Delete trainer_state.json
Browse files- trainer_state.json +0 -508
trainer_state.json
DELETED
|
@@ -1,508 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 3.0,
|
| 6 |
-
"eval_steps": 1000,
|
| 7 |
-
"global_step": 3180,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.047175374454534735,
|
| 14 |
-
"grad_norm": 0.6658920049667358,
|
| 15 |
-
"learning_rate": 2.0545073375262055e-06,
|
| 16 |
-
"loss": 0.7831,
|
| 17 |
-
"step": 50
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.09435074890906947,
|
| 21 |
-
"grad_norm": 0.6541114449501038,
|
| 22 |
-
"learning_rate": 4.150943396226416e-06,
|
| 23 |
-
"loss": 0.6891,
|
| 24 |
-
"step": 100
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.1415261233636042,
|
| 28 |
-
"grad_norm": 0.64142906665802,
|
| 29 |
-
"learning_rate": 6.247379454926625e-06,
|
| 30 |
-
"loss": 0.6727,
|
| 31 |
-
"step": 150
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 0.18870149781813894,
|
| 35 |
-
"grad_norm": 0.645492434501648,
|
| 36 |
-
"learning_rate": 8.343815513626834e-06,
|
| 37 |
-
"loss": 0.6608,
|
| 38 |
-
"step": 200
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.23587687227267368,
|
| 42 |
-
"grad_norm": 0.6750236749649048,
|
| 43 |
-
"learning_rate": 1.0440251572327045e-05,
|
| 44 |
-
"loss": 0.6656,
|
| 45 |
-
"step": 250
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 0.2830522467272084,
|
| 49 |
-
"grad_norm": 0.6192681789398193,
|
| 50 |
-
"learning_rate": 1.2536687631027256e-05,
|
| 51 |
-
"loss": 0.6613,
|
| 52 |
-
"step": 300
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"epoch": 0.33022762118174315,
|
| 56 |
-
"grad_norm": 0.6124400496482849,
|
| 57 |
-
"learning_rate": 1.4633123689727464e-05,
|
| 58 |
-
"loss": 0.6612,
|
| 59 |
-
"step": 350
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 0.3774029956362779,
|
| 63 |
-
"grad_norm": 0.5999819040298462,
|
| 64 |
-
"learning_rate": 1.6729559748427675e-05,
|
| 65 |
-
"loss": 0.6587,
|
| 66 |
-
"step": 400
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"epoch": 0.4245783700908126,
|
| 70 |
-
"grad_norm": 0.6131274104118347,
|
| 71 |
-
"learning_rate": 1.8825995807127882e-05,
|
| 72 |
-
"loss": 0.6613,
|
| 73 |
-
"step": 450
|
| 74 |
-
},
|
| 75 |
-
{
|
| 76 |
-
"epoch": 0.47175374454534735,
|
| 77 |
-
"grad_norm": 0.5893041491508484,
|
| 78 |
-
"learning_rate": 1.9996731117142877e-05,
|
| 79 |
-
"loss": 0.6615,
|
| 80 |
-
"step": 500
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.518929118999882,
|
| 84 |
-
"grad_norm": 0.546073853969574,
|
| 85 |
-
"learning_rate": 1.996500635384337e-05,
|
| 86 |
-
"loss": 0.6598,
|
| 87 |
-
"step": 550
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 0.5661044934544168,
|
| 91 |
-
"grad_norm": 0.5713281035423279,
|
| 92 |
-
"learning_rate": 1.9899637947477248e-05,
|
| 93 |
-
"loss": 0.6586,
|
| 94 |
-
"step": 600
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"epoch": 0.6132798679089515,
|
| 98 |
-
"grad_norm": 0.5555443167686462,
|
| 99 |
-
"learning_rate": 1.9800846593471427e-05,
|
| 100 |
-
"loss": 0.6602,
|
| 101 |
-
"step": 650
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.6604552423634863,
|
| 105 |
-
"grad_norm": 0.5415436625480652,
|
| 106 |
-
"learning_rate": 1.966896582909968e-05,
|
| 107 |
-
"loss": 0.6577,
|
| 108 |
-
"step": 700
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"epoch": 0.707630616818021,
|
| 112 |
-
"grad_norm": 0.5836868286132812,
|
| 113 |
-
"learning_rate": 1.9504440907401113e-05,
|
| 114 |
-
"loss": 0.6544,
|
| 115 |
-
"step": 750
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"epoch": 0.7548059912725558,
|
| 119 |
-
"grad_norm": 0.539230465888977,
|
| 120 |
-
"learning_rate": 1.9307827293926344e-05,
|
| 121 |
-
"loss": 0.6577,
|
| 122 |
-
"step": 800
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.8019813657270904,
|
| 126 |
-
"grad_norm": 0.5532299280166626,
|
| 127 |
-
"learning_rate": 1.9079788791386468e-05,
|
| 128 |
-
"loss": 0.6569,
|
| 129 |
-
"step": 850
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.8491567401816252,
|
| 133 |
-
"grad_norm": 0.5102491974830627,
|
| 134 |
-
"learning_rate": 1.8821095298536435e-05,
|
| 135 |
-
"loss": 0.6502,
|
| 136 |
-
"step": 900
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 0.8963321146361599,
|
| 140 |
-
"grad_norm": 0.5178349614143372,
|
| 141 |
-
"learning_rate": 1.853262021085921e-05,
|
| 142 |
-
"loss": 0.6559,
|
| 143 |
-
"step": 950
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"epoch": 0.9435074890906947,
|
| 147 |
-
"grad_norm": 0.5248669385910034,
|
| 148 |
-
"learning_rate": 1.821533747182645e-05,
|
| 149 |
-
"loss": 0.6549,
|
| 150 |
-
"step": 1000
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 0.9435074890906947,
|
| 154 |
-
"eval_loss": 0.6482675671577454,
|
| 155 |
-
"eval_runtime": 113.518,
|
| 156 |
-
"eval_samples_per_second": 66.395,
|
| 157 |
-
"eval_steps_per_second": 2.079,
|
| 158 |
-
"step": 1000
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 0.9906828635452294,
|
| 162 |
-
"grad_norm": 0.5705227255821228,
|
| 163 |
-
"learning_rate": 1.787031828469124e-05,
|
| 164 |
-
"loss": 0.6477,
|
| 165 |
-
"step": 1050
|
| 166 |
-
},
|
| 167 |
-
{
|
| 168 |
-
"epoch": 1.0377402995636278,
|
| 169 |
-
"grad_norm": 0.5323419570922852,
|
| 170 |
-
"learning_rate": 1.7498727495914378e-05,
|
| 171 |
-
"loss": 0.5975,
|
| 172 |
-
"step": 1100
|
| 173 |
-
},
|
| 174 |
-
{
|
| 175 |
-
"epoch": 1.0849156740181625,
|
| 176 |
-
"grad_norm": 0.5340594053268433,
|
| 177 |
-
"learning_rate": 1.710181966243447e-05,
|
| 178 |
-
"loss": 0.5819,
|
| 179 |
-
"step": 1150
|
| 180 |
-
},
|
| 181 |
-
{
|
| 182 |
-
"epoch": 1.1320910484726974,
|
| 183 |
-
"grad_norm": 0.49585267901420593,
|
| 184 |
-
"learning_rate": 1.6680934816059403e-05,
|
| 185 |
-
"loss": 0.5819,
|
| 186 |
-
"step": 1200
|
| 187 |
-
},
|
| 188 |
-
{
|
| 189 |
-
"epoch": 1.179266422927232,
|
| 190 |
-
"grad_norm": 0.500704824924469,
|
| 191 |
-
"learning_rate": 1.623749393927938e-05,
|
| 192 |
-
"loss": 0.5853,
|
| 193 |
-
"step": 1250
|
| 194 |
-
},
|
| 195 |
-
{
|
| 196 |
-
"epoch": 1.2264417973817667,
|
| 197 |
-
"grad_norm": 0.49850329756736755,
|
| 198 |
-
"learning_rate": 1.5772994167775986e-05,
|
| 199 |
-
"loss": 0.5844,
|
| 200 |
-
"step": 1300
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 1.2736171718363014,
|
| 204 |
-
"grad_norm": 0.5016334652900696,
|
| 205 |
-
"learning_rate": 1.5289003735824454e-05,
|
| 206 |
-
"loss": 0.5866,
|
| 207 |
-
"step": 1350
|
| 208 |
-
},
|
| 209 |
-
{
|
| 210 |
-
"epoch": 1.320792546290836,
|
| 211 |
-
"grad_norm": 0.5468031764030457,
|
| 212 |
-
"learning_rate": 1.4787156681654358e-05,
|
| 213 |
-
"loss": 0.5819,
|
| 214 |
-
"step": 1400
|
| 215 |
-
},
|
| 216 |
-
{
|
| 217 |
-
"epoch": 1.367967920745371,
|
| 218 |
-
"grad_norm": 0.5233626961708069,
|
| 219 |
-
"learning_rate": 1.426914733064444e-05,
|
| 220 |
-
"loss": 0.5849,
|
| 221 |
-
"step": 1450
|
| 222 |
-
},
|
| 223 |
-
{
|
| 224 |
-
"epoch": 1.4151432951999057,
|
| 225 |
-
"grad_norm": 0.5027523040771484,
|
| 226 |
-
"learning_rate": 1.373672457497717e-05,
|
| 227 |
-
"loss": 0.5844,
|
| 228 |
-
"step": 1500
|
| 229 |
-
},
|
| 230 |
-
{
|
| 231 |
-
"epoch": 1.4623186696544404,
|
| 232 |
-
"grad_norm": 0.5051872730255127,
|
| 233 |
-
"learning_rate": 1.3191685969066082e-05,
|
| 234 |
-
"loss": 0.5846,
|
| 235 |
-
"step": 1550
|
| 236 |
-
},
|
| 237 |
-
{
|
| 238 |
-
"epoch": 1.5094940441089753,
|
| 239 |
-
"grad_norm": 0.5215147733688354,
|
| 240 |
-
"learning_rate": 1.2635871660690677e-05,
|
| 241 |
-
"loss": 0.5829,
|
| 242 |
-
"step": 1600
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 1.55666941856351,
|
| 246 |
-
"grad_norm": 0.5042491555213928,
|
| 247 |
-
"learning_rate": 1.2071158178328547e-05,
|
| 248 |
-
"loss": 0.5854,
|
| 249 |
-
"step": 1650
|
| 250 |
-
},
|
| 251 |
-
{
|
| 252 |
-
"epoch": 1.6038447930180446,
|
| 253 |
-
"grad_norm": 0.5223520994186401,
|
| 254 |
-
"learning_rate": 1.1499452095659713e-05,
|
| 255 |
-
"loss": 0.5833,
|
| 256 |
-
"step": 1700
|
| 257 |
-
},
|
| 258 |
-
{
|
| 259 |
-
"epoch": 1.6510201674725793,
|
| 260 |
-
"grad_norm": 0.4980650246143341,
|
| 261 |
-
"learning_rate": 1.092268359463302e-05,
|
| 262 |
-
"loss": 0.5808,
|
| 263 |
-
"step": 1750
|
| 264 |
-
},
|
| 265 |
-
{
|
| 266 |
-
"epoch": 1.698195541927114,
|
| 267 |
-
"grad_norm": 0.5121437907218933,
|
| 268 |
-
"learning_rate": 1.0342799948826788e-05,
|
| 269 |
-
"loss": 0.5823,
|
| 270 |
-
"step": 1800
|
| 271 |
-
},
|
| 272 |
-
{
|
| 273 |
-
"epoch": 1.7453709163816487,
|
| 274 |
-
"grad_norm": 0.4902792274951935,
|
| 275 |
-
"learning_rate": 9.761758949105056e-06,
|
| 276 |
-
"loss": 0.5824,
|
| 277 |
-
"step": 1850
|
| 278 |
-
},
|
| 279 |
-
{
|
| 280 |
-
"epoch": 1.7925462908361836,
|
| 281 |
-
"grad_norm": 0.4946081042289734,
|
| 282 |
-
"learning_rate": 9.18152229376561e-06,
|
| 283 |
-
"loss": 0.5814,
|
| 284 |
-
"step": 1900
|
| 285 |
-
},
|
| 286 |
-
{
|
| 287 |
-
"epoch": 1.8397216652907182,
|
| 288 |
-
"grad_norm": 0.5029374361038208,
|
| 289 |
-
"learning_rate": 8.604048965495786e-06,
|
| 290 |
-
"loss": 0.5795,
|
| 291 |
-
"step": 1950
|
| 292 |
-
},
|
| 293 |
-
{
|
| 294 |
-
"epoch": 1.8868970397452531,
|
| 295 |
-
"grad_norm": 0.47936609387397766,
|
| 296 |
-
"learning_rate": 8.031288617496686e-06,
|
| 297 |
-
"loss": 0.5796,
|
| 298 |
-
"step": 2000
|
| 299 |
-
},
|
| 300 |
-
{
|
| 301 |
-
"epoch": 1.8868970397452531,
|
| 302 |
-
"eval_loss": 0.6402788162231445,
|
| 303 |
-
"eval_runtime": 113.5077,
|
| 304 |
-
"eval_samples_per_second": 66.401,
|
| 305 |
-
"eval_steps_per_second": 2.079,
|
| 306 |
-
"step": 2000
|
| 307 |
-
},
|
| 308 |
-
{
|
| 309 |
-
"epoch": 1.9340724141997878,
|
| 310 |
-
"grad_norm": 0.5165300369262695,
|
| 311 |
-
"learning_rate": 7.465174991105405e-06,
|
| 312 |
-
"loss": 0.5762,
|
| 313 |
-
"step": 2050
|
| 314 |
-
},
|
| 315 |
-
{
|
| 316 |
-
"epoch": 1.9812477886543225,
|
| 317 |
-
"grad_norm": 0.49469050765037537,
|
| 318 |
-
"learning_rate": 6.90761938713854e-06,
|
| 319 |
-
"loss": 0.5747,
|
| 320 |
-
"step": 2100
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 2.028305224672721,
|
| 324 |
-
"grad_norm": 0.5297324061393738,
|
| 325 |
-
"learning_rate": 6.360504212998903e-06,
|
| 326 |
-
"loss": 0.5171,
|
| 327 |
-
"step": 2150
|
| 328 |
-
},
|
| 329 |
-
{
|
| 330 |
-
"epoch": 2.0754805991272556,
|
| 331 |
-
"grad_norm": 0.5692960023880005,
|
| 332 |
-
"learning_rate": 5.825676627331614e-06,
|
| 333 |
-
"loss": 0.471,
|
| 334 |
-
"step": 2200
|
| 335 |
-
},
|
| 336 |
-
{
|
| 337 |
-
"epoch": 2.1226559735817903,
|
| 338 |
-
"grad_norm": 0.5705838799476624,
|
| 339 |
-
"learning_rate": 5.304942303686238e-06,
|
| 340 |
-
"loss": 0.4737,
|
| 341 |
-
"step": 2250
|
| 342 |
-
},
|
| 343 |
-
{
|
| 344 |
-
"epoch": 2.169831348036325,
|
| 345 |
-
"grad_norm": 0.5800350308418274,
|
| 346 |
-
"learning_rate": 4.800059334240049e-06,
|
| 347 |
-
"loss": 0.4734,
|
| 348 |
-
"step": 2300
|
| 349 |
-
},
|
| 350 |
-
{
|
| 351 |
-
"epoch": 2.2170067224908596,
|
| 352 |
-
"grad_norm": 0.5653440952301025,
|
| 353 |
-
"learning_rate": 4.3127322941645385e-06,
|
| 354 |
-
"loss": 0.4713,
|
| 355 |
-
"step": 2350
|
| 356 |
-
},
|
| 357 |
-
{
|
| 358 |
-
"epoch": 2.2641820969453947,
|
| 359 |
-
"grad_norm": 0.5822903513908386,
|
| 360 |
-
"learning_rate": 3.844606486674862e-06,
|
| 361 |
-
"loss": 0.4715,
|
| 362 |
-
"step": 2400
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 2.3113574713999294,
|
| 366 |
-
"grad_norm": 0.599399209022522,
|
| 367 |
-
"learning_rate": 3.3972623881920296e-06,
|
| 368 |
-
"loss": 0.47,
|
| 369 |
-
"step": 2450
|
| 370 |
-
},
|
| 371 |
-
{
|
| 372 |
-
"epoch": 2.358532845854464,
|
| 373 |
-
"grad_norm": 0.5810734033584595,
|
| 374 |
-
"learning_rate": 2.9722103123719324e-06,
|
| 375 |
-
"loss": 0.4661,
|
| 376 |
-
"step": 2500
|
| 377 |
-
},
|
| 378 |
-
{
|
| 379 |
-
"epoch": 2.405708220308999,
|
| 380 |
-
"grad_norm": 0.5994310975074768,
|
| 381 |
-
"learning_rate": 2.5708853110164346e-06,
|
| 382 |
-
"loss": 0.4686,
|
| 383 |
-
"step": 2550
|
| 384 |
-
},
|
| 385 |
-
{
|
| 386 |
-
"epoch": 2.4528835947635335,
|
| 387 |
-
"grad_norm": 0.577089786529541,
|
| 388 |
-
"learning_rate": 2.194642329081902e-06,
|
| 389 |
-
"loss": 0.4673,
|
| 390 |
-
"step": 2600
|
| 391 |
-
},
|
| 392 |
-
{
|
| 393 |
-
"epoch": 2.500058969218068,
|
| 394 |
-
"grad_norm": 0.578800618648529,
|
| 395 |
-
"learning_rate": 1.844751630142797e-06,
|
| 396 |
-
"loss": 0.4664,
|
| 397 |
-
"step": 2650
|
| 398 |
-
},
|
| 399 |
-
{
|
| 400 |
-
"epoch": 2.547234343672603,
|
| 401 |
-
"grad_norm": 0.5981403589248657,
|
| 402 |
-
"learning_rate": 1.5223945077547253e-06,
|
| 403 |
-
"loss": 0.4686,
|
| 404 |
-
"step": 2700
|
| 405 |
-
},
|
| 406 |
-
{
|
| 407 |
-
"epoch": 2.5944097181271375,
|
| 408 |
-
"grad_norm": 0.6022927165031433,
|
| 409 |
-
"learning_rate": 1.2286592971962152e-06,
|
| 410 |
-
"loss": 0.4676,
|
| 411 |
-
"step": 2750
|
| 412 |
-
},
|
| 413 |
-
{
|
| 414 |
-
"epoch": 2.641585092581672,
|
| 415 |
-
"grad_norm": 0.5744144916534424,
|
| 416 |
-
"learning_rate": 9.645377010542212e-07,
|
| 417 |
-
"loss": 0.4684,
|
| 418 |
-
"step": 2800
|
| 419 |
-
},
|
| 420 |
-
{
|
| 421 |
-
"epoch": 2.6887604670362073,
|
| 422 |
-
"grad_norm": 0.5747597813606262,
|
| 423 |
-
"learning_rate": 7.309214410588927e-07,
|
| 424 |
-
"loss": 0.4632,
|
| 425 |
-
"step": 2850
|
| 426 |
-
},
|
| 427 |
-
{
|
| 428 |
-
"epoch": 2.735935841490742,
|
| 429 |
-
"grad_norm": 0.5752475261688232,
|
| 430 |
-
"learning_rate": 5.285992474715796e-07,
|
| 431 |
-
"loss": 0.4643,
|
| 432 |
-
"step": 2900
|
| 433 |
-
},
|
| 434 |
-
{
|
| 435 |
-
"epoch": 2.7831112159452767,
|
| 436 |
-
"grad_norm": 0.5861772894859314,
|
| 437 |
-
"learning_rate": 3.5825419619046176e-07,
|
| 438 |
-
"loss": 0.465,
|
| 439 |
-
"step": 2950
|
| 440 |
-
},
|
| 441 |
-
{
|
| 442 |
-
"epoch": 2.8302865903998113,
|
| 443 |
-
"grad_norm": 0.5940195918083191,
|
| 444 |
-
"learning_rate": 2.2046140256418713e-07,
|
| 445 |
-
"loss": 0.4646,
|
| 446 |
-
"step": 3000
|
| 447 |
-
},
|
| 448 |
-
{
|
| 449 |
-
"epoch": 2.8302865903998113,
|
| 450 |
-
"eval_loss": 0.6850141286849976,
|
| 451 |
-
"eval_runtime": 113.5772,
|
| 452 |
-
"eval_samples_per_second": 66.36,
|
| 453 |
-
"eval_steps_per_second": 2.078,
|
| 454 |
-
"step": 3000
|
| 455 |
-
},
|
| 456 |
-
{
|
| 457 |
-
"epoch": 2.877461964854346,
|
| 458 |
-
"grad_norm": 0.5916756987571716,
|
| 459 |
-
"learning_rate": 1.1568607969963175e-07,
|
| 460 |
-
"loss": 0.4644,
|
| 461 |
-
"step": 3050
|
| 462 |
-
},
|
| 463 |
-
{
|
| 464 |
-
"epoch": 2.9246373393088807,
|
| 465 |
-
"grad_norm": 0.5982924103736877,
|
| 466 |
-
"learning_rate": 4.42819678192774e-08,
|
| 467 |
-
"loss": 0.4691,
|
| 468 |
-
"step": 3100
|
| 469 |
-
},
|
| 470 |
-
{
|
| 471 |
-
"epoch": 2.9718127137634154,
|
| 472 |
-
"grad_norm": 0.5770459175109863,
|
| 473 |
-
"learning_rate": 6.490139970976029e-09,
|
| 474 |
-
"loss": 0.4661,
|
| 475 |
-
"step": 3150
|
| 476 |
-
},
|
| 477 |
-
{
|
| 478 |
-
"epoch": 3.0,
|
| 479 |
-
"step": 3180,
|
| 480 |
-
"total_flos": 1.7575386520355865e+19,
|
| 481 |
-
"train_loss": 0.5721450322828953,
|
| 482 |
-
"train_runtime": 14690.2311,
|
| 483 |
-
"train_samples_per_second": 13.852,
|
| 484 |
-
"train_steps_per_second": 0.216
|
| 485 |
-
}
|
| 486 |
-
],
|
| 487 |
-
"logging_steps": 50,
|
| 488 |
-
"max_steps": 3180,
|
| 489 |
-
"num_input_tokens_seen": 0,
|
| 490 |
-
"num_train_epochs": 3,
|
| 491 |
-
"save_steps": 1000,
|
| 492 |
-
"stateful_callbacks": {
|
| 493 |
-
"TrainerControl": {
|
| 494 |
-
"args": {
|
| 495 |
-
"should_epoch_stop": false,
|
| 496 |
-
"should_evaluate": false,
|
| 497 |
-
"should_log": false,
|
| 498 |
-
"should_save": true,
|
| 499 |
-
"should_training_stop": true
|
| 500 |
-
},
|
| 501 |
-
"attributes": {}
|
| 502 |
-
}
|
| 503 |
-
},
|
| 504 |
-
"total_flos": 1.7575386520355865e+19,
|
| 505 |
-
"train_batch_size": 1,
|
| 506 |
-
"trial_name": null,
|
| 507 |
-
"trial_params": null
|
| 508 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|