss": 3.6214,
|
| 429 |
-
"step": 1200
|
| 430 |
-
},
|
| 431 |
-
{
|
| 432 |
-
"epoch": 0.1,
|
| 433 |
-
"grad_norm": 28.501314163208008,
|
| 434 |
-
"learning_rate": 2.9924460157702378e-05,
|
| 435 |
-
"loss": 3.6619,
|
| 436 |
-
"step": 1220
|
| 437 |
-
},
|
| 438 |
-
{
|
| 439 |
-
"epoch": 0.1,
|
| 440 |
-
"grad_norm": 30.773778915405273,
|
| 441 |
-
"learning_rate": 2.991951244738063e-05,
|
| 442 |
-
"loss": 3.6453,
|
| 443 |
-
"step": 1240
|
| 444 |
-
},
|
| 445 |
-
{
|
| 446 |
-
"epoch": 0.1,
|
| 447 |
-
"grad_norm": 24.701374053955078,
|
| 448 |
-
"learning_rate": 2.9914408238433095e-05,
|
| 449 |
-
"loss": 3.7282,
|
| 450 |
-
"step": 1260
|
| 451 |
-
},
|
| 452 |
-
{
|
| 453 |
-
"epoch": 0.1,
|
| 454 |
-
"grad_norm": 27.605117797851562,
|
| 455 |
-
"learning_rate": 2.990914758440052e-05,
|
| 456 |
-
"loss": 3.6635,
|
| 457 |
-
"step": 1280
|
| 458 |
-
},
|
| 459 |
-
{
|
| 460 |
-
"epoch": 0.1,
|
| 461 |
-
"grad_norm": 27.829086303710938,
|
| 462 |
-
"learning_rate": 2.9903730540464668e-05,
|
| 463 |
-
"loss": 3.5293,
|
| 464 |
-
"step": 1300
|
| 465 |
-
},
|
| 466 |
-
{
|
| 467 |
-
"epoch": 0.11,
|
| 468 |
-
"grad_norm": 40.916263580322266,
|
| 469 |
-
"learning_rate": 2.9898157163447767e-05,
|
| 470 |
-
"loss": 3.6976,
|
| 471 |
-
"step": 1320
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 0.11,
|
| 475 |
-
"grad_norm": 33.31068420410156,
|
| 476 |
-
"learning_rate": 2.9892427511811912e-05,
|
| 477 |
-
"loss": 3.548,
|
| 478 |
-
"step": 1340
|
| 479 |
-
},
|
| 480 |
-
{
|
| 481 |
-
"epoch": 0.11,
|
| 482 |
-
"grad_norm": 29.932533264160156,
|
| 483 |
-
"learning_rate": 2.9886541645658435e-05,
|
| 484 |
-
"loss": 3.7486,
|
| 485 |
-
"step": 1360
|
| 486 |
-
},
|
| 487 |
-
{
|
| 488 |
-
"epoch": 0.11,
|
| 489 |
-
"grad_norm": 35.59455490112305,
|
| 490 |
-
"learning_rate": 2.9880499626727284e-05,
|
| 491 |
-
"loss": 3.6342,
|
| 492 |
-
"step": 1380
|
| 493 |
-
},
|
| 494 |
-
{
|
| 495 |
-
"epoch": 0.11,
|
| 496 |
-
"grad_norm": 29.93869400024414,
|
| 497 |
-
"learning_rate": 2.9874301518396377e-05,
|
| 498 |
-
"loss": 3.6615,
|
| 499 |
-
"step": 1400
|
| 500 |
-
},
|
| 501 |
-
{
|
| 502 |
-
"epoch": 0.11,
|
| 503 |
-
"grad_norm": 43.417213439941406,
|
| 504 |
-
"learning_rate": 2.986794738568094e-05,
|
| 505 |
-
"loss": 3.607,
|
| 506 |
-
"step": 1420
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 0.12,
|
| 510 |
-
"grad_norm": 52.483917236328125,
|
| 511 |
-
"learning_rate": 2.9861437295232825e-05,
|
| 512 |
-
"loss": 3.5937,
|
| 513 |
-
"step": 1440
|
| 514 |
-
},
|
| 515 |
-
{
|
| 516 |
-
"epoch": 0.12,
|
| 517 |
-
"grad_norm": 30.312334060668945,
|
| 518 |
-
"learning_rate": 2.9854771315339787e-05,
|
| 519 |
-
"loss": 3.5991,
|
| 520 |
-
"step": 1460
|
| 521 |
-
},
|
| 522 |
-
{
|
| 523 |
-
"epoch": 0.12,
|
| 524 |
-
"grad_norm": 49.459136962890625,
|
| 525 |
-
"learning_rate": 2.984794951592481e-05,
|
| 526 |
-
"loss": 3.5261,
|
| 527 |
-
"step": 1480
|
| 528 |
-
},
|
| 529 |
-
{
|
| 530 |
-
"epoch": 0.12,
|
| 531 |
-
"grad_norm": 34.81111526489258,
|
| 532 |
-
"learning_rate": 2.984097196854534e-05,
|
| 533 |
-
"loss": 3.6818,
|
| 534 |
-
"step": 1500
|
| 535 |
-
},
|
| 536 |
-
{
|
| 537 |
-
"epoch": 0.12,
|
| 538 |
-
"grad_norm": 34.721946716308594,
|
| 539 |
-
"learning_rate": 2.9833838746392544e-05,
|
| 540 |
-
"loss": 3.5636,
|
| 541 |
-
"step": 1520
|
| 542 |
-
},
|
| 543 |
-
{
|
| 544 |
-
"epoch": 0.12,
|
| 545 |
-
"grad_norm": 31.46621322631836,
|
| 546 |
-
"learning_rate": 2.982654992429056e-05,
|
| 547 |
-
"loss": 3.5597,
|
| 548 |
-
"step": 1540
|
| 549 |
-
},
|
| 550 |
-
{
|
| 551 |
-
"epoch": 0.12,
|
| 552 |
-
"grad_norm": 38.78512191772461,
|
| 553 |
-
"learning_rate": 2.981910557869566e-05,
|
| 554 |
-
"loss": 3.661,
|
| 555 |
-
"step": 1560
|
| 556 |
-
},
|
| 557 |
-
{
|
| 558 |
-
"epoch": 0.13,
|
| 559 |
-
"grad_norm": 27.38837432861328,
|
| 560 |
-
"learning_rate": 2.981150578769553e-05,
|
| 561 |
-
"loss": 3.6173,
|
| 562 |
-
"step": 1580
|
| 563 |
-
},
|
| 564 |
-
{
|
| 565 |
-
"epoch": 0.13,
|
| 566 |
-
"grad_norm": 45.619632720947266,
|
| 567 |
-
"learning_rate": 2.980375063100836e-05,
|
| 568 |
-
"loss": 3.6632,
|
| 569 |
-
"step": 1600
|
| 570 |
-
},
|
| 571 |
-
{
|
| 572 |
-
"epoch": 0.13,
|
| 573 |
-
"grad_norm": 30.708433151245117,
|
| 574 |
-
"learning_rate": 2.979584018998209e-05,
|
| 575 |
-
"loss": 3.5165,
|
| 576 |
-
"step": 1620
|
| 577 |
-
},
|
| 578 |
-
{
|
| 579 |
-
"epoch": 0.13,
|
| 580 |
-
"grad_norm": 35.472938537597656,
|
| 581 |
-
"learning_rate": 2.97877745475935e-05,
|
| 582 |
-
"loss": 3.5157,
|
| 583 |
-
"step": 1640
|
| 584 |
-
},
|
| 585 |
-
{
|
| 586 |
-
"epoch": 0.13,
|
| 587 |
-
"grad_norm": 39.029415130615234,
|
| 588 |
-
"learning_rate": 2.9779553788447358e-05,
|
| 589 |
-
"loss": 3.6259,
|
| 590 |
-
"step": 1660
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 0.13,
|
| 594 |
-
"grad_norm": 57.90769577026367,
|
| 595 |
-
"learning_rate": 2.977117799877554e-05,
|
| 596 |
-
"loss": 3.6378,
|
| 597 |
-
"step": 1680
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 0.14,
|
| 601 |
-
"grad_norm": 36.95255661010742,
|
| 602 |
-
"learning_rate": 2.9762647266436115e-05,
|
| 603 |
-
"loss": 3.5845,
|
| 604 |
-
"step": 1700
|
| 605 |
-
},
|
| 606 |
-
{
|
| 607 |
-
"epoch": 0.14,
|
| 608 |
-
"grad_norm": 27.456787109375,
|
| 609 |
-
"learning_rate": 2.9753961680912432e-05,
|
| 610 |
-
"loss": 3.6647,
|
| 611 |
-
"step": 1720
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 0.14,
|
| 615 |
-
"grad_norm": 27.383285522460938,
|
| 616 |
-
"learning_rate": 2.9745121333312166e-05,
|
| 617 |
-
"loss": 3.6668,
|
| 618 |
-
"step": 1740
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 0.14,
|
| 622 |
-
"grad_norm": 26.555049896240234,
|
| 623 |
-
"learning_rate": 2.9736126316366385e-05,
|
| 624 |
-
"loss": 3.6617,
|
| 625 |
-
"step": 1760
|
| 626 |
-
},
|
| 627 |
-
{
|
| 628 |
-
"epoch": 0.14,
|
| 629 |
-
"grad_norm": 34.009620666503906,
|
| 630 |
-
"learning_rate": 2.9726976724428563e-05,
|
| 631 |
-
"loss": 3.572,
|
| 632 |
-
"step": 1780
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 0.14,
|
| 636 |
-
"grad_norm": 45.44181823730469,
|
| 637 |
-
"learning_rate": 2.9717672653473588e-05,
|
| 638 |
-
"loss": 3.6354,
|
| 639 |
-
"step": 1800
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 0.15,
|
| 643 |
-
"grad_norm": 30.79588508605957,
|
| 644 |
-
"learning_rate": 2.9708214201096758e-05,
|
| 645 |
-
"loss": 3.6953,
|
| 646 |
-
"step": 1820
|
| 647 |
-
},
|
| 648 |
-
{
|
| 649 |
-
"epoch": 0.15,
|
| 650 |
-
"grad_norm": 46.61872482299805,
|
| 651 |
-
"learning_rate": 2.9698601466512767e-05,
|
| 652 |
-
"loss": 3.5373,
|
| 653 |
-
"step": 1840
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 0.15,
|
| 657 |
-
"grad_norm": 42.86500930786133,
|
| 658 |
-
"learning_rate": 2.9688834550554647e-05,
|
| 659 |
-
"loss": 3.5982,
|
| 660 |
-
"step": 1860
|
| 661 |
-
},
|
| 662 |
-
{
|
| 663 |
-
"epoch": 0.15,
|
| 664 |
-
"grad_norm": 33.480289459228516,
|
| 665 |
-
"learning_rate": 2.9678913555672733e-05,
|
| 666 |
-
"loss": 3.6024,
|
| 667 |
-
"step": 1880
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 0.15,
|
| 671 |
-
"grad_norm": 36.41415786743164,
|
| 672 |
-
"learning_rate": 2.966883858593356e-05,
|
| 673 |
-
"loss": 3.4843,
|
| 674 |
-
"step": 1900
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 0.15,
|
| 678 |
-
"grad_norm": 41.39873123168945,
|
| 679 |
-
"learning_rate": 2.9658609747018796e-05,
|
| 680 |
-
"loss": 3.5257,
|
| 681 |
-
"step": 1920
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 0.16,
|
| 685 |
-
"grad_norm": 31.24024200439453,
|
| 686 |
-
"learning_rate": 2.964822714622412e-05,
|
| 687 |
-
"loss": 3.5927,
|
| 688 |
-
"step": 1940
|
| 689 |
-
},
|
| 690 |
-
{
|
| 691 |
-
"epoch": 0.16,
|
| 692 |
-
"grad_norm": 52.78026580810547,
|
| 693 |
-
"learning_rate": 2.9637690892458103e-05,
|
| 694 |
-
"loss": 3.4678,
|
| 695 |
-
"step": 1960
|
| 696 |
-
},
|
| 697 |
-
{
|
| 698 |
-
"epoch": 0.16,
|
| 699 |
-
"grad_norm": 27.40117835998535,
|
| 700 |
-
"learning_rate": 2.962700109624106e-05,
|
| 701 |
-
"loss": 3.5541,
|
| 702 |
-
"step": 1980
|
| 703 |
-
},
|
| 704 |
-
{
|
| 705 |
-
"epoch": 0.16,
|
| 706 |
-
"grad_norm": 23.172683715820312,
|
| 707 |
-
"learning_rate": 2.961615786970389e-05,
|
| 708 |
-
"loss": 3.5713,
|
| 709 |
-
"step": 2000
|
| 710 |
-
},
|
| 711 |
-
{
|
| 712 |
-
"epoch": 0.16,
|
| 713 |
-
"grad_norm": 24.177541732788086,
|
| 714 |
-
"learning_rate": 2.960516132658692e-05,
|
| 715 |
-
"loss": 3.585,
|
| 716 |
-
"step": 2020
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 0.16,
|
| 720 |
-
"grad_norm": 44.673912048339844,
|
| 721 |
-
"learning_rate": 2.9594011582238672e-05,
|
| 722 |
-
"loss": 3.5035,
|
| 723 |
-
"step": 2040
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 0.16,
|
| 727 |
-
"grad_norm": 35.91664505004883,
|
| 728 |
-
"learning_rate": 2.95827087536147e-05,
|
| 729 |
-
"loss": 3.6404,
|
| 730 |
-
"step": 2060
|
| 731 |
-
},
|
| 732 |
-
{
|
| 733 |
-
"epoch": 0.17,
|
| 734 |
-
"grad_norm": 27.3450870513916,
|
| 735 |
-
"learning_rate": 2.9571252959276313e-05,
|
| 736 |
-
"loss": 3.5121,
|
| 737 |
-
"step": 2080
|
| 738 |
-
},
|
| 739 |
-
{
|
| 740 |
-
"epoch": 0.17,
|
| 741 |
-
"grad_norm": 25.66405487060547,
|
| 742 |
-
"learning_rate": 2.955964431938939e-05,
|
| 743 |
-
"loss": 3.5009,
|
| 744 |
-
"step": 2100
|
| 745 |
-
},
|
| 746 |
-
{
|
| 747 |
-
"epoch": 0.17,
|
| 748 |
-
"grad_norm": 24.1674861907959,
|
| 749 |
-
"learning_rate": 2.9547882955723052e-05,
|
| 750 |
-
"loss": 3.5482,
|
| 751 |
-
"step": 2120
|
| 752 |
-
},
|
| 753 |
-
{
|
| 754 |
-
"epoch": 0.17,
|
| 755 |
-
"grad_norm": 49.72268295288086,
|
| 756 |
-
"learning_rate": 2.953596899164846e-05,
|
| 757 |
-
"loss": 3.4969,
|
| 758 |
-
"step": 2140
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 0.17,
|
| 762 |
-
"grad_norm": 26.238168716430664,
|
| 763 |
-
"learning_rate": 2.9523902552137436e-05,
|
| 764 |
-
"loss": 3.5541,
|
| 765 |
-
"step": 2160
|
| 766 |
-
},
|
| 767 |
-
{
|
| 768 |
-
"epoch": 0.17,
|
| 769 |
-
"grad_norm": 30.524545669555664,
|
| 770 |
-
"learning_rate": 2.951168376376124e-05,
|
| 771 |
-
"loss": 3.6343,
|
| 772 |
-
"step": 2180
|
| 773 |
-
},
|
| 774 |
-
{
|
| 775 |
-
"epoch": 0.18,
|
| 776 |
-
"grad_norm": 38.179908752441406,
|
| 777 |
-
"learning_rate": 2.9499312754689168e-05,
|
| 778 |
-
"loss": 3.4795,
|
| 779 |
-
"step": 2200
|
| 780 |
-
},
|
| 781 |
-
{
|
| 782 |
-
"epoch": 0.18,
|
| 783 |
-
"grad_norm": 32.98453903198242,
|
| 784 |
-
"learning_rate": 2.9486789654687256e-05,
|
| 785 |
-
"loss": 3.6333,
|
| 786 |
-
"step": 2220
|
| 787 |
-
},
|
| 788 |
-
{
|
| 789 |
-
"epoch": 0.18,
|
| 790 |
-
"grad_norm": 26.77848243713379,
|
| 791 |
-
"learning_rate": 2.94741145951169e-05,
|
| 792 |
-
"loss": 3.5654,
|
| 793 |
-
"step": 2240
|
| 794 |
-
},
|
| 795 |
-
{
|
| 796 |
-
"epoch": 0.18,
|
| 797 |
-
"grad_norm": 27.737852096557617,
|
| 798 |
-
"learning_rate": 2.9461287708933475e-05,
|
| 799 |
-
"loss": 3.5044,
|
| 800 |
-
"step": 2260
|
| 801 |
-
},
|
| 802 |
-
{
|
| 803 |
-
"epoch": 0.18,
|
| 804 |
-
"grad_norm": 34.2584342956543,
|
| 805 |
-
"learning_rate": 2.9448309130684944e-05,
|
| 806 |
-
"loss": 3.5979,
|
| 807 |
-
"step": 2280
|
| 808 |
-
},
|
| 809 |
-
{
|
| 810 |
-
"epoch": 0.18,
|
| 811 |
-
"grad_norm": 57.86616897583008,
|
| 812 |
-
"learning_rate": 2.9435178996510456e-05,
|
| 813 |
-
"loss": 3.5726,
|
| 814 |
-
"step": 2300
|
| 815 |
-
},
|
| 816 |
-
{
|
| 817 |
-
"epoch": 0.19,
|
| 818 |
-
"grad_norm": 37.64597702026367,
|
| 819 |
-
"learning_rate": 2.9421897444138902e-05,
|
| 820 |
-
"loss": 3.5913,
|
| 821 |
-
"step": 2320
|
| 822 |
-
},
|
| 823 |
-
{
|
| 824 |
-
"epoch": 0.19,
|
| 825 |
-
"grad_norm": 35.22037124633789,
|
| 826 |
-
"learning_rate": 2.9408464612887484e-05,
|
| 827 |
-
"loss": 3.5959,
|
| 828 |
-
"step": 2340
|
| 829 |
-
},
|
| 830 |
-
{
|
| 831 |
-
"epoch": 0.19,
|
| 832 |
-
"grad_norm": 31.878395080566406,
|
| 833 |
-
"learning_rate": 2.9394880643660242e-05,
|
| 834 |
-
"loss": 3.5974,
|
| 835 |
-
"step": 2360
|
| 836 |
-
},
|
| 837 |
-
{
|
| 838 |
-
"epoch": 0.19,
|
| 839 |
-
"grad_norm": 43.614994049072266,
|
| 840 |
-
"learning_rate": 2.938114567894659e-05,
|
| 841 |
-
"loss": 3.4834,
|
| 842 |
-
"step": 2380
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 0.19,
|
| 846 |
-
"grad_norm": 27.587766647338867,
|
| 847 |
-
"learning_rate": 2.9367259862819805e-05,
|
| 848 |
-
"loss": 3.6154,
|
| 849 |
-
"step": 2400
|
| 850 |
-
},
|
| 851 |
-
{
|
| 852 |
-
"epoch": 0.19,
|
| 853 |
-
"grad_norm": 30.223772048950195,
|
| 854 |
-
"learning_rate": 2.9353223340935533e-05,
|
| 855 |
-
"loss": 3.4871,
|
| 856 |
-
"step": 2420
|
| 857 |
-
},
|
| 858 |
-
{
|
| 859 |
-
"epoch": 0.2,
|
| 860 |
-
"grad_norm": 34.057884216308594,
|
| 861 |
-
"learning_rate": 2.933903626053024e-05,
|
| 862 |
-
"loss": 3.605,
|
| 863 |
-
"step": 2440
|
| 864 |
-
},
|
| 865 |
-
{
|
| 866 |
-
"epoch": 0.2,
|
| 867 |
-
"grad_norm": 39.219242095947266,
|
| 868 |
-
"learning_rate": 2.932469877041969e-05,
|
| 869 |
-
"loss": 3.6091,
|
| 870 |
-
"step": 2460
|
| 871 |
-
},
|
| 872 |
-
{
|
| 873 |
-
"epoch": 0.2,
|
| 874 |
-
"grad_norm": 33.33955001831055,
|
| 875 |
-
"learning_rate": 2.931021102099737e-05,
|
| 876 |
-
"loss": 3.4862,
|
| 877 |
-
"step": 2480
|
| 878 |
-
},
|
| 879 |
-
{
|
| 880 |
-
"epoch": 0.2,
|
| 881 |
-
"grad_norm": 37.07484436035156,
|
| 882 |
-
"learning_rate": 2.9295573164232913e-05,
|
| 883 |
-
"loss": 3.5267,
|
| 884 |
-
"step": 2500
|
| 885 |
-
},
|
| 886 |
-
{
|
| 887 |
-
"epoch": 0.2,
|
| 888 |
-
"grad_norm": 27.145864486694336,
|
| 889 |
-
"learning_rate": 2.9280785353670514e-05,
|
| 890 |
-
"loss": 3.4369,
|
| 891 |
-
"step": 2520
|
| 892 |
-
},
|
| 893 |
-
{
|
| 894 |
-
"epoch": 0.2,
|
| 895 |
-
"grad_norm": 30.31035041809082,
|
| 896 |
-
"learning_rate": 2.9265847744427305e-05,
|
| 897 |
-
"loss": 3.6056,
|
| 898 |
-
"step": 2540
|
| 899 |
-
},
|
| 900 |
-
{
|
| 901 |
-
"epoch": 0.2,
|
| 902 |
-
"grad_norm": 40.823490142822266,
|
| 903 |
-
"learning_rate": 2.925076049319174e-05,
|
| 904 |
-
"loss": 3.5916,
|
| 905 |
-
"step": 2560
|
| 906 |
-
},
|
| 907 |
-
{
|
| 908 |
-
"epoch": 0.21,
|
| 909 |
-
"grad_norm": 44.224796295166016,
|
| 910 |
-
"learning_rate": 2.9235523758221944e-05,
|
| 911 |
-
"loss": 3.5881,
|
| 912 |
-
"step": 2580
|
| 913 |
-
},
|
| 914 |
-
{
|
| 915 |
-
"epoch": 0.21,
|
| 916 |
-
"grad_norm": 33.34773254394531,
|
| 917 |
-
"learning_rate": 2.922013769934406e-05,
|
| 918 |
-
"loss": 3.5315,
|
| 919 |
-
"step": 2600
|
| 920 |
-
},
|
| 921 |
-
{
|
| 922 |
-
"epoch": 0.21,
|
| 923 |
-
"grad_norm": 25.755775451660156,
|
| 924 |
-
"learning_rate": 2.920460247795056e-05,
|
| 925 |
-
"loss": 3.621,
|
| 926 |
-
"step": 2620
|
| 927 |
-
},
|
| 928 |
-
{
|
| 929 |
-
"epoch": 0.21,
|
| 930 |
-
"grad_norm": 29.75731086730957,
|
| 931 |
-
"learning_rate": 2.918891825699857e-05,
|
| 932 |
-
"loss": 3.6067,
|
| 933 |
-
"step": 2640
|
| 934 |
-
},
|
| 935 |
-
{
|
| 936 |
-
"epoch": 0.21,
|
| 937 |
-
"grad_norm": 40.04263687133789,
|
| 938 |
-
"learning_rate": 2.9173085201008144e-05,
|
| 939 |
-
"loss": 3.5365,
|
| 940 |
-
"step": 2660
|
| 941 |
-
},
|
| 942 |
-
{
|
| 943 |
-
"epoch": 0.21,
|
| 944 |
-
"grad_norm": 26.394916534423828,
|
| 945 |
-
"learning_rate": 2.9157103476060547e-05,
|
| 946 |
-
"loss": 3.5944,
|
| 947 |
-
"step": 2680
|
| 948 |
-
},
|
| 949 |
-
{
|
| 950 |
-
"epoch": 0.22,
|
| 951 |
-
"grad_norm": 23.64137077331543,
|
| 952 |
-
"learning_rate": 2.914097324979651e-05,
|
| 953 |
-
"loss": 3.5048,
|
| 954 |
-
"step": 2700
|
| 955 |
-
},
|
| 956 |
-
{
|
| 957 |
-
"epoch": 0.22,
|
| 958 |
-
"grad_norm": 35.18840408325195,
|
| 959 |
-
"learning_rate": 2.9124694691414485e-05,
|
| 960 |
-
"loss": 3.5769,
|
| 961 |
-
"step": 2720
|
| 962 |
-
},
|
| 963 |
-
{
|
| 964 |
-
"epoch": 0.22,
|
| 965 |
-
"grad_norm": 25.55812644958496,
|
| 966 |
-
"learning_rate": 2.9108267971668828e-05,
|
| 967 |
-
"loss": 3.5082,
|
| 968 |
-
"step": 2740
|
| 969 |
-
},
|
| 970 |
-
{
|
| 971 |
-
"epoch": 0.22,
|
| 972 |
-
"grad_norm": 26.25998878479004,
|
| 973 |
-
"learning_rate": 2.909169326286807e-05,
|
| 974 |
-
"loss": 3.4967,
|
| 975 |
-
"step": 2760
|
| 976 |
-
},
|
| 977 |
-
{
|
| 978 |
-
"epoch": 0.22,
|
| 979 |
-
"grad_norm": 41.78467559814453,
|
| 980 |
-
"learning_rate": 2.9074970738873054e-05,
|
| 981 |
-
"loss": 3.6409,
|
| 982 |
-
"step": 2780
|
| 983 |
-
},
|
| 984 |
-
{
|
| 985 |
-
"epoch": 0.22,
|
| 986 |
-
"grad_norm": 27.53672218322754,
|
| 987 |
-
"learning_rate": 2.9058100575095156e-05,
|
| 988 |
-
"loss": 3.5903,
|
| 989 |
-
"step": 2800
|
| 990 |
-
},
|
| 991 |
-
{
|
| 992 |
-
"epoch": 0.23,
|
| 993 |
-
"grad_norm": 27.329477310180664,
|
| 994 |
-
"learning_rate": 2.90410829484944e-05,
|
| 995 |
-
"loss": 3.5148,
|
| 996 |
-
"step": 2820
|
| 997 |
-
},
|
| 998 |
-
{
|
| 999 |
-
"epoch": 0.23,
|
| 1000 |
-
"grad_norm": 34.077083587646484,
|
| 1001 |
-
"learning_rate": 2.902391803757764e-05,
|
| 1002 |
-
"loss": 3.5561,
|
| 1003 |
-
"step": 2840
|
| 1004 |
-
},
|
| 1005 |
-
{
|
| 1006 |
-
"epoch": 0.23,
|
| 1007 |
-
"grad_norm": 33.75153732299805,
|
| 1008 |
-
"learning_rate": 2.900660602239667e-05,
|
| 1009 |
-
"loss": 3.498,
|
| 1010 |
-
"step": 2860
|
| 1011 |
-
},
|
| 1012 |
-
{
|
| 1013 |
-
"epoch": 0.23,
|
| 1014 |
-
"grad_norm": 28.681190490722656,
|
| 1015 |
-
"learning_rate": 2.8989147084546335e-05,
|
| 1016 |
-
"loss": 3.501,
|
| 1017 |
-
"step": 2880
|
| 1018 |
-
},
|
| 1019 |
-
{
|
| 1020 |
-
"epoch": 0.23,
|
| 1021 |
-
"grad_norm": 41.04441833496094,
|
| 1022 |
-
"learning_rate": 2.8971541407162637e-05,
|
| 1023 |
-
"loss": 3.5707,
|
| 1024 |
-
"step": 2900
|
| 1025 |
-
},
|
| 1026 |
-
{
|
| 1027 |
-
"epoch": 0.23,
|
| 1028 |
-
"grad_norm": 31.303403854370117,
|
| 1029 |
-
"learning_rate": 2.8953789174920795e-05,
|
| 1030 |
-
"loss": 3.5748,
|
| 1031 |
-
"step": 2920
|
| 1032 |
-
},
|
| 1033 |
-
{
|
| 1034 |
-
"epoch": 0.24,
|
| 1035 |
-
"grad_norm": 36.61361312866211,
|
| 1036 |
-
"learning_rate": 2.8935890574033325e-05,
|
| 1037 |
-
"loss": 3.6372,
|
| 1038 |
-
"step": 2940
|
| 1039 |
-
},
|
| 1040 |
-
{
|
| 1041 |
-
"epoch": 0.24,
|
| 1042 |
-
"grad_norm": 24.437808990478516,
|
| 1043 |
-
"learning_rate": 2.8917845792248085e-05,
|
| 1044 |
-
"loss": 3.5737,
|
| 1045 |
-
"step": 2960
|
| 1046 |
-
},
|
| 1047 |
-
{
|
| 1048 |
-
"epoch": 0.24,
|
| 1049 |
-
"grad_norm": 47.12533187866211,
|
| 1050 |
-
"learning_rate": 2.8899655018846297e-05,
|
| 1051 |
-
"loss": 3.5125,
|
| 1052 |
-
"step": 2980
|
| 1053 |
-
},
|
| 1054 |
-
{
|
| 1055 |
-
"epoch": 0.24,
|
| 1056 |
-
"grad_norm": 23.667312622070312,
|
| 1057 |
-
"learning_rate": 2.8881318444640564e-05,
|
| 1058 |
-
"loss": 3.5043,
|
| 1059 |
-
"step": 3000
|
| 1060 |
-
},
|
| 1061 |
-
{
|
| 1062 |
-
"epoch": 0.24,
|
| 1063 |
-
"grad_norm": 26.047521591186523,
|
| 1064 |
-
"learning_rate": 2.8862836261972873e-05,
|
| 1065 |
-
"loss": 3.6236,
|
| 1066 |
-
"step": 3020
|
| 1067 |
-
},
|
| 1068 |
-
{
|
| 1069 |
-
"epoch": 0.24,
|
| 1070 |
-
"grad_norm": 39.91059494018555,
|
| 1071 |
-
"learning_rate": 2.8844208664712577e-05,
|
| 1072 |
-
"loss": 3.4851,
|
| 1073 |
-
"step": 3040
|
| 1074 |
-
},
|
| 1075 |
-
{
|
| 1076 |
-
"epoch": 0.24,
|
| 1077 |
-
"grad_norm": 57.92033004760742,
|
| 1078 |
-
"learning_rate": 2.882543584825435e-05,
|
| 1079 |
-
"loss": 3.5578,
|
| 1080 |
-
"step": 3060
|
| 1081 |
-
},
|
| 1082 |
-
{
|
| 1083 |
-
"epoch": 0.25,
|
| 1084 |
-
"grad_norm": 99.40699005126953,
|
| 1085 |
-
"learning_rate": 2.880651800951616e-05,
|
| 1086 |
-
"loss": 3.577,
|
| 1087 |
-
"step": 3080
|
| 1088 |
-
},
|
| 1089 |
-
{
|
| 1090 |
-
"epoch": 0.25,
|
| 1091 |
-
"grad_norm": 30.7738094329834,
|
| 1092 |
-
"learning_rate": 2.8787455346937182e-05,
|
| 1093 |
-
"loss": 3.5683,
|
| 1094 |
-
"step": 3100
|
| 1095 |
-
},
|
| 1096 |
-
{
|
| 1097 |
-
"epoch": 0.25,
|
| 1098 |
-
"grad_norm": 55.61745834350586,
|
| 1099 |
-
"learning_rate": 2.876824806047573e-05,
|
| 1100 |
-
"loss": 3.4959,
|
| 1101 |
-
"step": 3120
|
| 1102 |
-
},
|
| 1103 |
-
{
|
| 1104 |
-
"epoch": 0.25,
|
| 1105 |
-
"grad_norm": 23.431480407714844,
|
| 1106 |
-
"learning_rate": 2.8748896351607145e-05,
|
| 1107 |
-
"loss": 3.5882,
|
| 1108 |
-
"step": 3140
|
| 1109 |
-
},
|
| 1110 |
-
{
|
| 1111 |
-
"epoch": 0.25,
|
| 1112 |
-
"grad_norm": 20.944149017333984,
|
| 1113 |
-
"learning_rate": 2.8729400423321693e-05,
|
| 1114 |
-
"loss": 3.6096,
|
| 1115 |
-
"step": 3160
|
| 1116 |
-
},
|
| 1117 |
-
{
|
| 1118 |
-
"epoch": 0.25,
|
| 1119 |
-
"grad_norm": 46.77301788330078,
|
| 1120 |
-
"learning_rate": 2.8709760480122443e-05,
|
| 1121 |
-
"loss": 3.5665,
|
| 1122 |
-
"step": 3180
|
| 1123 |
-
},
|
| 1124 |
-
{
|
| 1125 |
-
"epoch": 0.26,
|
| 1126 |
-
"grad_norm": 36.645328521728516,
|
| 1127 |
-
"learning_rate": 2.8689976728023103e-05,
|
| 1128 |
-
"loss": 3.5087,
|
| 1129 |
-
"step": 3200
|
| 1130 |
-
},
|
| 1131 |
-
{
|
| 1132 |
-
"epoch": 0.26,
|
| 1133 |
-
"grad_norm": 32.61063003540039,
|
| 1134 |
-
"learning_rate": 2.8670049374545873e-05,
|
| 1135 |
-
"loss": 3.5054,
|
| 1136 |
-
"step": 3220
|
| 1137 |
-
},
|
| 1138 |
-
{
|
| 1139 |
-
"epoch": 0.26,
|
| 1140 |
-
"grad_norm": 47.47910690307617,
|
| 1141 |
-
"learning_rate": 2.8649978628719256e-05,
|
| 1142 |
-
"loss": 3.5674,
|
| 1143 |
-
"step": 3240
|
| 1144 |
-
},
|
| 1145 |
-
{
|
| 1146 |
-
"epoch": 0.26,
|
| 1147 |
-
"grad_norm": 40.817115783691406,
|
| 1148 |
-
"learning_rate": 2.8629764701075885e-05,
|
| 1149 |
-
"loss": 3.4504,
|
| 1150 |
-
"step": 3260
|
| 1151 |
-
},
|
| 1152 |
-
{
|
| 1153 |
-
"epoch": 0.26,
|
| 1154 |
-
"grad_norm": 32.290626525878906,
|
| 1155 |
-
"learning_rate": 2.8609407803650295e-05,
|
| 1156 |
-
"loss": 3.4699,
|
| 1157 |
-
"step": 3280
|
| 1158 |
-
},
|
| 1159 |
-
{
|
| 1160 |
-
"epoch": 0.26,
|
| 1161 |
-
"grad_norm": 26.942697525024414,
|
| 1162 |
-
"learning_rate": 2.8588908149976702e-05,
|
| 1163 |
-
"loss": 3.5642,
|
| 1164 |
-
"step": 3300
|
| 1165 |
-
},
|
| 1166 |
-
{
|
| 1167 |
-
"epoch": 0.27,
|
| 1168 |
-
"grad_norm": 30.925710678100586,
|
| 1169 |
-
"learning_rate": 2.856826595508678e-05,
|
| 1170 |
-
"loss": 3.6097,
|
| 1171 |
-
"step": 3320
|
| 1172 |
-
},
|
| 1173 |
-
{
|
| 1174 |
-
"epoch": 0.27,
|
| 1175 |
-
"grad_norm": 42.991546630859375,
|
| 1176 |
-
"learning_rate": 2.8547481435507382e-05,
|
| 1177 |
-
"loss": 3.5292,
|
| 1178 |
-
"step": 3340
|
| 1179 |
-
},
|
| 1180 |
-
{
|
| 1181 |
-
"epoch": 0.27,
|
| 1182 |
-
"grad_norm": 23.902685165405273,
|
| 1183 |
-
"learning_rate": 2.852655480925828e-05,
|
| 1184 |
-
"loss": 3.5195,
|
| 1185 |
-
"step": 3360
|
| 1186 |
-
},
|
| 1187 |
-
{
|
| 1188 |
-
"epoch": 0.27,
|
| 1189 |
-
"grad_norm": 36.46931076049805,
|
| 1190 |
-
"learning_rate": 2.8505486295849884e-05,
|
| 1191 |
-
"loss": 3.5451,
|
| 1192 |
-
"step": 3380
|
| 1193 |
-
},
|
| 1194 |
-
{
|
| 1195 |
-
"epoch": 0.27,
|
| 1196 |
-
"grad_norm": 35.51163101196289,
|
| 1197 |
-
"learning_rate": 2.848427611628093e-05,
|
| 1198 |
-
"loss": 3.4868,
|
| 1199 |
-
"step": 3400
|
| 1200 |
-
},
|
| 1201 |
-
{
|
| 1202 |
-
"epoch": 0.27,
|
| 1203 |
-
"grad_norm": 28.35199546813965,
|
| 1204 |
-
"learning_rate": 2.8462924493036168e-05,
|
| 1205 |
-
"loss": 3.4984,
|
| 1206 |
-
"step": 3420
|
| 1207 |
-
},
|
| 1208 |
-
{
|
| 1209 |
-
"epoch": 0.28,
|
| 1210 |
-
"grad_norm": 40.32368850708008,
|
| 1211 |
-
"learning_rate": 2.8441431650084018e-05,
|
| 1212 |
-
"loss": 3.4839,
|
| 1213 |
-
"step": 3440
|
| 1214 |
-
},
|
| 1215 |
-
{
|
| 1216 |
-
"epoch": 0.28,
|
| 1217 |
-
"grad_norm": 46.23302459716797,
|
| 1218 |
-
"learning_rate": 2.841979781287424e-05,
|
| 1219 |
-
"loss": 3.5411,
|
| 1220 |
-
"step": 3460
|
| 1221 |
-
},
|
| 1222 |
-
{
|
| 1223 |
-
"epoch": 0.28,
|
| 1224 |
-
"grad_norm": 36.41411209106445,
|
| 1225 |
-
"learning_rate": 2.8398023208335537e-05,
|
| 1226 |
-
"loss": 3.5701,
|
| 1227 |
-
"step": 3480
|
| 1228 |
-
},
|
| 1229 |
-
{
|
| 1230 |
-
"epoch": 0.28,
|
| 1231 |
-
"grad_norm": 29.031347274780273,
|
| 1232 |
-
"learning_rate": 2.8376108064873216e-05,
|
| 1233 |
-
"loss": 3.5228,
|
| 1234 |
-
"step": 3500
|
| 1235 |
-
},
|
| 1236 |
-
{
|
| 1237 |
-
"epoch": 0.28,
|
| 1238 |
-
"grad_norm": 32.50440216064453,
|
| 1239 |
-
"learning_rate": 2.835405261236676e-05,
|
| 1240 |
-
"loss": 3.479,
|
| 1241 |
-
"step": 3520
|
| 1242 |
-
},
|
| 1243 |
-
{
|
| 1244 |
-
"epoch": 0.28,
|
| 1245 |
-
"grad_norm": 30.749858856201172,
|
| 1246 |
-
"learning_rate": 2.833185708216743e-05,
|
| 1247 |
-
"loss": 3.53,
|
| 1248 |
-
"step": 3540
|
| 1249 |
-
},
|
| 1250 |
-
{
|
| 1251 |
-
"epoch": 0.28,
|
| 1252 |
-
"grad_norm": 27.148874282836914,
|
| 1253 |
-
"learning_rate": 2.8309521707095835e-05,
|
| 1254 |
-
"loss": 3.4933,
|
| 1255 |
-
"step": 3560
|
| 1256 |
-
},
|
| 1257 |
-
{
|
| 1258 |
-
"epoch": 0.29,
|
| 1259 |
-
"grad_norm": 28.534822463989258,
|
| 1260 |
-
"learning_rate": 2.8287046721439487e-05,
|
| 1261 |
-
"loss": 3.5881,
|
| 1262 |
-
"step": 3580
|
| 1263 |
-
},
|
| 1264 |
-
{
|
| 1265 |
-
"epoch": 0.29,
|
| 1266 |
-
"grad_norm": 53.12272644042969,
|
| 1267 |
-
"learning_rate": 2.8264432360950355e-05,
|
| 1268 |
-
"loss": 3.5626,
|
| 1269 |
-
"step": 3600
|
| 1270 |
-
},
|
| 1271 |
-
{
|
| 1272 |
-
"epoch": 0.29,
|
| 1273 |
-
"grad_norm": 27.4322509765625,
|
| 1274 |
-
"learning_rate": 2.8241678862842374e-05,
|
| 1275 |
-
"loss": 3.4831,
|
| 1276 |
-
"step": 3620
|
| 1277 |
-
},
|
| 1278 |
-
{
|
| 1279 |
-
"epoch": 0.29,
|
| 1280 |
-
"grad_norm": 31.194787979125977,
|
| 1281 |
-
"learning_rate": 2.8218786465788984e-05,
|
| 1282 |
-
"loss": 3.6001,
|
| 1283 |
-
"step": 3640
|
| 1284 |
-
},
|
| 1285 |
-
{
|
| 1286 |
-
"epoch": 0.29,
|
| 1287 |
-
"grad_norm": 27.951440811157227,
|
| 1288 |
-
"learning_rate": 2.8195755409920584e-05,
|
| 1289 |
-
"loss": 3.4387,
|
| 1290 |
-
"step": 3660
|
| 1291 |
-
},
|
| 1292 |
-
{
|
| 1293 |
-
"epoch": 0.29,
|
| 1294 |
-
"grad_norm": 43.26001739501953,
|
| 1295 |
-
"learning_rate": 2.8172585936822056e-05,
|
| 1296 |
-
"loss": 3.5127,
|
| 1297 |
-
"step": 3680
|
| 1298 |
-
},
|
| 1299 |
-
{
|
| 1300 |
-
"epoch": 0.3,
|
| 1301 |
-
"grad_norm": 30.359540939331055,
|
| 1302 |
-
"learning_rate": 2.814927828953022e-05,
|
| 1303 |
-
"loss": 3.4761,
|
| 1304 |
-
"step": 3700
|
| 1305 |
-
},
|
| 1306 |
-
{
|
| 1307 |
-
"epoch": 0.3,
|
| 1308 |
-
"grad_norm": 29.243453979492188,
|
| 1309 |
-
"learning_rate": 2.812583271253125e-05,
|
| 1310 |
-
"loss": 3.6265,
|
| 1311 |
-
"step": 3720
|
| 1312 |
-
},
|
| 1313 |
-
{
|
| 1314 |
-
"epoch": 0.3,
|
| 1315 |
-
"grad_norm": 25.85703468322754,
|
| 1316 |
-
"learning_rate": 2.8102249451758162e-05,
|
| 1317 |
-
"loss": 3.5619,
|
| 1318 |
-
"step": 3740
|
| 1319 |
-
},
|
| 1320 |
-
{
|
| 1321 |
-
"epoch": 0.3,
|
| 1322 |
-
"grad_norm": 32.1122932434082,
|
| 1323 |
-
"learning_rate": 2.8078528754588207e-05,
|
| 1324 |
-
"loss": 3.5173,
|
| 1325 |
-
"step": 3760
|
| 1326 |
-
},
|
| 1327 |
-
{
|
| 1328 |
-
"epoch": 0.3,
|
| 1329 |
-
"grad_norm": 28.02115821838379,
|
| 1330 |
-
"learning_rate": 2.805467086984027e-05,
|
| 1331 |
-
"loss": 3.4905,
|
| 1332 |
-
"step": 3780
|
| 1333 |
-
},
|
| 1334 |
-
{
|
| 1335 |
-
"epoch": 0.3,
|
| 1336 |
-
"grad_norm": 41.776817321777344,
|
| 1337 |
-
"learning_rate": 2.803067604777227e-05,
|
| 1338 |
-
"loss": 3.5289,
|
| 1339 |
-
"step": 3800
|
| 1340 |
-
},
|
| 1341 |
-
{
|
| 1342 |
-
"epoch": 0.31,
|
| 1343 |
-
"grad_norm": 33.876312255859375,
|
| 1344 |
-
"learning_rate": 2.8006544540078535e-05,
|
| 1345 |
-
"loss": 3.5414,
|
| 1346 |
-
"step": 3820
|
| 1347 |
-
},
|
| 1348 |
-
{
|
| 1349 |
-
"epoch": 0.31,
|
| 1350 |
-
"grad_norm": 33.06633758544922,
|
| 1351 |
-
"learning_rate": 2.798227659988717e-05,
|
| 1352 |
-
"loss": 3.5586,
|
| 1353 |
-
"step": 3840
|
| 1354 |
-
},
|
| 1355 |
-
{
|
| 1356 |
-
"epoch": 0.31,
|
| 1357 |
-
"grad_norm": 24.628082275390625,
|
| 1358 |
-
"learning_rate": 2.7957872481757377e-05,
|
| 1359 |
-
"loss": 3.5455,
|
| 1360 |
-
"step": 3860
|
| 1361 |
-
},
|
| 1362 |
-
{
|
| 1363 |
-
"epoch": 0.31,
|
| 1364 |
-
"grad_norm": 23.779598236083984,
|
| 1365 |
-
"learning_rate": 2.793333244167681e-05,
|
| 1366 |
-
"loss": 3.5345,
|
| 1367 |
-
"step": 3880
|
| 1368 |
-
},
|
| 1369 |
-
{
|
| 1370 |
-
"epoch": 0.31,
|
| 1371 |
-
"grad_norm": 29.29547119140625,
|
| 1372 |
-
"learning_rate": 2.790865673705888e-05,
|
| 1373 |
-
"loss": 3.4588,
|
| 1374 |
-
"step": 3900
|
| 1375 |
-
},
|
| 1376 |
-
{
|
| 1377 |
-
"epoch": 0.31,
|
| 1378 |
-
"grad_norm": 29.236722946166992,
|
| 1379 |
-
"learning_rate": 2.7883845626740046e-05,
|
| 1380 |
-
"loss": 3.5286,
|
| 1381 |
-
"step": 3920
|
| 1382 |
-
},
|
| 1383 |
-
{
|
| 1384 |
-
"epoch": 0.32,
|
| 1385 |
-
"grad_norm": 36.6258430480957,
|
| 1386 |
-
"learning_rate": 2.7858899370977123e-05,
|
| 1387 |
-
"loss": 3.589,
|
| 1388 |
-
"step": 3940
|
| 1389 |
-
},
|
| 1390 |
-
{
|
| 1391 |
-
"epoch": 0.32,
|
| 1392 |
-
"grad_norm": 48.60535430908203,
|
| 1393 |
-
"learning_rate": 2.783381823144452e-05,
|
| 1394 |
-
"loss": 3.6398,
|
| 1395 |
-
"step": 3960
|
| 1396 |
-
},
|
| 1397 |
-
{
|
| 1398 |
-
"epoch": 0.32,
|
| 1399 |
-
"grad_norm": 28.959075927734375,
|
| 1400 |
-
"learning_rate": 2.780860247123153e-05,
|
| 1401 |
-
"loss": 3.4484,
|
| 1402 |
-
"step": 3980
|
| 1403 |
-
},
|
| 1404 |
-
{
|
| 1405 |
-
"epoch": 0.32,
|
| 1406 |
-
"grad_norm": 30.727558135986328,
|
| 1407 |
-
"learning_rate": 2.778325235483954e-05,
|
| 1408 |
-
"loss": 3.5112,
|
| 1409 |
-
"step": 4000
|
| 1410 |
-
},
|
| 1411 |
-
{
|
| 1412 |
-
"epoch": 0.32,
|
| 1413 |
-
"grad_norm": 32.5091667175293,
|
| 1414 |
-
"learning_rate": 2.775776814817928e-05,
|
| 1415 |
-
"loss": 3.4266,
|
| 1416 |
-
"step": 4020
|
| 1417 |
-
},
|
| 1418 |
-
{
|
| 1419 |
-
"epoch": 0.32,
|
| 1420 |
-
"grad_norm": 37.957637786865234,
|
| 1421 |
-
"learning_rate": 2.7732150118568016e-05,
|
| 1422 |
-
"loss": 3.6768,
|
| 1423 |
-
"step": 4040
|
| 1424 |
-
},
|
| 1425 |
-
{
|
| 1426 |
-
"epoch": 0.32,
|
| 1427 |
-
"grad_norm": 42.639320373535156,
|
| 1428 |
-
"learning_rate": 2.770639853472676e-05,
|
| 1429 |
-
"loss": 3.5102,
|
| 1430 |
-
"step": 4060
|
| 1431 |
-
},
|
| 1432 |
-
{
|
| 1433 |
-
"epoch": 0.33,
|
| 1434 |
-
"grad_norm": 29.558870315551758,
|
| 1435 |
-
"learning_rate": 2.768051366677744e-05,
|
| 1436 |
-
"loss": 3.5354,
|
| 1437 |
-
"step": 4080
|
| 1438 |
-
},
|
| 1439 |
-
{
|
| 1440 |
-
"epoch": 0.33,
|
| 1441 |
-
"grad_norm": 23.12784194946289,
|
| 1442 |
-
"learning_rate": 2.765449578624007e-05,
|
| 1443 |
-
"loss": 3.5432,
|
| 1444 |
-
"step": 4100
|
| 1445 |
-
},
|
| 1446 |
-
{
|
| 1447 |
-
"epoch": 0.33,
|
| 1448 |
-
"grad_norm": 27.601444244384766,
|
| 1449 |
-
"learning_rate": 2.7628345166029907e-05,
|
| 1450 |
-
"loss": 3.5672,
|
| 1451 |
-
"step": 4120
|
| 1452 |
-
},
|
| 1453 |
-
{
|
| 1454 |
-
"epoch": 0.33,
|
| 1455 |
-
"grad_norm": 26.26235008239746,
|
| 1456 |
-
"learning_rate": 2.760206208045458e-05,
|
| 1457 |
-
"loss": 3.5635,
|
| 1458 |
-
"step": 4140
|
| 1459 |
-
},
|
| 1460 |
-
{
|
| 1461 |
-
"epoch": 0.33,
|
| 1462 |
-
"grad_norm": 57.84916305541992,
|
| 1463 |
-
"learning_rate": 2.7575646805211224e-05,
|
| 1464 |
-
"loss": 3.5254,
|
| 1465 |
-
"step": 4160
|
| 1466 |
-
},
|
| 1467 |
-
{
|
| 1468 |
-
"epoch": 0.33,
|
| 1469 |
-
"grad_norm": 25.554025650024414,
|
| 1470 |
-
"learning_rate": 2.7549099617383573e-05,
|
| 1471 |
-
"loss": 3.5142,
|
| 1472 |
-
"step": 4180
|
| 1473 |
-
},
|
| 1474 |
-
{
|
| 1475 |
-
"epoch": 0.34,
|
| 1476 |
-
"grad_norm": 38.82815170288086,
|
| 1477 |
-
"learning_rate": 2.7522420795439067e-05,
|
| 1478 |
-
"loss": 3.6104,
|
| 1479 |
-
"step": 4200
|
| 1480 |
-
},
|
| 1481 |
-
{
|
| 1482 |
-
"epoch": 0.34,
|
| 1483 |
-
"grad_norm": 28.093948364257812,
|
| 1484 |
-
"learning_rate": 2.7495610619225925e-05,
|
| 1485 |
-
"loss": 3.5265,
|
| 1486 |
-
"step": 4220
|
| 1487 |
-
},
|
| 1488 |
-
{
|
| 1489 |
-
"epoch": 0.34,
|
| 1490 |
-
"grad_norm": 26.187891006469727,
|
| 1491 |
-
"learning_rate": 2.746866936997021e-05,
|
| 1492 |
-
"loss": 3.4307,
|
| 1493 |
-
"step": 4240
|
| 1494 |
-
},
|
| 1495 |
-
{
|
| 1496 |
-
"epoch": 0.34,
|
| 1497 |
-
"grad_norm": 35.642738342285156,
|
| 1498 |
-
"learning_rate": 2.7441597330272874e-05,
|
| 1499 |
-
"loss": 3.5501,
|
| 1500 |
-
"step": 4260
|
| 1501 |
-
},
|
| 1502 |
-
{
|
| 1503 |
-
"epoch": 0.34,
|
| 1504 |
-
"grad_norm": 32.99201965332031,
|
| 1505 |
-
"learning_rate": 2.7414394784106812e-05,
|
| 1506 |
-
"loss": 3.4463,
|
| 1507 |
-
"step": 4280
|
| 1508 |
-
},
|
| 1509 |
-
{
|
| 1510 |
-
"epoch": 0.34,
|
| 1511 |
-
"grad_norm": 28.848899841308594,
|
| 1512 |
-
"learning_rate": 2.7387062016813845e-05,
|
| 1513 |
-
"loss": 3.5128,
|
| 1514 |
-
"step": 4300
|
| 1515 |
-
},
|
| 1516 |
-
{
|
| 1517 |
-
"epoch": 0.35,
|
| 1518 |
-
"grad_norm": 30.502288818359375,
|
| 1519 |
-
"learning_rate": 2.7359599315101788e-05,
|
| 1520 |
-
"loss": 3.4909,
|
| 1521 |
-
"step": 4320
|
| 1522 |
-
},
|
| 1523 |
-
{
|
| 1524 |
-
"epoch": 0.35,
|
| 1525 |
-
"grad_norm": 27.91356658935547,
|
| 1526 |
-
"learning_rate": 2.7332006967041373e-05,
|
| 1527 |
-
"loss": 3.53,
|
| 1528 |
-
"step": 4340
|
| 1529 |
-
},
|
| 1530 |
-
{
|
| 1531 |
-
"epoch": 0.35,
|
| 1532 |
-
"grad_norm": 47.296627044677734,
|
| 1533 |
-
"learning_rate": 2.7304285262063274e-05,
|
| 1534 |
-
"loss": 3.4793,
|
| 1535 |
-
"step": 4360
|
| 1536 |
-
},
|
| 1537 |
-
{
|
| 1538 |
-
"epoch": 0.35,
|
| 1539 |
-
"grad_norm": 33.32771682739258,
|
| 1540 |
-
"learning_rate": 2.7276434490955074e-05,
|
| 1541 |
-
"loss": 3.4695,
|
| 1542 |
-
"step": 4380
|
| 1543 |
-
},
|
| 1544 |
-
{
|
| 1545 |
-
"epoch": 0.35,
|
| 1546 |
-
"grad_norm": 36.65375518798828,
|
| 1547 |
-
"learning_rate": 2.7248454945858164e-05,
|
| 1548 |
-
"loss": 3.4502,
|
| 1549 |
-
"step": 4400
|
| 1550 |
-
},
|
| 1551 |
-
{
|
| 1552 |
-
"epoch": 0.35,
|
| 1553 |
-
"grad_norm": 62.65798568725586,
|
| 1554 |
-
"learning_rate": 2.7220346920264743e-05,
|
| 1555 |
-
"loss": 3.5893,
|
| 1556 |
-
"step": 4420
|
| 1557 |
-
},
|
| 1558 |
-
{
|
| 1559 |
-
"epoch": 0.36,
|
| 1560 |
-
"grad_norm": 26.921863555908203,
|
| 1561 |
-
"learning_rate": 2.71921107090147e-05,
|
| 1562 |
-
"loss": 3.4381,
|
| 1563 |
-
"step": 4440
|
| 1564 |
-
},
|
| 1565 |
-
{
|
| 1566 |
-
"epoch": 0.36,
|
| 1567 |
-
"grad_norm": 35.91081619262695,
|
| 1568 |
-
"learning_rate": 2.7163746608292525e-05,
|
| 1569 |
-
"loss": 3.5292,
|
| 1570 |
-
"step": 4460
|
| 1571 |
-
},
|
| 1572 |
-
{
|
| 1573 |
-
"epoch": 0.36,
|
| 1574 |
-
"grad_norm": 42.172306060791016,
|
| 1575 |
-
"learning_rate": 2.7135254915624213e-05,
|
| 1576 |
-
"loss": 3.5314,
|
| 1577 |
-
"step": 4480
|
| 1578 |
-
},
|
| 1579 |
-
{
|
| 1580 |
-
"epoch": 0.36,
|
| 1581 |
-
"grad_norm": 65.17137908935547,
|
| 1582 |
-
"learning_rate": 2.710663592987414e-05,
|
| 1583 |
-
"loss": 3.518,
|
| 1584 |
-
"step": 4500
|
| 1585 |
-
},
|
| 1586 |
-
{
|
| 1587 |
-
"epoch": 0.36,
|
| 1588 |
-
"grad_norm": 32.53944396972656,
|
| 1589 |
-
"learning_rate": 2.7077889951241924e-05,
|
| 1590 |
-
"loss": 3.5562,
|
| 1591 |
-
"step": 4520
|
| 1592 |
-
},
|
| 1593 |
-
{
|
| 1594 |
-
"epoch": 0.36,
|
| 1595 |
-
"grad_norm": 25.663211822509766,
|
| 1596 |
-
"learning_rate": 2.704901728125928e-05,
|
| 1597 |
-
"loss": 3.5537,
|
| 1598 |
-
"step": 4540
|
| 1599 |
-
},
|
| 1600 |
-
{
|
| 1601 |
-
"epoch": 0.36,
|
| 1602 |
-
"grad_norm": 23.626951217651367,
|
| 1603 |
-
"learning_rate": 2.702001822278685e-05,
|
| 1604 |
-
"loss": 3.5525,
|
| 1605 |
-
"step": 4560
|
| 1606 |
-
},
|
| 1607 |
-
{
|
| 1608 |
-
"epoch": 0.37,
|
| 1609 |
-
"grad_norm": 30.527162551879883,
|
| 1610 |
-
"learning_rate": 2.699089308001104e-05,
|
| 1611 |
-
"loss": 3.4913,
|
| 1612 |
-
"step": 4580
|
| 1613 |
-
},
|
| 1614 |
-
{
|
| 1615 |
-
"epoch": 0.37,
|
| 1616 |
-
"grad_norm": 37.62814712524414,
|
| 1617 |
-
"learning_rate": 2.696164215844081e-05,
|
| 1618 |
-
"loss": 3.5342,
|
| 1619 |
-
"step": 4600
|
| 1620 |
-
},
|
| 1621 |
-
{
|
| 1622 |
-
"epoch": 0.37,
|
| 1623 |
-
"grad_norm": 26.47550392150879,
|
| 1624 |
-
"learning_rate": 2.6932265764904494e-05,
|
| 1625 |
-
"loss": 3.4708,
|
| 1626 |
-
"step": 4620
|
| 1627 |
-
},
|
| 1628 |
-
{
|
| 1629 |
-
"epoch": 0.37,
|
| 1630 |
-
"grad_norm": 30.779155731201172,
|
| 1631 |
-
"learning_rate": 2.6902764207546553e-05,
|
| 1632 |
-
"loss": 3.5078,
|
| 1633 |
-
"step": 4640
|
| 1634 |
-
},
|
| 1635 |
-
{
|
| 1636 |
-
"epoch": 0.37,
|
| 1637 |
-
"grad_norm": 34.16841506958008,
|
| 1638 |
-
"learning_rate": 2.6873137795824367e-05,
|
| 1639 |
-
"loss": 3.4754,
|
| 1640 |
-
"step": 4660
|
| 1641 |
-
},
|
| 1642 |
-
{
|
| 1643 |
-
"epoch": 0.37,
|
| 1644 |
-
"grad_norm": 36.18644714355469,
|
| 1645 |
-
"learning_rate": 2.6843386840504972e-05,
|
| 1646 |
-
"loss": 3.4413,
|
| 1647 |
-
"step": 4680
|
| 1648 |
-
},
|
| 1649 |
-
{
|
| 1650 |
-
"epoch": 0.38,
|
| 1651 |
-
"grad_norm": 34.17078399658203,
|
| 1652 |
-
"learning_rate": 2.6813511653661817e-05,
|
| 1653 |
-
"loss": 3.4916,
|
| 1654 |
-
"step": 4700
|
| 1655 |
-
},
|
| 1656 |
-
{
|
| 1657 |
-
"epoch": 0.38,
|
| 1658 |
-
"grad_norm": 24.693265914916992,
|
| 1659 |
-
"learning_rate": 2.678351254867147e-05,
|
| 1660 |
-
"loss": 3.4072,
|
| 1661 |
-
"step": 4720
|
| 1662 |
-
},
|
| 1663 |
-
{
|
| 1664 |
-
"epoch": 0.38,
|
| 1665 |
-
"grad_norm": 27.831270217895508,
|
| 1666 |
-
"learning_rate": 2.675338984021035e-05,
|
| 1667 |
-
"loss": 3.5353,
|
| 1668 |
-
"step": 4740
|
| 1669 |
-
},
|
| 1670 |
-
{
|
| 1671 |
-
"epoch": 0.38,
|
| 1672 |
-
"grad_norm": 18.52642059326172,
|
| 1673 |
-
"learning_rate": 2.672314384425142e-05,
|
| 1674 |
-
"loss": 3.4582,
|
| 1675 |
-
"step": 4760
|
| 1676 |
-
},
|
| 1677 |
-
{
|
| 1678 |
-
"epoch": 0.38,
|
| 1679 |
-
"grad_norm": 44.86159133911133,
|
| 1680 |
-
"learning_rate": 2.669277487806085e-05,
|
| 1681 |
-
"loss": 3.4384,
|
| 1682 |
-
"step": 4780
|
| 1683 |
-
},
|
| 1684 |
-
{
|
| 1685 |
-
"epoch": 0.38,
|
| 1686 |
-
"grad_norm": 28.123258590698242,
|
| 1687 |
-
"learning_rate": 2.6662283260194743e-05,
|
| 1688 |
-
"loss": 3.5766,
|
| 1689 |
-
"step": 4800
|
| 1690 |
-
},
|
| 1691 |
-
{
|
| 1692 |
-
"epoch": 0.39,
|
| 1693 |
-
"grad_norm": 27.150848388671875,
|
| 1694 |
-
"learning_rate": 2.6631669310495725e-05,
|
| 1695 |
-
"loss": 3.5095,
|
| 1696 |
-
"step": 4820
|
| 1697 |
-
},
|
| 1698 |
-
{
|
| 1699 |
-
"epoch": 0.39,
|
| 1700 |
-
"grad_norm": 43.018043518066406,
|
| 1701 |
-
"learning_rate": 2.660093335008966e-05,
|
| 1702 |
-
"loss": 3.4795,
|
| 1703 |
-
"step": 4840
|
| 1704 |
-
},
|
| 1705 |
-
{
|
| 1706 |
-
"epoch": 0.39,
|
| 1707 |
-
"grad_norm": 29.27479362487793,
|
| 1708 |
-
"learning_rate": 2.6570075701382213e-05,
|
| 1709 |
-
"loss": 3.5236,
|
| 1710 |
-
"step": 4860
|
| 1711 |
-
},
|
| 1712 |
-
{
|
| 1713 |
-
"epoch": 0.39,
|
| 1714 |
-
"grad_norm": 19.501262664794922,
|
| 1715 |
-
"learning_rate": 2.653909668805553e-05,
|
| 1716 |
-
"loss": 3.5479,
|
| 1717 |
-
"step": 4880
|
| 1718 |
-
},
|
| 1719 |
-
{
|
| 1720 |
-
"epoch": 0.39,
|
| 1721 |
-
"grad_norm": 57.28257369995117,
|
| 1722 |
-
"learning_rate": 2.6507996635064792e-05,
|
| 1723 |
-
"loss": 3.5156,
|
| 1724 |
-
"step": 4900
|
| 1725 |
-
},
|
| 1726 |
-
{
|
| 1727 |
-
"epoch": 0.39,
|
| 1728 |
-
"grad_norm": 27.764036178588867,
|
| 1729 |
-
"learning_rate": 2.647677586863484e-05,
|
| 1730 |
-
"loss": 3.5222,
|
| 1731 |
-
"step": 4920
|
| 1732 |
-
},
|
| 1733 |
-
{
|
| 1734 |
-
"epoch": 0.4,
|
| 1735 |
-
"grad_norm": 33.74861526489258,
|
| 1736 |
-
"learning_rate": 2.644543471625675e-05,
|
| 1737 |
-
"loss": 3.4773,
|
| 1738 |
-
"step": 4940
|
| 1739 |
-
},
|
| 1740 |
-
{
|
| 1741 |
-
"epoch": 0.4,
|
| 1742 |
-
"grad_norm": 25.404314041137695,
|
| 1743 |
-
"learning_rate": 2.6413973506684366e-05,
|
| 1744 |
-
"loss": 3.4646,
|
| 1745 |
-
"step": 4960
|
| 1746 |
-
},
|
| 1747 |
-
{
|
| 1748 |
-
"epoch": 0.4,
|
| 1749 |
-
"grad_norm": 33.307674407958984,
|
| 1750 |
-
"learning_rate": 2.63823925699309e-05,
|
| 1751 |
-
"loss": 3.4975,
|
| 1752 |
-
"step": 4980
|
| 1753 |
-
},
|
| 1754 |
-
{
|
| 1755 |
-
"epoch": 0.4,
|
| 1756 |
-
"grad_norm": 28.22442054748535,
|
| 1757 |
-
"learning_rate": 2.6350692237265428e-05,
|
| 1758 |
-
"loss": 3.4797,
|
| 1759 |
-
"step": 5000
|
| 1760 |
-
},
|
| 1761 |
-
{
|
| 1762 |
-
"epoch": 0.4,
|
| 1763 |
-
"grad_norm": 26.52558135986328,
|
| 1764 |
-
"learning_rate": 2.6318872841209446e-05,
|
| 1765 |
-
"loss": 3.4309,
|
| 1766 |
-
"step": 5020
|
| 1767 |
-
},
|
| 1768 |
-
{
|
| 1769 |
-
"epoch": 0.4,
|
| 1770 |
-
"grad_norm": 36.679386138916016,
|
| 1771 |
-
"learning_rate": 2.6286934715533353e-05,
|
| 1772 |
-
"loss": 3.585,
|
| 1773 |
-
"step": 5040
|
| 1774 |
-
},
|
| 1775 |
-
{
|
| 1776 |
-
"epoch": 0.4,
|
| 1777 |
-
"grad_norm": 42.78778839111328,
|
| 1778 |
-
"learning_rate": 2.6254878195252985e-05,
|
| 1779 |
-
"loss": 3.4239,
|
| 1780 |
-
"step": 5060
|
| 1781 |
-
},
|
| 1782 |
-
{
|
| 1783 |
-
"epoch": 0.41,
|
| 1784 |
-
"grad_norm": 34.719482421875,
|
| 1785 |
-
"learning_rate": 2.622270361662606e-05,
|
| 1786 |
-
"loss": 3.4777,
|
| 1787 |
-
"step": 5080
|
| 1788 |
-
},
|
| 1789 |
-
{
|
| 1790 |
-
"epoch": 0.41,
|
| 1791 |
-
"grad_norm": 33.207427978515625,
|
| 1792 |
-
"learning_rate": 2.619041131714869e-05,
|
| 1793 |
-
"loss": 3.5593,
|
| 1794 |
-
"step": 5100
|
| 1795 |
-
},
|
| 1796 |
-
{
|
| 1797 |
-
"epoch": 0.41,
|
| 1798 |
-
"grad_norm": 35.62514877319336,
|
| 1799 |
-
"learning_rate": 2.6158001635551818e-05,
|
| 1800 |
-
"loss": 3.5606,
|
| 1801 |
-
"step": 5120
|
| 1802 |
-
},
|
| 1803 |
-
{
|
| 1804 |
-
"epoch": 0.41,
|
| 1805 |
-
"grad_norm": 31.691574096679688,
|
| 1806 |
-
"learning_rate": 2.6125474911797664e-05,
|
| 1807 |
-
"loss": 3.4959,
|
| 1808 |
-
"step": 5140
|
| 1809 |
-
},
|
| 1810 |
-
{
|
| 1811 |
-
"epoch": 0.41,
|
| 1812 |
-
"grad_norm": 34.012420654296875,
|
| 1813 |
-
"learning_rate": 2.6092831487076163e-05,
|
| 1814 |
-
"loss": 3.57,
|
| 1815 |
-
"step": 5160
|
| 1816 |
-
},
|
| 1817 |
-
{
|
| 1818 |
-
"epoch": 0.41,
|
| 1819 |
-
"grad_norm": 36.75544357299805,
|
| 1820 |
-
"learning_rate": 2.6060071703801406e-05,
|
| 1821 |
-
"loss": 3.4718,
|
| 1822 |
-
"step": 5180
|
| 1823 |
-
},
|
| 1824 |
-
{
|
| 1825 |
-
"epoch": 0.42,
|
| 1826 |
-
"grad_norm": 37.18219757080078,
|
| 1827 |
-
"learning_rate": 2.6027195905608006e-05,
|
| 1828 |
-
"loss": 3.5332,
|
| 1829 |
-
"step": 5200
|
| 1830 |
-
},
|
| 1831 |
-
{
|
| 1832 |
-
"epoch": 0.42,
|
| 1833 |
-
"grad_norm": 32.344398498535156,
|
| 1834 |
-
"learning_rate": 2.599420443734754e-05,
|
| 1835 |
-
"loss": 3.5154,
|
| 1836 |
-
"step": 5220
|
| 1837 |
-
},
|
| 1838 |
-
{
|
| 1839 |
-
"epoch": 0.42,
|
| 1840 |
-
"grad_norm": 26.169748306274414,
|
| 1841 |
-
"learning_rate": 2.596109764508489e-05,
|
| 1842 |
-
"loss": 3.5462,
|
| 1843 |
-
"step": 5240
|
| 1844 |
-
},
|
| 1845 |
-
{
|
| 1846 |
-
"epoch": 0.42,
|
| 1847 |
-
"grad_norm": 33.38447570800781,
|
| 1848 |
-
"learning_rate": 2.592787587609465e-05,
|
| 1849 |
-
"loss": 3.5658,
|
| 1850 |
-
"step": 5260
|
| 1851 |
-
},
|
| 1852 |
-
{
|
| 1853 |
-
"epoch": 0.42,
|
| 1854 |
-
"grad_norm": 43.4962158203125,
|
| 1855 |
-
"learning_rate": 2.589453947885745e-05,
|
| 1856 |
-
"loss": 3.5018,
|
| 1857 |
-
"step": 5280
|
| 1858 |
-
},
|
| 1859 |
-
{
|
| 1860 |
-
"epoch": 0.42,
|
| 1861 |
-
"grad_norm": 34.955291748046875,
|
| 1862 |
-
"learning_rate": 2.5861088803056324e-05,
|
| 1863 |
-
"loss": 3.4988,
|
| 1864 |
-
"step": 5300
|
| 1865 |
-
},
|
| 1866 |
-
{
|
| 1867 |
-
"epoch": 0.43,
|
| 1868 |
-
"grad_norm": 29.175064086914062,
|
| 1869 |
-
"learning_rate": 2.5827524199573033e-05,
|
| 1870 |
-
"loss": 3.475,
|
| 1871 |
-
"step": 5320
|
| 1872 |
-
},
|
| 1873 |
-
{
|
| 1874 |
-
"epoch": 0.43,
|
| 1875 |
-
"grad_norm": 40.04597091674805,
|
| 1876 |
-
"learning_rate": 2.5793846020484383e-05,
|
| 1877 |
-
"loss": 3.4903,
|
| 1878 |
-
"step": 5340
|
| 1879 |
-
},
|
| 1880 |
-
{
|
| 1881 |
-
"epoch": 0.43,
|
| 1882 |
-
"grad_norm": 27.712465286254883,
|
| 1883 |
-
"learning_rate": 2.5760054619058537e-05,
|
| 1884 |
-
"loss": 3.5108,
|
| 1885 |
-
"step": 5360
|
| 1886 |
-
},
|
| 1887 |
-
{
|
| 1888 |
-
"epoch": 0.43,
|
| 1889 |
-
"grad_norm": 43.66648864746094,
|
| 1890 |
-
"learning_rate": 2.5726150349751306e-05,
|
| 1891 |
-
"loss": 3.4656,
|
| 1892 |
-
"step": 5380
|
| 1893 |
-
},
|
| 1894 |
-
{
|
| 1895 |
-
"epoch": 0.43,
|
| 1896 |
-
"grad_norm": 28.305545806884766,
|
| 1897 |
-
"learning_rate": 2.569213356820244e-05,
|
| 1898 |
-
"loss": 3.5766,
|
| 1899 |
-
"step": 5400
|
| 1900 |
-
},
|
| 1901 |
-
{
|
| 1902 |
-
"epoch": 0.43,
|
| 1903 |
-
"grad_norm": 34.897857666015625,
|
| 1904 |
-
"learning_rate": 2.565800463123187e-05,
|
| 1905 |
-
"loss": 3.4286,
|
| 1906 |
-
"step": 5420
|
| 1907 |
-
},
|
| 1908 |
-
{
|
| 1909 |
-
"epoch": 0.44,
|
| 1910 |
-
"grad_norm": 43.083229064941406,
|
| 1911 |
-
"learning_rate": 2.5623763896835997e-05,
|
| 1912 |
-
"loss": 3.4292,
|
| 1913 |
-
"step": 5440
|
| 1914 |
-
},
|
| 1915 |
-
{
|
| 1916 |
-
"epoch": 0.44,
|
| 1917 |
-
"grad_norm": 35.71794128417969,
|
| 1918 |
-
"learning_rate": 2.5589411724183926e-05,
|
| 1919 |
-
"loss": 3.5542,
|
| 1920 |
-
"step": 5460
|
| 1921 |
-
},
|
| 1922 |
-
{
|
| 1923 |
-
"epoch": 0.44,
|
| 1924 |
-
"grad_norm": 23.70340347290039,
|
| 1925 |
-
"learning_rate": 2.555494847361369e-05,
|
| 1926 |
-
"loss": 3.5276,
|
| 1927 |
-
"step": 5480
|
| 1928 |
-
},
|
| 1929 |
-
{
|
| 1930 |
-
"epoch": 0.44,
|
| 1931 |
-
"grad_norm": 55.299556732177734,
|
| 1932 |
-
"learning_rate": 2.552037450662849e-05,
|
| 1933 |
-
"loss": 3.5644,
|
| 1934 |
-
"step": 5500
|
| 1935 |
-
},
|
| 1936 |
-
{
|
| 1937 |
-
"epoch": 0.44,
|
| 1938 |
-
"grad_norm": 22.433879852294922,
|
| 1939 |
-
"learning_rate": 2.5485690185892864e-05,
|
| 1940 |
-
"loss": 3.4475,
|
| 1941 |
-
"step": 5520
|
| 1942 |
-
},
|
| 1943 |
-
{
|
| 1944 |
-
"epoch": 0.44,
|
| 1945 |
-
"grad_norm": 26.397438049316406,
|
| 1946 |
-
"learning_rate": 2.545089587522893e-05,
|
| 1947 |
-
"loss": 3.507,
|
| 1948 |
-
"step": 5540
|
| 1949 |
-
},
|
| 1950 |
-
{
|
| 1951 |
-
"epoch": 0.44,
|
| 1952 |
-
"grad_norm": 25.061750411987305,
|
| 1953 |
-
"learning_rate": 2.5415991939612545e-05,
|
| 1954 |
-
"loss": 3.4521,
|
| 1955 |
-
"step": 5560
|
| 1956 |
-
},
|
| 1957 |
-
{
|
| 1958 |
-
"epoch": 0.45,
|
| 1959 |
-
"grad_norm": 35.5067138671875,
|
| 1960 |
-
"learning_rate": 2.5380978745169473e-05,
|
| 1961 |
-
"loss": 3.4894,
|
| 1962 |
-
"step": 5580
|
| 1963 |
-
},
|
| 1964 |
-
{
|
| 1965 |
-
"epoch": 0.45,
|
| 1966 |
-
"grad_norm": 27.13253402709961,
|
| 1967 |
-
"learning_rate": 2.5345856659171567e-05,
|
| 1968 |
-
"loss": 3.4033,
|
| 1969 |
-
"step": 5600
|
| 1970 |
-
},
|
| 1971 |
-
{
|
| 1972 |
-
"epoch": 0.45,
|
| 1973 |
-
"grad_norm": 34.580299377441406,
|
| 1974 |
-
"learning_rate": 2.5310626050032873e-05,
|
| 1975 |
-
"loss": 3.5598,
|
| 1976 |
-
"step": 5620
|
| 1977 |
-
},
|
| 1978 |
-
{
|
| 1979 |
-
"epoch": 0.45,
|
| 1980 |
-
"grad_norm": 21.44734764099121,
|
| 1981 |
-
"learning_rate": 2.527528728730582e-05,
|
| 1982 |
-
"loss": 3.5189,
|
| 1983 |
-
"step": 5640
|
| 1984 |
-
},
|
| 1985 |
-
{
|
| 1986 |
-
"epoch": 0.45,
|
| 1987 |
-
"grad_norm": 26.825183868408203,
|
| 1988 |
-
"learning_rate": 2.5239840741677307e-05,
|
| 1989 |
-
"loss": 3.6052,
|
| 1990 |
-
"step": 5660
|
| 1991 |
-
},
|
| 1992 |
-
{
|
| 1993 |
-
"epoch": 0.45,
|
| 1994 |
-
"grad_norm": 29.19519805908203,
|
| 1995 |
-
"learning_rate": 2.5204286784964823e-05,
|
| 1996 |
-
"loss": 3.4724,
|
| 1997 |
-
"step": 5680
|
| 1998 |
-
},
|
| 1999 |
-
{
|
| 2000 |
-
"epoch": 0.46,
|
| 2001 |
-
"grad_norm": 37.9902229309082,
|
| 2002 |
-
"learning_rate": 2.516862579011255e-05,
|
| 2003 |
-
"loss": 3.3665,
|
| 2004 |
-
"step": 5700
|
| 2005 |
-
},
|
| 2006 |
-
{
|
| 2007 |
-
"epoch": 0.46,
|
| 2008 |
-
"grad_norm": 28.95624542236328,
|
| 2009 |
-
"learning_rate": 2.5132858131187446e-05,
|
| 2010 |
-
"loss": 3.4688,
|
| 2011 |
-
"step": 5720
|
| 2012 |
-
},
|
| 2013 |
-
{
|
| 2014 |
-
"epoch": 0.46,
|
| 2015 |
-
"grad_norm": 34.07016372680664,
|
| 2016 |
-
"learning_rate": 2.509698418337534e-05,
|
| 2017 |
-
"loss": 3.4938,
|
| 2018 |
-
"step": 5740
|
| 2019 |
-
},
|
| 2020 |
-
{
|
| 2021 |
-
"epoch": 0.46,
|
| 2022 |
-
"grad_norm": 32.694766998291016,
|
| 2023 |
-
"learning_rate": 2.5061004322976953e-05,
|
| 2024 |
-
"loss": 3.4351,
|
| 2025 |
-
"step": 5760
|
| 2026 |
-
},
|
| 2027 |
-
{
|
| 2028 |
-
"epoch": 0.46,
|
| 2029 |
-
"grad_norm": 92.71965789794922,
|
| 2030 |
-
"learning_rate": 2.5024918927404005e-05,
|
| 2031 |
-
"loss": 3.5194,
|
| 2032 |
-
"step": 5780
|
| 2033 |
-
},
|
| 2034 |
-
{
|
| 2035 |
-
"epoch": 0.46,
|
| 2036 |
-
"grad_norm": 59.62504959106445,
|
| 2037 |
-
"learning_rate": 2.4988728375175216e-05,
|
| 2038 |
-
"loss": 3.5436,
|
| 2039 |
-
"step": 5800
|
| 2040 |
-
},
|
| 2041 |
-
{
|
| 2042 |
-
"epoch": 0.47,
|
| 2043 |
-
"grad_norm": 40.504127502441406,
|
| 2044 |
-
"learning_rate": 2.495243304591236e-05,
|
| 2045 |
-
"loss": 3.561,
|
| 2046 |
-
"step": 5820
|
| 2047 |
-
},
|
| 2048 |
-
{
|
| 2049 |
-
"epoch": 0.47,
|
| 2050 |
-
"grad_norm": 43.67660140991211,
|
| 2051 |
-
"learning_rate": 2.4916033320336263e-05,
|
| 2052 |
-
"loss": 3.3979,
|
| 2053 |
-
"step": 5840
|
| 2054 |
-
},
|
| 2055 |
-
{
|
| 2056 |
-
"epoch": 0.47,
|
| 2057 |
-
"grad_norm": 38.79692077636719,
|
| 2058 |
-
"learning_rate": 2.487952958026282e-05,
|
| 2059 |
-
"loss": 3.5581,
|
| 2060 |
-
"step": 5860
|
| 2061 |
-
},
|
| 2062 |
-
{
|
| 2063 |
-
"epoch": 0.47,
|
| 2064 |
-
"grad_norm": 34.00471878051758,
|
| 2065 |
-
"learning_rate": 2.4842922208598996e-05,
|
| 2066 |
-
"loss": 3.3891,
|
| 2067 |
-
"step": 5880
|
| 2068 |
-
},
|
| 2069 |
-
{
|
| 2070 |
-
"epoch": 0.47,
|
| 2071 |
-
"grad_norm": 51.836395263671875,
|
| 2072 |
-
"learning_rate": 2.480621158933879e-05,
|
| 2073 |
-
"loss": 3.5146,
|
| 2074 |
-
"step": 5900
|
| 2075 |
-
},
|
| 2076 |
-
{
|
| 2077 |
-
"epoch": 0.47,
|
| 2078 |
-
"grad_norm": 34.47429656982422,
|
| 2079 |
-
"learning_rate": 2.476939810755923e-05,
|
| 2080 |
-
"loss": 3.4988,
|
| 2081 |
-
"step": 5920
|
| 2082 |
-
},
|
| 2083 |
-
{
|
| 2084 |
-
"epoch": 0.48,
|
| 2085 |
-
"grad_norm": 27.388093948364258,
|
| 2086 |
-
"learning_rate": 2.4732482149416325e-05,
|
| 2087 |
-
"loss": 3.4577,
|
| 2088 |
-
"step": 5940
|
| 2089 |
-
},
|
| 2090 |
-
{
|
| 2091 |
-
"epoch": 0.48,
|
| 2092 |
-
"grad_norm": 22.31077766418457,
|
| 2093 |
-
"learning_rate": 2.4695464102141002e-05,
|
| 2094 |
-
"loss": 3.4816,
|
| 2095 |
-
"step": 5960
|
| 2096 |
-
},
|
| 2097 |
-
{
|
| 2098 |
-
"epoch": 0.48,
|
| 2099 |
-
"grad_norm": 33.66743469238281,
|
| 2100 |
-
"learning_rate": 2.4658344354035063e-05,
|
| 2101 |
-
"loss": 3.5148,
|
| 2102 |
-
"step": 5980
|
| 2103 |
-
},
|
| 2104 |
-
{
|
| 2105 |
-
"epoch": 0.48,
|
| 2106 |
-
"grad_norm": 30.200748443603516,
|
| 2107 |
-
"learning_rate": 2.46211232944671e-05,
|
| 2108 |
-
"loss": 3.6094,
|
| 2109 |
-
"step": 6000
|
| 2110 |
-
},
|
| 2111 |
-
{
|
| 2112 |
-
"epoch": 0.48,
|
| 2113 |
-
"grad_norm": 36.24076843261719,
|
| 2114 |
-
"learning_rate": 2.4583801313868417e-05,
|
| 2115 |
-
"loss": 3.4601,
|
| 2116 |
-
"step": 6020
|
| 2117 |
-
},
|
| 2118 |
-
{
|
| 2119 |
-
"epoch": 0.48,
|
| 2120 |
-
"grad_norm": 36.80402374267578,
|
| 2121 |
-
"learning_rate": 2.4546378803728922e-05,
|
| 2122 |
-
"loss": 3.5053,
|
| 2123 |
-
"step": 6040
|
| 2124 |
-
},
|
| 2125 |
-
{
|
| 2126 |
-
"epoch": 0.48,
|
| 2127 |
-
"grad_norm": 25.655963897705078,
|
| 2128 |
-
"learning_rate": 2.450885615659305e-05,
|
| 2129 |
-
"loss": 3.4791,
|
| 2130 |
-
"step": 6060
|
| 2131 |
-
},
|
| 2132 |
-
{
|
| 2133 |
-
"epoch": 0.49,
|
| 2134 |
-
"grad_norm": 47.66796112060547,
|
| 2135 |
-
"learning_rate": 2.447123376605561e-05,
|
| 2136 |
-
"loss": 3.4535,
|
| 2137 |
-
"step": 6080
|
| 2138 |
-
},
|
| 2139 |
-
{
|
| 2140 |
-
"epoch": 0.49,
|
| 2141 |
-
"grad_norm": 39.040924072265625,
|
| 2142 |
-
"learning_rate": 2.4433512026757668e-05,
|
| 2143 |
-
"loss": 3.5125,
|
| 2144 |
-
"step": 6100
|
| 2145 |
-
},
|
| 2146 |
-
{
|
| 2147 |
-
"epoch": 0.49,
|
| 2148 |
-
"grad_norm": 27.91790008544922,
|
| 2149 |
-
"learning_rate": 2.439569133438243e-05,
|
| 2150 |
-
"loss": 3.5131,
|
| 2151 |
-
"step": 6120
|
| 2152 |
-
},
|
| 2153 |
-
{
|
| 2154 |
-
"epoch": 0.49,
|
| 2155 |
-
"grad_norm": 42.502357482910156,
|
| 2156 |
-
"learning_rate": 2.435777208565106e-05,
|
| 2157 |
-
"loss": 3.5424,
|
| 2158 |
-
"step": 6140
|
| 2159 |
-
},
|
| 2160 |
-
{
|
| 2161 |
-
"epoch": 0.49,
|
| 2162 |
-
"grad_norm": 26.83557891845703,
|
| 2163 |
-
"learning_rate": 2.431975467831853e-05,
|
| 2164 |
-
"loss": 3.4445,
|
| 2165 |
-
"step": 6160
|
| 2166 |
-
},
|
| 2167 |
-
{
|
| 2168 |
-
"epoch": 0.49,
|
| 2169 |
-
"grad_norm": 39.8357048034668,
|
| 2170 |
-
"learning_rate": 2.4281639511169457e-05,
|
| 2171 |
-
"loss": 3.5702,
|
| 2172 |
-
"step": 6180
|
| 2173 |
-
},
|
| 2174 |
-
{
|
| 2175 |
-
"epoch": 0.5,
|
| 2176 |
-
"grad_norm": 31.746124267578125,
|
| 2177 |
-
"learning_rate": 2.424342698401391e-05,
|
| 2178 |
-
"loss": 3.4539,
|
| 2179 |
-
"step": 6200
|
| 2180 |
-
},
|
| 2181 |
-
{
|
| 2182 |
-
"epoch": 0.5,
|
| 2183 |
-
"grad_norm": 53.038482666015625,
|
| 2184 |
-
"learning_rate": 2.4205117497683213e-05,
|
| 2185 |
-
"loss": 3.5491,
|
| 2186 |
-
"step": 6220
|
| 2187 |
-
},
|
| 2188 |
-
{
|
| 2189 |
-
"epoch": 0.5,
|
| 2190 |
-
"grad_norm": 28.155752182006836,
|
| 2191 |
-
"learning_rate": 2.4166711454025754e-05,
|
| 2192 |
-
"loss": 3.4353,
|
| 2193 |
-
"step": 6240
|
| 2194 |
-
},
|
| 2195 |
-
{
|
| 2196 |
-
"epoch": 0.5,
|
| 2197 |
-
"grad_norm": 42.623130798339844,
|
| 2198 |
-
"learning_rate": 2.4128209255902753e-05,
|
| 2199 |
-
"loss": 3.4348,
|
| 2200 |
-
"step": 6260
|
| 2201 |
-
},
|
| 2202 |
-
{
|
| 2203 |
-
"epoch": 0.5,
|
| 2204 |
-
"grad_norm": 44.352596282958984,
|
| 2205 |
-
"learning_rate": 2.408961130718405e-05,
|
| 2206 |
-
"loss": 3.4637,
|
| 2207 |
-
"step": 6280
|
| 2208 |
-
},
|
| 2209 |
-
{
|
| 2210 |
-
"epoch": 0.5,
|
| 2211 |
-
"grad_norm": 35.380767822265625,
|
| 2212 |
-
"learning_rate": 2.405091801274387e-05,
|
| 2213 |
-
"loss": 3.4403,
|
| 2214 |
-
"step": 6300
|
| 2215 |
-
},
|
| 2216 |
-
{
|
| 2217 |
-
"epoch": 0.51,
|
| 2218 |
-
"grad_norm": 67.38970184326172,
|
| 2219 |
-
"learning_rate": 2.4012129778456556e-05,
|
| 2220 |
-
"loss": 3.5062,
|
| 2221 |
-
"step": 6320
|
| 2222 |
-
},
|
| 2223 |
-
{
|
| 2224 |
-
"epoch": 0.51,
|
| 2225 |
-
"grad_norm": 36.066741943359375,
|
| 2226 |
-
"learning_rate": 2.397324701119233e-05,
|
| 2227 |
-
"loss": 3.6017,
|
| 2228 |
-
"step": 6340
|
| 2229 |
-
},
|
| 2230 |
-
{
|
| 2231 |
-
"epoch": 0.51,
|
| 2232 |
-
"grad_norm": 29.4312744140625,
|
| 2233 |
-
"learning_rate": 2.3934270118813024e-05,
|
| 2234 |
-
"loss": 3.4738,
|
| 2235 |
-
"step": 6360
|
| 2236 |
-
},
|
| 2237 |
-
{
|
| 2238 |
-
"epoch": 0.51,
|
| 2239 |
-
"grad_norm": 24.581098556518555,
|
| 2240 |
-
"learning_rate": 2.3895199510167793e-05,
|
| 2241 |
-
"loss": 3.4775,
|
| 2242 |
-
"step": 6380
|
| 2243 |
-
},
|
| 2244 |
-
{
|
| 2245 |
-
"epoch": 0.51,
|
| 2246 |
-
"grad_norm": 22.933914184570312,
|
| 2247 |
-
"learning_rate": 2.385603559508884e-05,
|
| 2248 |
-
"loss": 3.5718,
|
| 2249 |
-
"step": 6400
|
| 2250 |
-
},
|
| 2251 |
-
{
|
| 2252 |
-
"epoch": 0.51,
|
| 2253 |
-
"grad_norm": 30.199399948120117,
|
| 2254 |
-
"learning_rate": 2.3816778784387097e-05,
|
| 2255 |
-
"loss": 3.4446,
|
| 2256 |
-
"step": 6420
|
| 2257 |
-
},
|
| 2258 |
-
{
|
| 2259 |
-
"epoch": 0.52,
|
| 2260 |
-
"grad_norm": 53.968971252441406,
|
| 2261 |
-
"learning_rate": 2.3777429489847935e-05,
|
| 2262 |
-
"loss": 3.5161,
|
| 2263 |
-
"step": 6440
|
| 2264 |
-
},
|
| 2265 |
-
{
|
| 2266 |
-
"epoch": 0.52,
|
| 2267 |
-
"grad_norm": 34.33303451538086,
|
| 2268 |
-
"learning_rate": 2.3737988124226834e-05,
|
| 2269 |
-
"loss": 3.422,
|
| 2270 |
-
"step": 6460
|
| 2271 |
-
},
|
| 2272 |
-
{
|
| 2273 |
-
"epoch": 0.52,
|
| 2274 |
-
"grad_norm": 28.162084579467773,
|
| 2275 |
-
"learning_rate": 2.3698455101245052e-05,
|
| 2276 |
-
"loss": 3.5403,
|
| 2277 |
-
"step": 6480
|
| 2278 |
-
},
|
| 2279 |
-
{
|
| 2280 |
-
"epoch": 0.52,
|
| 2281 |
-
"grad_norm": 28.261327743530273,
|
| 2282 |
-
"learning_rate": 2.3658830835585294e-05,
|
| 2283 |
-
"loss": 3.5093,
|
| 2284 |
-
"step": 6500
|
| 2285 |
-
},
|
| 2286 |
-
{
|
| 2287 |
-
"epoch": 0.52,
|
| 2288 |
-
"grad_norm": 37.4576530456543,
|
| 2289 |
-
"learning_rate": 2.361911574288736e-05,
|
| 2290 |
-
"loss": 3.5952,
|
| 2291 |
-
"step": 6520
|
| 2292 |
-
},
|
| 2293 |
-
{
|
| 2294 |
-
"epoch": 0.52,
|
| 2295 |
-
"grad_norm": 41.03891372680664,
|
| 2296 |
-
"learning_rate": 2.3579310239743776e-05,
|
| 2297 |
-
"loss": 3.5194,
|
| 2298 |
-
"step": 6540
|
| 2299 |
-
},
|
| 2300 |
-
{
|
| 2301 |
-
"epoch": 0.52,
|
| 2302 |
-
"grad_norm": 26.61548614501953,
|
| 2303 |
-
"learning_rate": 2.353941474369544e-05,
|
| 2304 |
-
"loss": 3.4933,
|
| 2305 |
-
"step": 6560
|
| 2306 |
-
},
|
| 2307 |
-
{
|
| 2308 |
-
"epoch": 0.53,
|
| 2309 |
-
"grad_norm": 33.990203857421875,
|
| 2310 |
-
"learning_rate": 2.3499429673227224e-05,
|
| 2311 |
-
"loss": 3.4745,
|
| 2312 |
-
"step": 6580
|
| 2313 |
-
},
|
| 2314 |
-
{
|
| 2315 |
-
"epoch": 0.53,
|
| 2316 |
-
"grad_norm": 39.59190368652344,
|
| 2317 |
-
"learning_rate": 2.3459355447763596e-05,
|
| 2318 |
-
"loss": 3.4875,
|
| 2319 |
-
"step": 6600
|
| 2320 |
-
},
|
| 2321 |
-
{
|
| 2322 |
-
"epoch": 0.53,
|
| 2323 |
-
"grad_norm": 28.182842254638672,
|
| 2324 |
-
"learning_rate": 2.341919248766422e-05,
|
| 2325 |
-
"loss": 3.5085,
|
| 2326 |
-
"step": 6620
|
| 2327 |
-
},
|
| 2328 |
-
{
|
| 2329 |
-
"epoch": 0.53,
|
| 2330 |
-
"grad_norm": 27.481103897094727,
|
| 2331 |
-
"learning_rate": 2.3378941214219545e-05,
|
| 2332 |
-
"loss": 3.5401,
|
| 2333 |
-
"step": 6640
|
| 2334 |
-
},
|
| 2335 |
-
{
|
| 2336 |
-
"epoch": 0.53,
|
| 2337 |
-
"grad_norm": 29.300914764404297,
|
| 2338 |
-
"learning_rate": 2.3338602049646372e-05,
|
| 2339 |
-
"loss": 3.4601,
|
| 2340 |
-
"step": 6660
|
| 2341 |
-
},
|
| 2342 |
-
{
|
| 2343 |
-
"epoch": 0.53,
|
| 2344 |
-
"grad_norm": 27.341121673583984,
|
| 2345 |
-
"learning_rate": 2.329817541708346e-05,
|
| 2346 |
-
"loss": 3.4163,
|
| 2347 |
-
"step": 6680
|
| 2348 |
-
},
|
| 2349 |
-
{
|
| 2350 |
-
"epoch": 0.54,
|
| 2351 |
-
"grad_norm": 21.337862014770508,
|
| 2352 |
-
"learning_rate": 2.3257661740587055e-05,
|
| 2353 |
-
"loss": 3.3443,
|
| 2354 |
-
"step": 6700
|
| 2355 |
-
},
|
| 2356 |
-
{
|
| 2357 |
-
"epoch": 0.54,
|
| 2358 |
-
"grad_norm": 38.06822204589844,
|
| 2359 |
-
"learning_rate": 2.3217061445126444e-05,
|
| 2360 |
-
"loss": 3.4762,
|
| 2361 |
-
"step": 6720
|
| 2362 |
-
},
|
| 2363 |
-
{
|
| 2364 |
-
"epoch": 0.54,
|
| 2365 |
-
"grad_norm": 27.79935646057129,
|
| 2366 |
-
"learning_rate": 2.3176374956579525e-05,
|
| 2367 |
-
"loss": 3.4748,
|
| 2368 |
-
"step": 6740
|
| 2369 |
-
},
|
| 2370 |
-
{
|
| 2371 |
-
"epoch": 0.54,
|
| 2372 |
-
"grad_norm": 41.864219665527344,
|
| 2373 |
-
"learning_rate": 2.3135602701728302e-05,
|
| 2374 |
-
"loss": 3.4859,
|
| 2375 |
-
"step": 6760
|
| 2376 |
-
},
|
| 2377 |
-
{
|
| 2378 |
-
"epoch": 0.54,
|
| 2379 |
-
"grad_norm": 48.69860076904297,
|
| 2380 |
-
"learning_rate": 2.3094745108254437e-05,
|
| 2381 |
-
"loss": 3.5111,
|
| 2382 |
-
"step": 6780
|
| 2383 |
-
},
|
| 2384 |
-
{
|
| 2385 |
-
"epoch": 0.54,
|
| 2386 |
-
"grad_norm": 26.23602867126465,
|
| 2387 |
-
"learning_rate": 2.305380260473476e-05,
|
| 2388 |
-
"loss": 3.4637,
|
| 2389 |
-
"step": 6800
|
| 2390 |
-
},
|
| 2391 |
-
{
|
| 2392 |
-
"epoch": 0.55,
|
| 2393 |
-
"grad_norm": 32.71681213378906,
|
| 2394 |
-
"learning_rate": 2.3012775620636747e-05,
|
| 2395 |
-
"loss": 3.4752,
|
| 2396 |
-
"step": 6820
|
| 2397 |
-
},
|
| 2398 |
-
{
|
| 2399 |
-
"epoch": 0.55,
|
| 2400 |
-
"grad_norm": 55.637813568115234,
|
| 2401 |
-
"learning_rate": 2.2971664586314055e-05,
|
| 2402 |
-
"loss": 3.531,
|
| 2403 |
-
"step": 6840
|
| 2404 |
-
},
|
| 2405 |
-
{
|
| 2406 |
-
"epoch": 0.55,
|
| 2407 |
-
"grad_norm": 29.17162322998047,
|
| 2408 |
-
"learning_rate": 2.293046993300198e-05,
|
| 2409 |
-
"loss": 3.567,
|
| 2410 |
-
"step": 6860
|
| 2411 |
-
},
|
| 2412 |
-
{
|
| 2413 |
-
"epoch": 0.55,
|
| 2414 |
-
"grad_norm": 21.349533081054688,
|
| 2415 |
-
"learning_rate": 2.288919209281294e-05,
|
| 2416 |
-
"loss": 3.4752,
|
| 2417 |
-
"step": 6880
|
| 2418 |
-
},
|
| 2419 |
-
{
|
| 2420 |
-
"epoch": 0.55,
|
| 2421 |
-
"grad_norm": 25.281597137451172,
|
| 2422 |
-
"learning_rate": 2.284783149873195e-05,
|
| 2423 |
-
"loss": 3.4934,
|
| 2424 |
-
"step": 6900
|
| 2425 |
-
},
|
| 2426 |
-
{
|
| 2427 |
-
"epoch": 0.55,
|
| 2428 |
-
"grad_norm": 32.13029861450195,
|
| 2429 |
-
"learning_rate": 2.2806388584612067e-05,
|
| 2430 |
-
"loss": 3.4793,
|
| 2431 |
-
"step": 6920
|
| 2432 |
-
},
|
| 2433 |
-
{
|
| 2434 |
-
"epoch": 0.56,
|
| 2435 |
-
"grad_norm": 21.672082901000977,
|
| 2436 |
-
"learning_rate": 2.2764863785169857e-05,
|
| 2437 |
-
"loss": 3.4366,
|
| 2438 |
-
"step": 6940
|
| 2439 |
-
},
|
| 2440 |
-
{
|
| 2441 |
-
"epoch": 0.56,
|
| 2442 |
-
"grad_norm": 29.802305221557617,
|
| 2443 |
-
"learning_rate": 2.2723257535980804e-05,
|
| 2444 |
-
"loss": 3.5174,
|
| 2445 |
-
"step": 6960
|
| 2446 |
-
},
|
| 2447 |
-
{
|
| 2448 |
-
"epoch": 0.56,
|
| 2449 |
-
"grad_norm": 43.310577392578125,
|
| 2450 |
-
"learning_rate": 2.2681570273474783e-05,
|
| 2451 |
-
"loss": 3.4745,
|
| 2452 |
-
"step": 6980
|
| 2453 |
-
},
|
| 2454 |
-
{
|
| 2455 |
-
"epoch": 0.56,
|
| 2456 |
-
"grad_norm": 32.417236328125,
|
| 2457 |
-
"learning_rate": 2.2639802434931447e-05,
|
| 2458 |
-
"loss": 3.438,
|
| 2459 |
-
"step": 7000
|
| 2460 |
-
},
|
| 2461 |
-
{
|
| 2462 |
-
"epoch": 0.56,
|
| 2463 |
-
"grad_norm": 42.29374313354492,
|
| 2464 |
-
"learning_rate": 2.259795445847566e-05,
|
| 2465 |
-
"loss": 3.5194,
|
| 2466 |
-
"step": 7020
|
| 2467 |
-
},
|
| 2468 |
-
{
|
| 2469 |
-
"epoch": 0.56,
|
| 2470 |
-
"grad_norm": 51.19217300415039,
|
| 2471 |
-
"learning_rate": 2.2556026783072896e-05,
|
| 2472 |
-
"loss": 3.496,
|
| 2473 |
-
"step": 7040
|
| 2474 |
-
},
|
| 2475 |
-
{
|
| 2476 |
-
"epoch": 0.56,
|
| 2477 |
-
"grad_norm": 24.400171279907227,
|
| 2478 |
-
"learning_rate": 2.251401984852463e-05,
|
| 2479 |
-
"loss": 3.4013,
|
| 2480 |
-
"step": 7060
|
| 2481 |
-
},
|
| 2482 |
-
{
|
| 2483 |
-
"epoch": 0.57,
|
| 2484 |
-
"grad_norm": 26.298309326171875,
|
| 2485 |
-
"learning_rate": 2.2471934095463724e-05,
|
| 2486 |
-
"loss": 3.5607,
|
| 2487 |
-
"step": 7080
|
| 2488 |
-
},
|
| 2489 |
-
{
|
| 2490 |
-
"epoch": 0.57,
|
| 2491 |
-
"grad_norm": 27.419946670532227,
|
| 2492 |
-
"learning_rate": 2.2429769965349818e-05,
|
| 2493 |
-
"loss": 3.4593,
|
| 2494 |
-
"step": 7100
|
| 2495 |
-
},
|
| 2496 |
-
{
|
| 2497 |
-
"epoch": 0.57,
|
| 2498 |
-
"grad_norm": 29.470266342163086,
|
| 2499 |
-
"learning_rate": 2.2387527900464676e-05,
|
| 2500 |
-
"loss": 3.4388,
|
| 2501 |
-
"step": 7120
|
| 2502 |
-
},
|
| 2503 |
-
{
|
| 2504 |
-
"epoch": 0.57,
|
| 2505 |
-
"grad_norm": 21.410829544067383,
|
| 2506 |
-
"learning_rate": 2.2345208343907577e-05,
|
| 2507 |
-
"loss": 3.5141,
|
| 2508 |
-
"step": 7140
|
| 2509 |
-
},
|
| 2510 |
-
{
|
| 2511 |
-
"epoch": 0.57,
|
| 2512 |
-
"grad_norm": 32.501766204833984,
|
| 2513 |
-
"learning_rate": 2.2302811739590642e-05,
|
| 2514 |
-
"loss": 3.4647,
|
| 2515 |
-
"step": 7160
|
| 2516 |
-
},
|
| 2517 |
-
{
|
| 2518 |
-
"epoch": 0.57,
|
| 2519 |
-
"grad_norm": 24.69274139404297,
|
| 2520 |
-
"learning_rate": 2.2260338532234194e-05,
|
| 2521 |
-
"loss": 3.4781,
|
| 2522 |
-
"step": 7180
|
| 2523 |
-
},
|
| 2524 |
-
{
|
| 2525 |
-
"epoch": 0.58,
|
| 2526 |
-
"grad_norm": 25.721759796142578,
|
| 2527 |
-
"learning_rate": 2.2217789167362078e-05,
|
| 2528 |
-
"loss": 3.4405,
|
| 2529 |
-
"step": 7200
|
| 2530 |
-
},
|
| 2531 |
-
{
|
| 2532 |
-
"epoch": 0.58,
|
| 2533 |
-
"grad_norm": 34.66562271118164,
|
| 2534 |
-
"learning_rate": 2.217516409129699e-05,
|
| 2535 |
-
"loss": 3.5408,
|
| 2536 |
-
"step": 7220
|
| 2537 |
-
},
|
| 2538 |
-
{
|
| 2539 |
-
"epoch": 0.58,
|
| 2540 |
-
"grad_norm": 42.050819396972656,
|
| 2541 |
-
"learning_rate": 2.2132463751155815e-05,
|
| 2542 |
-
"loss": 3.4422,
|
| 2543 |
-
"step": 7240
|
| 2544 |
-
},
|
| 2545 |
-
{
|
| 2546 |
-
"epoch": 0.58,
|
| 2547 |
-
"grad_norm": 26.5496883392334,
|
| 2548 |
-
"learning_rate": 2.2089688594844917e-05,
|
| 2549 |
-
"loss": 3.4953,
|
| 2550 |
-
"step": 7260
|
| 2551 |
-
},
|
| 2552 |
-
{
|
| 2553 |
-
"epoch": 0.58,
|
| 2554 |
-
"grad_norm": 28.221229553222656,
|
| 2555 |
-
"learning_rate": 2.2046839071055436e-05,
|
| 2556 |
-
"loss": 3.3627,
|
| 2557 |
-
"step": 7280
|
| 2558 |
-
},
|
| 2559 |
-
{
|
| 2560 |
-
"epoch": 0.58,
|
| 2561 |
-
"grad_norm": 33.21296310424805,
|
| 2562 |
-
"learning_rate": 2.2003915629258607e-05,
|
| 2563 |
-
"loss": 3.383,
|
| 2564 |
-
"step": 7300
|
| 2565 |
-
},
|
| 2566 |
-
{
|
| 2567 |
-
"epoch": 0.59,
|
| 2568 |
-
"grad_norm": 34.35666275024414,
|
| 2569 |
-
"learning_rate": 2.196091871970103e-05,
|
| 2570 |
-
"loss": 3.5508,
|
| 2571 |
-
"step": 7320
|
| 2572 |
-
},
|
| 2573 |
-
{
|
| 2574 |
-
"epoch": 0.59,
|
| 2575 |
-
"grad_norm": 25.02892303466797,
|
| 2576 |
-
"learning_rate": 2.1917848793399926e-05,
|
| 2577 |
-
"loss": 3.395,
|
| 2578 |
-
"step": 7340
|
| 2579 |
-
},
|
| 2580 |
-
{
|
| 2581 |
-
"epoch": 0.59,
|
| 2582 |
-
"grad_norm": 25.45415687561035,
|
| 2583 |
-
"learning_rate": 2.187470630213845e-05,
|
| 2584 |
-
"loss": 3.4692,
|
| 2585 |
-
"step": 7360
|
| 2586 |
-
},
|
| 2587 |
-
{
|
| 2588 |
-
"epoch": 0.59,
|
| 2589 |
-
"grad_norm": 30.72098159790039,
|
| 2590 |
-
"learning_rate": 2.1831491698460923e-05,
|
| 2591 |
-
"loss": 3.5714,
|
| 2592 |
-
"step": 7380
|
| 2593 |
-
},
|
| 2594 |
-
{
|
| 2595 |
-
"epoch": 0.59,
|
| 2596 |
-
"grad_norm": 32.58151626586914,
|
| 2597 |
-
"learning_rate": 2.1788205435668086e-05,
|
| 2598 |
-
"loss": 3.4358,
|
| 2599 |
-
"step": 7400
|
| 2600 |
-
},
|
| 2601 |
-
{
|
| 2602 |
-
"epoch": 0.59,
|
| 2603 |
-
"grad_norm": 32.444644927978516,
|
| 2604 |
-
"learning_rate": 2.1744847967812352e-05,
|
| 2605 |
-
"loss": 3.5389,
|
| 2606 |
-
"step": 7420
|
| 2607 |
-
},
|
| 2608 |
-
{
|
| 2609 |
-
"epoch": 0.6,
|
| 2610 |
-
"grad_norm": 39.33094787597656,
|
| 2611 |
-
"learning_rate": 2.1701419749693036e-05,
|
| 2612 |
-
"loss": 3.4003,
|
| 2613 |
-
"step": 7440
|
| 2614 |
-
},
|
| 2615 |
-
{
|
| 2616 |
-
"epoch": 0.6,
|
| 2617 |
-
"grad_norm": 24.18759536743164,
|
| 2618 |
-
"learning_rate": 2.1657921236851607e-05,
|
| 2619 |
-
"loss": 3.4501,
|
| 2620 |
-
"step": 7460
|
| 2621 |
-
},
|
| 2622 |
-
{
|
| 2623 |
-
"epoch": 0.6,
|
| 2624 |
-
"grad_norm": 24.81824493408203,
|
| 2625 |
-
"learning_rate": 2.1614352885566874e-05,
|
| 2626 |
-
"loss": 3.5004,
|
| 2627 |
-
"step": 7480
|
| 2628 |
-
},
|
| 2629 |
-
{
|
| 2630 |
-
"epoch": 0.6,
|
| 2631 |
-
"grad_norm": 94.60846710205078,
|
| 2632 |
-
"learning_rate": 2.1570715152850237e-05,
|
| 2633 |
-
"loss": 3.3695,
|
| 2634 |
-
"step": 7500
|
| 2635 |
-
},
|
| 2636 |
-
{
|
| 2637 |
-
"epoch": 0.6,
|
| 2638 |
-
"grad_norm": 39.35719299316406,
|
| 2639 |
-
"learning_rate": 2.1527008496440848e-05,
|
| 2640 |
-
"loss": 3.4058,
|
| 2641 |
-
"step": 7520
|
| 2642 |
-
},
|
| 2643 |
-
{
|
| 2644 |
-
"epoch": 0.6,
|
| 2645 |
-
"grad_norm": 23.86483383178711,
|
| 2646 |
-
"learning_rate": 2.1483233374800863e-05,
|
| 2647 |
-
"loss": 3.403,
|
| 2648 |
-
"step": 7540
|
| 2649 |
-
},
|
| 2650 |
-
{
|
| 2651 |
-
"epoch": 0.6,
|
| 2652 |
-
"grad_norm": 36.54322052001953,
|
| 2653 |
-
"learning_rate": 2.143939024711059e-05,
|
| 2654 |
-
"loss": 3.504,
|
| 2655 |
-
"step": 7560
|
| 2656 |
-
},
|
| 2657 |
-
{
|
| 2658 |
-
"epoch": 0.61,
|
| 2659 |
-
"grad_norm": 32.654876708984375,
|
| 2660 |
-
"learning_rate": 2.139547957326369e-05,
|
| 2661 |
-
"loss": 3.421,
|
| 2662 |
-
"step": 7580
|
| 2663 |
-
},
|
| 2664 |
-
{
|
| 2665 |
-
"epoch": 0.61,
|
| 2666 |
-
"grad_norm": 31.50335121154785,
|
| 2667 |
-
"learning_rate": 2.1351501813862358e-05,
|
| 2668 |
-
"loss": 3.5079,
|
| 2669 |
-
"step": 7600
|
| 2670 |
-
},
|
| 2671 |
-
{
|
| 2672 |
-
"epoch": 0.61,
|
| 2673 |
-
"grad_norm": 27.86897850036621,
|
| 2674 |
-
"learning_rate": 2.130745743021247e-05,
|
| 2675 |
-
"loss": 3.5605,
|
| 2676 |
-
"step": 7620
|
| 2677 |
-
},
|
| 2678 |
-
{
|
| 2679 |
-
"epoch": 0.61,
|
| 2680 |
-
"grad_norm": 29.914051055908203,
|
| 2681 |
-
"learning_rate": 2.1263346884318778e-05,
|
| 2682 |
-
"loss": 3.4826,
|
| 2683 |
-
"step": 7640
|
| 2684 |
-
},
|
| 2685 |
-
{
|
| 2686 |
-
"epoch": 0.61,
|
| 2687 |
-
"grad_norm": 30.09800910949707,
|
| 2688 |
-
"learning_rate": 2.121917063888004e-05,
|
| 2689 |
-
"loss": 3.5523,
|
| 2690 |
-
"step": 7660
|
| 2691 |
-
},
|
| 2692 |
-
{
|
| 2693 |
-
"epoch": 0.61,
|
| 2694 |
-
"grad_norm": 37.76126480102539,
|
| 2695 |
-
"learning_rate": 2.117492915728416e-05,
|
| 2696 |
-
"loss": 3.6012,
|
| 2697 |
-
"step": 7680
|
| 2698 |
-
},
|
| 2699 |
-
{
|
| 2700 |
-
"epoch": 0.62,
|
| 2701 |
-
"grad_norm": 23.360584259033203,
|
| 2702 |
-
"learning_rate": 2.1130622903603344e-05,
|
| 2703 |
-
"loss": 3.45,
|
| 2704 |
-
"step": 7700
|
| 2705 |
-
},
|
| 2706 |
-
{
|
| 2707 |
-
"epoch": 0.62,
|
| 2708 |
-
"grad_norm": 31.356124877929688,
|
| 2709 |
-
"learning_rate": 2.1086252342589235e-05,
|
| 2710 |
-
"loss": 3.4666,
|
| 2711 |
-
"step": 7720
|
| 2712 |
-
},
|
| 2713 |
-
{
|
| 2714 |
-
"epoch": 0.62,
|
| 2715 |
-
"grad_norm": 36.73623275756836,
|
| 2716 |
-
"learning_rate": 2.1041817939668006e-05,
|
| 2717 |
-
"loss": 3.5774,
|
| 2718 |
-
"step": 7740
|
| 2719 |
-
},
|
| 2720 |
-
{
|
| 2721 |
-
"epoch": 0.62,
|
| 2722 |
-
"grad_norm": 29.439634323120117,
|
| 2723 |
-
"learning_rate": 2.0997320160935536e-05,
|
| 2724 |
-
"loss": 3.3953,
|
| 2725 |
-
"step": 7760
|
| 2726 |
-
},
|
| 2727 |
-
{
|
| 2728 |
-
"epoch": 0.62,
|
| 2729 |
-
"grad_norm": 22.339067459106445,
|
| 2730 |
-
"learning_rate": 2.095275947315246e-05,
|
| 2731 |
-
"loss": 3.4628,
|
| 2732 |
-
"step": 7780
|
| 2733 |
-
},
|
| 2734 |
-
{
|
| 2735 |
-
"epoch": 0.62,
|
| 2736 |
-
"grad_norm": 16.739225387573242,
|
| 2737 |
-
"learning_rate": 2.0908136343739308e-05,
|
| 2738 |
-
"loss": 3.4708,
|
| 2739 |
-
"step": 7800
|
| 2740 |
-
},
|
| 2741 |
-
{
|
| 2742 |
-
"epoch": 0.63,
|
| 2743 |
-
"grad_norm": 27.203229904174805,
|
| 2744 |
-
"learning_rate": 2.0863451240771592e-05,
|
| 2745 |
-
"loss": 3.4326,
|
| 2746 |
-
"step": 7820
|
| 2747 |
-
},
|
| 2748 |
-
{
|
| 2749 |
-
"epoch": 0.63,
|
| 2750 |
-
"grad_norm": 30.963396072387695,
|
| 2751 |
-
"learning_rate": 2.0818704632974896e-05,
|
| 2752 |
-
"loss": 3.445,
|
| 2753 |
-
"step": 7840
|
| 2754 |
-
},
|
| 2755 |
-
{
|
| 2756 |
-
"epoch": 0.63,
|
| 2757 |
-
"grad_norm": 23.80015754699707,
|
| 2758 |
-
"learning_rate": 2.0773896989719967e-05,
|
| 2759 |
-
"loss": 3.5519,
|
| 2760 |
-
"step": 7860
|
| 2761 |
-
},
|
| 2762 |
-
{
|
| 2763 |
-
"epoch": 0.63,
|
| 2764 |
-
"grad_norm": 39.84937286376953,
|
| 2765 |
-
"learning_rate": 2.072902878101778e-05,
|
| 2766 |
-
"loss": 3.4384,
|
| 2767 |
-
"step": 7880
|
| 2768 |
-
},
|
| 2769 |
-
{
|
| 2770 |
-
"epoch": 0.63,
|
| 2771 |
-
"grad_norm": 25.7357120513916,
|
| 2772 |
-
"learning_rate": 2.06841004775146e-05,
|
| 2773 |
-
"loss": 3.4322,
|
| 2774 |
-
"step": 7900
|
| 2775 |
-
},
|
| 2776 |
-
{
|
| 2777 |
-
"epoch": 0.63,
|
| 2778 |
-
"grad_norm": 25.275423049926758,
|
| 2779 |
-
"learning_rate": 2.0639112550487085e-05,
|
| 2780 |
-
"loss": 3.4426,
|
| 2781 |
-
"step": 7920
|
| 2782 |
-
},
|
| 2783 |
-
{
|
| 2784 |
-
"epoch": 0.64,
|
| 2785 |
-
"grad_norm": 28.167940139770508,
|
| 2786 |
-
"learning_rate": 2.0594065471837302e-05,
|
| 2787 |
-
"loss": 3.5217,
|
| 2788 |
-
"step": 7940
|
| 2789 |
-
},
|
| 2790 |
-
{
|
| 2791 |
-
"epoch": 0.64,
|
| 2792 |
-
"grad_norm": 44.70423889160156,
|
| 2793 |
-
"learning_rate": 2.0548959714087783e-05,
|
| 2794 |
-
"loss": 3.5207,
|
| 2795 |
-
"step": 7960
|
| 2796 |
-
},
|
| 2797 |
-
{
|
| 2798 |
-
"epoch": 0.64,
|
| 2799 |
-
"grad_norm": 30.5069580078125,
|
| 2800 |
-
"learning_rate": 2.0503795750376583e-05,
|
| 2801 |
-
"loss": 3.4576,
|
| 2802 |
-
"step": 7980
|
| 2803 |
-
},
|
| 2804 |
-
{
|
| 2805 |
-
"epoch": 0.64,
|
| 2806 |
-
"grad_norm": 24.197166442871094,
|
| 2807 |
-
"learning_rate": 2.0458574054452316e-05,
|
| 2808 |
-
"loss": 3.4993,
|
| 2809 |
-
"step": 8000
|
| 2810 |
-
},
|
| 2811 |
-
{
|
| 2812 |
-
"epoch": 0.64,
|
| 2813 |
-
"grad_norm": 31.20002555847168,
|
| 2814 |
-
"learning_rate": 2.0413295100669167e-05,
|
| 2815 |
-
"loss": 3.4158,
|
| 2816 |
-
"step": 8020
|
| 2817 |
-
},
|
| 2818 |
-
{
|
| 2819 |
-
"epoch": 0.64,
|
| 2820 |
-
"grad_norm": 33.78310012817383,
|
| 2821 |
-
"learning_rate": 2.036795936398194e-05,
|
| 2822 |
-
"loss": 3.4395,
|
| 2823 |
-
"step": 8040
|
| 2824 |
-
},
|
| 2825 |
-
{
|
| 2826 |
-
"epoch": 0.64,
|
| 2827 |
-
"grad_norm": 28.749357223510742,
|
| 2828 |
-
"learning_rate": 2.0322567319941062e-05,
|
| 2829 |
-
"loss": 3.451,
|
| 2830 |
-
"step": 8060
|
| 2831 |
-
},
|
| 2832 |
-
{
|
| 2833 |
-
"epoch": 0.65,
|
| 2834 |
-
"grad_norm": 19.698383331298828,
|
| 2835 |
-
"learning_rate": 2.0277119444687586e-05,
|
| 2836 |
-
"loss": 3.5069,
|
| 2837 |
-
"step": 8080
|
| 2838 |
-
},
|
| 2839 |
-
{
|
| 2840 |
-
"epoch": 0.65,
|
| 2841 |
-
"grad_norm": 25.00635528564453,
|
| 2842 |
-
"learning_rate": 2.0231616214948232e-05,
|
| 2843 |
-
"loss": 3.3704,
|
| 2844 |
-
"step": 8100
|
| 2845 |
-
},
|
| 2846 |
-
{
|
| 2847 |
-
"epoch": 0.65,
|
| 2848 |
-
"grad_norm": 31.135547637939453,
|
| 2849 |
-
"learning_rate": 2.0186058108030343e-05,
|
| 2850 |
-
"loss": 3.4219,
|
| 2851 |
-
"step": 8120
|
| 2852 |
-
},
|
| 2853 |
-
{
|
| 2854 |
-
"epoch": 0.65,
|
| 2855 |
-
"grad_norm": 29.921775817871094,
|
| 2856 |
-
"learning_rate": 2.01404456018169e-05,
|
| 2857 |
-
"loss": 3.431,
|
| 2858 |
-
"step": 8140
|
| 2859 |
-
},
|
| 2860 |
-
{
|
| 2861 |
-
"epoch": 0.65,
|
| 2862 |
-
"grad_norm": 20.908395767211914,
|
| 2863 |
-
"learning_rate": 2.0094779174761507e-05,
|
| 2864 |
-
"loss": 3.4647,
|
| 2865 |
-
"step": 8160
|
| 2866 |
-
},
|
| 2867 |
-
{
|
| 2868 |
-
"epoch": 0.65,
|
| 2869 |
-
"grad_norm": 17.694931030273438,
|
| 2870 |
-
"learning_rate": 2.0049059305883383e-05,
|
| 2871 |
-
"loss": 3.4588,
|
| 2872 |
-
"step": 8180
|
| 2873 |
-
},
|
| 2874 |
-
{
|
| 2875 |
-
"epoch": 0.66,
|
| 2876 |
-
"grad_norm": 29.230735778808594,
|
| 2877 |
-
"learning_rate": 2.000328647476231e-05,
|
| 2878 |
-
"loss": 3.4788,
|
| 2879 |
-
"step": 8200
|
| 2880 |
-
},
|
| 2881 |
-
{
|
| 2882 |
-
"epoch": 0.66,
|
| 2883 |
-
"grad_norm": 41.20121383666992,
|
| 2884 |
-
"learning_rate": 1.995746116153363e-05,
|
| 2885 |
-
"loss": 3.4812,
|
| 2886 |
-
"step": 8220
|
| 2887 |
-
},
|
| 2888 |
-
{
|
| 2889 |
-
"epoch": 0.66,
|
| 2890 |
-
"grad_norm": 39.98432540893555,
|
| 2891 |
-
"learning_rate": 1.99115838468832e-05,
|
| 2892 |
-
"loss": 3.4172,
|
| 2893 |
-
"step": 8240
|
| 2894 |
-
},
|
| 2895 |
-
{
|
| 2896 |
-
"epoch": 0.66,
|
| 2897 |
-
"grad_norm": 41.378902435302734,
|
| 2898 |
-
"learning_rate": 1.9865655012042337e-05,
|
| 2899 |
-
"loss": 3.5478,
|
| 2900 |
-
"step": 8260
|
| 2901 |
-
},
|
| 2902 |
-
{
|
| 2903 |
-
"epoch": 0.66,
|
| 2904 |
-
"grad_norm": 42.58726501464844,
|
| 2905 |
-
"learning_rate": 1.9819675138782785e-05,
|
| 2906 |
-
"loss": 3.4849,
|
| 2907 |
-
"step": 8280
|
| 2908 |
-
},
|
| 2909 |
-
{
|
| 2910 |
-
"epoch": 0.66,
|
| 2911 |
-
"grad_norm": 22.349281311035156,
|
| 2912 |
-
"learning_rate": 1.9773644709411662e-05,
|
| 2913 |
-
"loss": 3.4835,
|
| 2914 |
-
"step": 8300
|
| 2915 |
-
},
|
| 2916 |
-
{
|
| 2917 |
-
"epoch": 0.67,
|
| 2918 |
-
"grad_norm": 31.974403381347656,
|
| 2919 |
-
"learning_rate": 1.9727564206766382e-05,
|
| 2920 |
-
"loss": 3.3919,
|
| 2921 |
-
"step": 8320
|
| 2922 |
-
},
|
| 2923 |
-
{
|
| 2924 |
-
"epoch": 0.67,
|
| 2925 |
-
"grad_norm": 20.418291091918945,
|
| 2926 |
-
"learning_rate": 1.9681434114209617e-05,
|
| 2927 |
-
"loss": 3.4415,
|
| 2928 |
-
"step": 8340
|
| 2929 |
-
},
|
| 2930 |
-
{
|
| 2931 |
-
"epoch": 0.67,
|
| 2932 |
-
"grad_norm": 30.5143985748291,
|
| 2933 |
-
"learning_rate": 1.963525491562421e-05,
|
| 2934 |
-
"loss": 3.4806,
|
| 2935 |
-
"step": 8360
|
| 2936 |
-
},
|
| 2937 |
-
{
|
| 2938 |
-
"epoch": 0.67,
|
| 2939 |
-
"grad_norm": 27.770360946655273,
|
| 2940 |
-
"learning_rate": 1.958902709540811e-05,
|
| 2941 |
-
"loss": 3.4379,
|
| 2942 |
-
"step": 8380
|
| 2943 |
-
},
|
| 2944 |
-
{
|
| 2945 |
-
"epoch": 0.67,
|
| 2946 |
-
"grad_norm": 24.266944885253906,
|
| 2947 |
-
"learning_rate": 1.954275113846926e-05,
|
| 2948 |
-
"loss": 3.4933,
|
| 2949 |
-
"step": 8400
|
| 2950 |
-
},
|
| 2951 |
-
{
|
| 2952 |
-
"epoch": 0.67,
|
| 2953 |
-
"grad_norm": 43.798301696777344,
|
| 2954 |
-
"learning_rate": 1.9496427530220567e-05,
|
| 2955 |
-
"loss": 3.4107,
|
| 2956 |
-
"step": 8420
|
| 2957 |
-
},
|
| 2958 |
-
{
|
| 2959 |
-
"epoch": 0.68,
|
| 2960 |
-
"grad_norm": 32.54145431518555,
|
| 2961 |
-
"learning_rate": 1.9450056756574753e-05,
|
| 2962 |
-
"loss": 3.507,
|
| 2963 |
-
"step": 8440
|
| 2964 |
-
},
|
| 2965 |
-
{
|
| 2966 |
-
"epoch": 0.68,
|
| 2967 |
-
"grad_norm": 29.06185531616211,
|
| 2968 |
-
"learning_rate": 1.9403639303939293e-05,
|
| 2969 |
-
"loss": 3.4434,
|
| 2970 |
-
"step": 8460
|
| 2971 |
-
},
|
| 2972 |
-
{
|
| 2973 |
-
"epoch": 0.68,
|
| 2974 |
-
"grad_norm": 26.419170379638672,
|
| 2975 |
-
"learning_rate": 1.93571756592113e-05,
|
| 2976 |
-
"loss": 3.4684,
|
| 2977 |
-
"step": 8480
|
| 2978 |
-
},
|
| 2979 |
-
{
|
| 2980 |
-
"epoch": 0.68,
|
| 2981 |
-
"grad_norm": 31.831510543823242,
|
| 2982 |
-
"learning_rate": 1.9310666309772426e-05,
|
| 2983 |
-
"loss": 3.4565,
|
| 2984 |
-
"step": 8500
|
| 2985 |
-
},
|
| 2986 |
-
{
|
| 2987 |
-
"epoch": 0.68,
|
| 2988 |
-
"grad_norm": 33.18935775756836,
|
| 2989 |
-
"learning_rate": 1.926411174348373e-05,
|
| 2990 |
-
"loss": 3.4568,
|
| 2991 |
-
"step": 8520
|
| 2992 |
-
},
|
| 2993 |
-
{
|
| 2994 |
-
"epoch": 0.68,
|
| 2995 |
-
"grad_norm": 25.60289192199707,
|
| 2996 |
-
"learning_rate": 1.9217512448680586e-05,
|
| 2997 |
-
"loss": 3.3513,
|
| 2998 |
-
"step": 8540
|
| 2999 |
-
},
|
| 3000 |
-
{
|
| 3001 |
-
"epoch": 0.68,
|
| 3002 |
-
"grad_norm": 27.03973960876465,
|
| 3003 |
-
"learning_rate": 1.917086891416755e-05,
|
| 3004 |
-
"loss": 3.4551,
|
| 3005 |
-
"step": 8560
|
| 3006 |
-
},
|
| 3007 |
-
{
|
| 3008 |
-
"epoch": 0.69,
|
| 3009 |
-
"grad_norm": 39.20319366455078,
|
| 3010 |
-
"learning_rate": 1.9124181629213228e-05,
|
| 3011 |
-
"loss": 3.4217,
|
| 3012 |
-
"step": 8580
|
| 3013 |
-
},
|
| 3014 |
-
{
|
| 3015 |
-
"epoch": 0.69,
|
| 3016 |
-
"grad_norm": 25.287826538085938,
|
| 3017 |
-
"learning_rate": 1.9077451083545144e-05,
|
| 3018 |
-
"loss": 3.4172,
|
| 3019 |
-
"step": 8600
|
| 3020 |
-
},
|
| 3021 |
-
{
|
| 3022 |
-
"epoch": 0.69,
|
| 3023 |
-
"grad_norm": 51.33893585205078,
|
| 3024 |
-
"learning_rate": 1.903067776734461e-05,
|
| 3025 |
-
"loss": 3.4487,
|
| 3026 |
-
"step": 8620
|
| 3027 |
-
},
|
| 3028 |
-
{
|
| 3029 |
-
"epoch": 0.69,
|
| 3030 |
-
"grad_norm": 23.41849708557129,
|
| 3031 |
-
"learning_rate": 1.8983862171241577e-05,
|
| 3032 |
-
"loss": 3.4675,
|
| 3033 |
-
"step": 8640
|
| 3034 |
-
},
|
| 3035 |
-
{
|
| 3036 |
-
"epoch": 0.69,
|
| 3037 |
-
"grad_norm": 37.373104095458984,
|
| 3038 |
-
"learning_rate": 1.8937004786309504e-05,
|
| 3039 |
-
"loss": 3.4782,
|
| 3040 |
-
"step": 8660
|
| 3041 |
-
},
|
| 3042 |
-
{
|
| 3043 |
-
"epoch": 0.69,
|
| 3044 |
-
"grad_norm": 26.743480682373047,
|
| 3045 |
-
"learning_rate": 1.8890106104060177e-05,
|
| 3046 |
-
"loss": 3.432,
|
| 3047 |
-
"step": 8680
|
| 3048 |
-
},
|
| 3049 |
-
{
|
| 3050 |
-
"epoch": 0.7,
|
| 3051 |
-
"grad_norm": 39.051124572753906,
|
| 3052 |
-
"learning_rate": 1.8843166616438585e-05,
|
| 3053 |
-
"loss": 3.4937,
|
| 3054 |
-
"step": 8700
|
| 3055 |
-
},
|
| 3056 |
-
{
|
| 3057 |
-
"epoch": 0.7,
|
| 3058 |
-
"grad_norm": 32.51453399658203,
|
| 3059 |
-
"learning_rate": 1.8796186815817743e-05,
|
| 3060 |
-
"loss": 3.4618,
|
| 3061 |
-
"step": 8720
|
| 3062 |
-
},
|
| 3063 |
-
{
|
| 3064 |
-
"epoch": 0.7,
|
| 3065 |
-
"grad_norm": 22.23604393005371,
|
| 3066 |
-
"learning_rate": 1.874916719499353e-05,
|
| 3067 |
-
"loss": 3.5259,
|
| 3068 |
-
"step": 8740
|
| 3069 |
-
},
|
| 3070 |
-
{
|
| 3071 |
-
"epoch": 0.7,
|
| 3072 |
-
"grad_norm": 27.168733596801758,
|
| 3073 |
-
"learning_rate": 1.8702108247179512e-05,
|
| 3074 |
-
"loss": 3.4829,
|
| 3075 |
-
"step": 8760
|
| 3076 |
-
},
|
| 3077 |
-
{
|
| 3078 |
-
"epoch": 0.7,
|
| 3079 |
-
"grad_norm": 28.35675621032715,
|
| 3080 |
-
"learning_rate": 1.8655010466001794e-05,
|
| 3081 |
-
"loss": 3.4133,
|
| 3082 |
-
"step": 8780
|
| 3083 |
-
},
|
| 3084 |
-
{
|
| 3085 |
-
"epoch": 0.7,
|
| 3086 |
-
"grad_norm": 30.800825119018555,
|
| 3087 |
-
"learning_rate": 1.8607874345493806e-05,
|
| 3088 |
-
"loss": 3.4527,
|
| 3089 |
-
"step": 8800
|
| 3090 |
-
},
|
| 3091 |
-
{
|
| 3092 |
-
"epoch": 0.71,
|
| 3093 |
-
"grad_norm": 26.275304794311523,
|
| 3094 |
-
"learning_rate": 1.856070038009115e-05,
|
| 3095 |
-
"loss": 3.4386,
|
| 3096 |
-
"step": 8820
|
| 3097 |
-
},
|
| 3098 |
-
{
|
| 3099 |
-
"epoch": 0.71,
|
| 3100 |
-
"grad_norm": 27.828401565551758,
|
| 3101 |
-
"learning_rate": 1.85134890646264e-05,
|
| 3102 |
-
"loss": 3.4806,
|
| 3103 |
-
"step": 8840
|
| 3104 |
-
},
|
| 3105 |
-
{
|
| 3106 |
-
"epoch": 0.71,
|
| 3107 |
-
"grad_norm": 31.448450088500977,
|
| 3108 |
-
"learning_rate": 1.846624089432392e-05,
|
| 3109 |
-
"loss": 3.4659,
|
| 3110 |
-
"step": 8860
|
| 3111 |
-
},
|
| 3112 |
-
{
|
| 3113 |
-
"epoch": 0.71,
|
| 3114 |
-
"grad_norm": 30.396820068359375,
|
| 3115 |
-
"learning_rate": 1.8418956364794655e-05,
|
| 3116 |
-
"loss": 3.4717,
|
| 3117 |
-
"step": 8880
|
| 3118 |
-
},
|
| 3119 |
-
{
|
| 3120 |
-
"epoch": 0.71,
|
| 3121 |
-
"grad_norm": 65.90203857421875,
|
| 3122 |
-
"learning_rate": 1.8371635972030942e-05,
|
| 3123 |
-
"loss": 3.4938,
|
| 3124 |
-
"step": 8900
|
| 3125 |
-
},
|
| 3126 |
-
{
|
| 3127 |
-
"epoch": 0.71,
|
| 3128 |
-
"grad_norm": 25.6701602935791,
|
| 3129 |
-
"learning_rate": 1.8324280212401316e-05,
|
| 3130 |
-
"loss": 3.3931,
|
| 3131 |
-
"step": 8920
|
| 3132 |
-
},
|
| 3133 |
-
{
|
| 3134 |
-
"epoch": 0.72,
|
| 3135 |
-
"grad_norm": 27.28550148010254,
|
| 3136 |
-
"learning_rate": 1.8276889582645278e-05,
|
| 3137 |
-
"loss": 3.426,
|
| 3138 |
-
"step": 8940
|
| 3139 |
-
},
|
| 3140 |
-
{
|
| 3141 |
-
"epoch": 0.72,
|
| 3142 |
-
"grad_norm": 26.73450469970703,
|
| 3143 |
-
"learning_rate": 1.8229464579868124e-05,
|
| 3144 |
-
"loss": 3.4074,
|
| 3145 |
-
"step": 8960
|
| 3146 |
-
},
|
| 3147 |
-
{
|
| 3148 |
-
"epoch": 0.72,
|
| 3149 |
-
"grad_norm": 30.16777992248535,
|
| 3150 |
-
"learning_rate": 1.818200570153568e-05,
|
| 3151 |
-
"loss": 3.4204,
|
| 3152 |
-
"step": 8980
|
| 3153 |
-
},
|
| 3154 |
-
{
|
| 3155 |
-
"epoch": 0.72,
|
| 3156 |
-
"grad_norm": 30.022031784057617,
|
| 3157 |
-
"learning_rate": 1.813451344546913e-05,
|
| 3158 |
-
"loss": 3.4905,
|
| 3159 |
-
"step": 9000
|
| 3160 |
-
},
|
| 3161 |
-
{
|
| 3162 |
-
"epoch": 0.72,
|
| 3163 |
-
"grad_norm": 34.67860412597656,
|
| 3164 |
-
"learning_rate": 1.8086988309839755e-05,
|
| 3165 |
-
"loss": 3.4165,
|
| 3166 |
-
"step": 9020
|
| 3167 |
-
},
|
| 3168 |
-
{
|
| 3169 |
-
"epoch": 0.72,
|
| 3170 |
-
"grad_norm": 26.23653793334961,
|
| 3171 |
-
"learning_rate": 1.8039430793163753e-05,
|
| 3172 |
-
"loss": 3.5014,
|
| 3173 |
-
"step": 9040
|
| 3174 |
-
},
|
| 3175 |
-
{
|
| 3176 |
-
"epoch": 0.72,
|
| 3177 |
-
"grad_norm": 21.284992218017578,
|
| 3178 |
-
"learning_rate": 1.7991841394296962e-05,
|
| 3179 |
-
"loss": 3.4575,
|
| 3180 |
-
"step": 9060
|
| 3181 |
-
},
|
| 3182 |
-
{
|
| 3183 |
-
"epoch": 0.73,
|
| 3184 |
-
"grad_norm": 38.38702392578125,
|
| 3185 |
-
"learning_rate": 1.7944220612429664e-05,
|
| 3186 |
-
"loss": 3.4258,
|
| 3187 |
-
"step": 9080
|
| 3188 |
-
},
|
| 3189 |
-
{
|
| 3190 |
-
"epoch": 0.73,
|
| 3191 |
-
"grad_norm": 23.724790573120117,
|
| 3192 |
-
"learning_rate": 1.789656894708132e-05,
|
| 3193 |
-
"loss": 3.4175,
|
| 3194 |
-
"step": 9100
|
| 3195 |
-
},
|
| 3196 |
-
{
|
| 3197 |
-
"epoch": 0.73,
|
| 3198 |
-
"grad_norm": 22.310564041137695,
|
| 3199 |
-
"learning_rate": 1.7848886898095356e-05,
|
| 3200 |
-
"loss": 3.4606,
|
| 3201 |
-
"step": 9120
|
| 3202 |
-
},
|
| 3203 |
-
{
|
| 3204 |
-
"epoch": 0.73,
|
| 3205 |
-
"grad_norm": 31.924890518188477,
|
| 3206 |
-
"learning_rate": 1.7801174965633898e-05,
|
| 3207 |
-
"loss": 3.3977,
|
| 3208 |
-
"step": 9140
|
| 3209 |
-
},
|
| 3210 |
-
{
|
| 3211 |
-
"epoch": 0.73,
|
| 3212 |
-
"grad_norm": 32.8968620300293,
|
| 3213 |
-
"learning_rate": 1.775343365017254e-05,
|
| 3214 |
-
"loss": 3.5308,
|
| 3215 |
-
"step": 9160
|
| 3216 |
-
},
|
| 3217 |
-
{
|
| 3218 |
-
"epoch": 0.73,
|
| 3219 |
-
"grad_norm": 23.338972091674805,
|
| 3220 |
-
"learning_rate": 1.77056634524951e-05,
|
| 3221 |
-
"loss": 3.5206,
|
| 3222 |
-
"step": 9180
|
| 3223 |
-
},
|
| 3224 |
-
{
|
| 3225 |
-
"epoch": 0.74,
|
| 3226 |
-
"grad_norm": 29.5195369720459,
|
| 3227 |
-
"learning_rate": 1.7657864873688343e-05,
|
| 3228 |
-
"loss": 3.4515,
|
| 3229 |
-
"step": 9200
|
| 3230 |
-
},
|
| 3231 |
-
{
|
| 3232 |
-
"epoch": 0.74,
|
| 3233 |
-
"grad_norm": 41.386260986328125,
|
| 3234 |
-
"learning_rate": 1.761003841513674e-05,
|
| 3235 |
-
"loss": 3.4684,
|
| 3236 |
-
"step": 9220
|
| 3237 |
-
},
|
| 3238 |
-
{
|
| 3239 |
-
"epoch": 0.74,
|
| 3240 |
-
"grad_norm": 32.939849853515625,
|
| 3241 |
-
"learning_rate": 1.7562184578517207e-05,
|
| 3242 |
-
"loss": 3.4478,
|
| 3243 |
-
"step": 9240
|
| 3244 |
-
},
|
| 3245 |
-
{
|
| 3246 |
-
"epoch": 0.74,
|
| 3247 |
-
"grad_norm": 22.82186508178711,
|
| 3248 |
-
"learning_rate": 1.751430386579385e-05,
|
| 3249 |
-
"loss": 3.4226,
|
| 3250 |
-
"step": 9260
|
| 3251 |
-
},
|
| 3252 |
-
{
|
| 3253 |
-
"epoch": 0.74,
|
| 3254 |
-
"grad_norm": 25.349609375,
|
| 3255 |
-
"learning_rate": 1.7466396779212695e-05,
|
| 3256 |
-
"loss": 3.4381,
|
| 3257 |
-
"step": 9280
|
| 3258 |
-
},
|
| 3259 |
-
{
|
| 3260 |
-
"epoch": 0.74,
|
| 3261 |
-
"grad_norm": 26.957141876220703,
|
| 3262 |
-
"learning_rate": 1.7418463821296398e-05,
|
| 3263 |
-
"loss": 3.4075,
|
| 3264 |
-
"step": 9300
|
| 3265 |
-
},
|
| 3266 |
-
{
|
| 3267 |
-
"epoch": 0.75,
|
| 3268 |
-
"grad_norm": 52.64670944213867,
|
| 3269 |
-
"learning_rate": 1.7370505494839012e-05,
|
| 3270 |
-
"loss": 3.4542,
|
| 3271 |
-
"step": 9320
|
| 3272 |
-
},
|
| 3273 |
-
{
|
| 3274 |
-
"epoch": 0.75,
|
| 3275 |
-
"grad_norm": 30.222015380859375,
|
| 3276 |
-
"learning_rate": 1.73225223029007e-05,
|
| 3277 |
-
"loss": 3.5017,
|
| 3278 |
-
"step": 9340
|
| 3279 |
-
},
|
| 3280 |
-
{
|
| 3281 |
-
"epoch": 0.75,
|
| 3282 |
-
"grad_norm": 33.9267578125,
|
| 3283 |
-
"learning_rate": 1.7274514748802418e-05,
|
| 3284 |
-
"loss": 3.4322,
|
| 3285 |
-
"step": 9360
|
| 3286 |
-
},
|
| 3287 |
-
{
|
| 3288 |
-
"epoch": 0.75,
|
| 3289 |
-
"grad_norm": 29.3850040435791,
|
| 3290 |
-
"learning_rate": 1.7226483336120707e-05,
|
| 3291 |
-
"loss": 3.443,
|
| 3292 |
-
"step": 9380
|
| 3293 |
-
},
|
| 3294 |
-
{
|
| 3295 |
-
"epoch": 0.75,
|
| 3296 |
-
"grad_norm": 19.676841735839844,
|
| 3297 |
-
"learning_rate": 1.7178428568682357e-05,
|
| 3298 |
-
"loss": 3.443,
|
| 3299 |
-
"step": 9400
|
| 3300 |
}
|
| 3301 |
],
|
| 3302 |
"logging_steps": 20,
|
| 3303 |
"max_steps": 20000,
|
| 3304 |
"num_input_tokens_seen": 0,
|
| 3305 |
-
"num_train_epochs":
|
| 3306 |
"save_steps": 20,
|
| 3307 |
-
"total_flos":
|
| 3308 |
"train_batch_size": 8,
|
| 3309 |
"trial_name": null,
|
| 3310 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.0002830335536277826,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 20,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 58.19491958618164,
|
| 14 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 15 |
+
"loss": 4.5462,
|
| 16 |
"step": 20
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
}
|
| 18 |
],
|
| 19 |
"logging_steps": 20,
|
| 20 |
"max_steps": 20000,
|
| 21 |
"num_input_tokens_seen": 0,
|
| 22 |
+
"num_train_epochs": 1,
|
| 23 |
"save_steps": 20,
|
| 24 |
+
"total_flos": 41401061670912.0,
|
| 25 |
"train_batch_size": 8,
|
| 26 |
"trial_name": null,
|
| 27 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81b8007143b3dd4d9fc05510e0e4b454b7d10dc1b483b5df5fa0fb37f29b72e7
|
| 3 |
size 4984
|