Delete folder checkpoint-6440 with huggingface_hub
Browse files
checkpoint-6440/model.safetensors
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:d5e1fb06effa3e083fe2a19cd5d1f785d29749d6b4039d87028728634ff2ab84
|
| 3 |
-
size 1749099756
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6440/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:20c2e805f3dd16a00054e87fec4f666262363ff1eff61f37699c6ca56af373c4
|
| 3 |
-
size 3498435905
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6440/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:b87807198463d30a7e8e7df7649c0bf956339a84c2ad65d1450fa4f0145975ee
|
| 3 |
-
size 14645
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6440/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:52efdb72a390ba9eb971747212b8e3830d9a27a6caf97f048e884ef95171d007
|
| 3 |
-
size 1465
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6440/trainer_state.json
DELETED
|
@@ -1,975 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": 6440,
|
| 3 |
-
"best_metric": 0.836709384042993,
|
| 4 |
-
"best_model_checkpoint": "/content/outputs/frame2/checkpoint-6440",
|
| 5 |
-
"epoch": 5.0,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 6440,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.03881987577639751,
|
| 14 |
-
"grad_norm": 6.954199314117432,
|
| 15 |
-
"learning_rate": 1.2661498708010337e-06,
|
| 16 |
-
"loss": 7.1416796875,
|
| 17 |
-
"step": 50
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.07763975155279502,
|
| 21 |
-
"grad_norm": 6.37876033782959,
|
| 22 |
-
"learning_rate": 2.558139534883721e-06,
|
| 23 |
-
"loss": 7.1059375,
|
| 24 |
-
"step": 100
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.11645962732919254,
|
| 28 |
-
"grad_norm": 6.456676006317139,
|
| 29 |
-
"learning_rate": 3.850129198966409e-06,
|
| 30 |
-
"loss": 7.01484375,
|
| 31 |
-
"step": 150
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 0.15527950310559005,
|
| 35 |
-
"grad_norm": 7.245243549346924,
|
| 36 |
-
"learning_rate": 5.1421188630490955e-06,
|
| 37 |
-
"loss": 6.883046875,
|
| 38 |
-
"step": 200
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.19409937888198758,
|
| 42 |
-
"grad_norm": 7.722668647766113,
|
| 43 |
-
"learning_rate": 6.434108527131784e-06,
|
| 44 |
-
"loss": 6.7165234375,
|
| 45 |
-
"step": 250
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 0.2329192546583851,
|
| 49 |
-
"grad_norm": 8.763988494873047,
|
| 50 |
-
"learning_rate": 7.72609819121447e-06,
|
| 51 |
-
"loss": 6.39888671875,
|
| 52 |
-
"step": 300
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"epoch": 0.2717391304347826,
|
| 56 |
-
"grad_norm": 9.204142570495605,
|
| 57 |
-
"learning_rate": 9.018087855297158e-06,
|
| 58 |
-
"loss": 6.1780078125,
|
| 59 |
-
"step": 350
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 0.3105590062111801,
|
| 63 |
-
"grad_norm": 9.30582046508789,
|
| 64 |
-
"learning_rate": 9.980175119775319e-06,
|
| 65 |
-
"loss": 5.985400390625,
|
| 66 |
-
"step": 400
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"epoch": 0.34937888198757766,
|
| 70 |
-
"grad_norm": 10.375268936157227,
|
| 71 |
-
"learning_rate": 9.897571452172478e-06,
|
| 72 |
-
"loss": 5.727763671875,
|
| 73 |
-
"step": 450
|
| 74 |
-
},
|
| 75 |
-
{
|
| 76 |
-
"epoch": 0.38819875776397517,
|
| 77 |
-
"grad_norm": 10.128670692443848,
|
| 78 |
-
"learning_rate": 9.814967784569636e-06,
|
| 79 |
-
"loss": 5.58036376953125,
|
| 80 |
-
"step": 500
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.42701863354037267,
|
| 84 |
-
"grad_norm": 10.368155479431152,
|
| 85 |
-
"learning_rate": 9.732364116966793e-06,
|
| 86 |
-
"loss": 5.530234375,
|
| 87 |
-
"step": 550
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 0.4658385093167702,
|
| 91 |
-
"grad_norm": 10.306320190429688,
|
| 92 |
-
"learning_rate": 9.649760449363953e-06,
|
| 93 |
-
"loss": 5.449205322265625,
|
| 94 |
-
"step": 600
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"epoch": 0.5046583850931677,
|
| 98 |
-
"grad_norm": 10.807398796081543,
|
| 99 |
-
"learning_rate": 9.56715678176111e-06,
|
| 100 |
-
"loss": 5.28119140625,
|
| 101 |
-
"step": 650
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"epoch": 0.5434782608695652,
|
| 105 |
-
"grad_norm": 10.359151840209961,
|
| 106 |
-
"learning_rate": 9.48455311415827e-06,
|
| 107 |
-
"loss": 5.224617919921875,
|
| 108 |
-
"step": 700
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"epoch": 0.5822981366459627,
|
| 112 |
-
"grad_norm": 10.162821769714355,
|
| 113 |
-
"learning_rate": 9.401949446555427e-06,
|
| 114 |
-
"loss": 4.982539672851562,
|
| 115 |
-
"step": 750
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"epoch": 0.6211180124223602,
|
| 119 |
-
"grad_norm": 11.337475776672363,
|
| 120 |
-
"learning_rate": 9.319345778952586e-06,
|
| 121 |
-
"loss": 4.874409484863281,
|
| 122 |
-
"step": 800
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.6599378881987578,
|
| 126 |
-
"grad_norm": 11.128438949584961,
|
| 127 |
-
"learning_rate": 9.236742111349744e-06,
|
| 128 |
-
"loss": 4.720284423828125,
|
| 129 |
-
"step": 850
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"epoch": 0.6987577639751553,
|
| 133 |
-
"grad_norm": 11.113153457641602,
|
| 134 |
-
"learning_rate": 9.154138443746903e-06,
|
| 135 |
-
"loss": 4.625420837402344,
|
| 136 |
-
"step": 900
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"epoch": 0.7375776397515528,
|
| 140 |
-
"grad_norm": 11.993304252624512,
|
| 141 |
-
"learning_rate": 9.071534776144063e-06,
|
| 142 |
-
"loss": 4.4766607666015625,
|
| 143 |
-
"step": 950
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"epoch": 0.7763975155279503,
|
| 147 |
-
"grad_norm": 11.786646842956543,
|
| 148 |
-
"learning_rate": 8.98893110854122e-06,
|
| 149 |
-
"loss": 4.373000793457031,
|
| 150 |
-
"step": 1000
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 0.8152173913043478,
|
| 154 |
-
"grad_norm": 11.01918888092041,
|
| 155 |
-
"learning_rate": 8.906327440938378e-06,
|
| 156 |
-
"loss": 4.122265625,
|
| 157 |
-
"step": 1050
|
| 158 |
-
},
|
| 159 |
-
{
|
| 160 |
-
"epoch": 0.8540372670807453,
|
| 161 |
-
"grad_norm": 14.887975692749023,
|
| 162 |
-
"learning_rate": 8.823723773335536e-06,
|
| 163 |
-
"loss": 4.2356741333007815,
|
| 164 |
-
"step": 1100
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.8928571428571429,
|
| 168 |
-
"grad_norm": 11.712749481201172,
|
| 169 |
-
"learning_rate": 8.741120105732695e-06,
|
| 170 |
-
"loss": 4.030892333984375,
|
| 171 |
-
"step": 1150
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 0.9316770186335404,
|
| 175 |
-
"grad_norm": 11.420888900756836,
|
| 176 |
-
"learning_rate": 8.658516438129854e-06,
|
| 177 |
-
"loss": 3.809114685058594,
|
| 178 |
-
"step": 1200
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 0.9704968944099379,
|
| 182 |
-
"grad_norm": 11.520773887634277,
|
| 183 |
-
"learning_rate": 8.575912770527012e-06,
|
| 184 |
-
"loss": 3.6759454345703126,
|
| 185 |
-
"step": 1250
|
| 186 |
-
},
|
| 187 |
-
{
|
| 188 |
-
"epoch": 1.0,
|
| 189 |
-
"eval_accuracy": 0.5010334849111203,
|
| 190 |
-
"eval_loss": 3.0984573364257812,
|
| 191 |
-
"eval_runtime": 9.2642,
|
| 192 |
-
"eval_samples_per_second": 261.114,
|
| 193 |
-
"eval_steps_per_second": 16.407,
|
| 194 |
-
"step": 1288
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 1.0093167701863355,
|
| 198 |
-
"grad_norm": 12.707167625427246,
|
| 199 |
-
"learning_rate": 8.493309102924171e-06,
|
| 200 |
-
"loss": 3.6287490844726564,
|
| 201 |
-
"step": 1300
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 1.0481366459627328,
|
| 205 |
-
"grad_norm": 11.602799415588379,
|
| 206 |
-
"learning_rate": 8.410705435321329e-06,
|
| 207 |
-
"loss": 3.4154583740234377,
|
| 208 |
-
"step": 1350
|
| 209 |
-
},
|
| 210 |
-
{
|
| 211 |
-
"epoch": 1.0869565217391304,
|
| 212 |
-
"grad_norm": 10.797930717468262,
|
| 213 |
-
"learning_rate": 8.328101767718486e-06,
|
| 214 |
-
"loss": 3.2645416259765625,
|
| 215 |
-
"step": 1400
|
| 216 |
-
},
|
| 217 |
-
{
|
| 218 |
-
"epoch": 1.125776397515528,
|
| 219 |
-
"grad_norm": 14.686330795288086,
|
| 220 |
-
"learning_rate": 8.245498100115646e-06,
|
| 221 |
-
"loss": 3.0807373046875,
|
| 222 |
-
"step": 1450
|
| 223 |
-
},
|
| 224 |
-
{
|
| 225 |
-
"epoch": 1.1645962732919255,
|
| 226 |
-
"grad_norm": 11.849090576171875,
|
| 227 |
-
"learning_rate": 8.162894432512805e-06,
|
| 228 |
-
"loss": 3.089696350097656,
|
| 229 |
-
"step": 1500
|
| 230 |
-
},
|
| 231 |
-
{
|
| 232 |
-
"epoch": 1.203416149068323,
|
| 233 |
-
"grad_norm": 10.781908988952637,
|
| 234 |
-
"learning_rate": 8.080290764909963e-06,
|
| 235 |
-
"loss": 2.9422247314453127,
|
| 236 |
-
"step": 1550
|
| 237 |
-
},
|
| 238 |
-
{
|
| 239 |
-
"epoch": 1.2422360248447206,
|
| 240 |
-
"grad_norm": 12.094367980957031,
|
| 241 |
-
"learning_rate": 7.99768709730712e-06,
|
| 242 |
-
"loss": 2.857108154296875,
|
| 243 |
-
"step": 1600
|
| 244 |
-
},
|
| 245 |
-
{
|
| 246 |
-
"epoch": 1.281055900621118,
|
| 247 |
-
"grad_norm": 13.91542911529541,
|
| 248 |
-
"learning_rate": 7.91508342970428e-06,
|
| 249 |
-
"loss": 2.9898529052734375,
|
| 250 |
-
"step": 1650
|
| 251 |
-
},
|
| 252 |
-
{
|
| 253 |
-
"epoch": 1.3198757763975155,
|
| 254 |
-
"grad_norm": 11.836753845214844,
|
| 255 |
-
"learning_rate": 7.832479762101437e-06,
|
| 256 |
-
"loss": 2.714930114746094,
|
| 257 |
-
"step": 1700
|
| 258 |
-
},
|
| 259 |
-
{
|
| 260 |
-
"epoch": 1.358695652173913,
|
| 261 |
-
"grad_norm": 12.638890266418457,
|
| 262 |
-
"learning_rate": 7.749876094498597e-06,
|
| 263 |
-
"loss": 2.671651611328125,
|
| 264 |
-
"step": 1750
|
| 265 |
-
},
|
| 266 |
-
{
|
| 267 |
-
"epoch": 1.3975155279503104,
|
| 268 |
-
"grad_norm": 13.112387657165527,
|
| 269 |
-
"learning_rate": 7.667272426895754e-06,
|
| 270 |
-
"loss": 2.6401641845703123,
|
| 271 |
-
"step": 1800
|
| 272 |
-
},
|
| 273 |
-
{
|
| 274 |
-
"epoch": 1.4363354037267082,
|
| 275 |
-
"grad_norm": 12.418985366821289,
|
| 276 |
-
"learning_rate": 7.584668759292914e-06,
|
| 277 |
-
"loss": 2.4186663818359375,
|
| 278 |
-
"step": 1850
|
| 279 |
-
},
|
| 280 |
-
{
|
| 281 |
-
"epoch": 1.4751552795031055,
|
| 282 |
-
"grad_norm": 10.9830904006958,
|
| 283 |
-
"learning_rate": 7.502065091690072e-06,
|
| 284 |
-
"loss": 2.366060943603516,
|
| 285 |
-
"step": 1900
|
| 286 |
-
},
|
| 287 |
-
{
|
| 288 |
-
"epoch": 1.513975155279503,
|
| 289 |
-
"grad_norm": 12.341230392456055,
|
| 290 |
-
"learning_rate": 7.41946142408723e-06,
|
| 291 |
-
"loss": 2.284035949707031,
|
| 292 |
-
"step": 1950
|
| 293 |
-
},
|
| 294 |
-
{
|
| 295 |
-
"epoch": 1.5527950310559007,
|
| 296 |
-
"grad_norm": 8.954991340637207,
|
| 297 |
-
"learning_rate": 7.336857756484389e-06,
|
| 298 |
-
"loss": 2.3693133544921876,
|
| 299 |
-
"step": 2000
|
| 300 |
-
},
|
| 301 |
-
{
|
| 302 |
-
"epoch": 1.591614906832298,
|
| 303 |
-
"grad_norm": 22.260467529296875,
|
| 304 |
-
"learning_rate": 7.2542540888815475e-06,
|
| 305 |
-
"loss": 2.2734780883789063,
|
| 306 |
-
"step": 2050
|
| 307 |
-
},
|
| 308 |
-
{
|
| 309 |
-
"epoch": 1.6304347826086958,
|
| 310 |
-
"grad_norm": 13.057415008544922,
|
| 311 |
-
"learning_rate": 7.171650421278705e-06,
|
| 312 |
-
"loss": 2.126544647216797,
|
| 313 |
-
"step": 2100
|
| 314 |
-
},
|
| 315 |
-
{
|
| 316 |
-
"epoch": 1.6692546583850931,
|
| 317 |
-
"grad_norm": 9.723812103271484,
|
| 318 |
-
"learning_rate": 7.0890467536758636e-06,
|
| 319 |
-
"loss": 2.184694519042969,
|
| 320 |
-
"step": 2150
|
| 321 |
-
},
|
| 322 |
-
{
|
| 323 |
-
"epoch": 1.7080745341614907,
|
| 324 |
-
"grad_norm": 14.918439865112305,
|
| 325 |
-
"learning_rate": 7.006443086073023e-06,
|
| 326 |
-
"loss": 2.0987617492675783,
|
| 327 |
-
"step": 2200
|
| 328 |
-
},
|
| 329 |
-
{
|
| 330 |
-
"epoch": 1.7468944099378882,
|
| 331 |
-
"grad_norm": 16.97414207458496,
|
| 332 |
-
"learning_rate": 6.9238394184701805e-06,
|
| 333 |
-
"loss": 2.126544952392578,
|
| 334 |
-
"step": 2250
|
| 335 |
-
},
|
| 336 |
-
{
|
| 337 |
-
"epoch": 1.7857142857142856,
|
| 338 |
-
"grad_norm": 10.108041763305664,
|
| 339 |
-
"learning_rate": 6.841235750867339e-06,
|
| 340 |
-
"loss": 2.0426617431640626,
|
| 341 |
-
"step": 2300
|
| 342 |
-
},
|
| 343 |
-
{
|
| 344 |
-
"epoch": 1.8245341614906834,
|
| 345 |
-
"grad_norm": 17.742259979248047,
|
| 346 |
-
"learning_rate": 6.758632083264498e-06,
|
| 347 |
-
"loss": 1.9366094970703125,
|
| 348 |
-
"step": 2350
|
| 349 |
-
},
|
| 350 |
-
{
|
| 351 |
-
"epoch": 1.8633540372670807,
|
| 352 |
-
"grad_norm": 16.957307815551758,
|
| 353 |
-
"learning_rate": 6.676028415661656e-06,
|
| 354 |
-
"loss": 2.1593305969238283,
|
| 355 |
-
"step": 2400
|
| 356 |
-
},
|
| 357 |
-
{
|
| 358 |
-
"epoch": 1.9021739130434783,
|
| 359 |
-
"grad_norm": 11.384793281555176,
|
| 360 |
-
"learning_rate": 6.593424748058814e-06,
|
| 361 |
-
"loss": 1.7849739074707032,
|
| 362 |
-
"step": 2450
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 1.9409937888198758,
|
| 366 |
-
"grad_norm": 12.84943962097168,
|
| 367 |
-
"learning_rate": 6.510821080455972e-06,
|
| 368 |
-
"loss": 1.9132063293457031,
|
| 369 |
-
"step": 2500
|
| 370 |
-
},
|
| 371 |
-
{
|
| 372 |
-
"epoch": 1.9798136645962732,
|
| 373 |
-
"grad_norm": 20.355567932128906,
|
| 374 |
-
"learning_rate": 6.428217412853131e-06,
|
| 375 |
-
"loss": 2.026173553466797,
|
| 376 |
-
"step": 2550
|
| 377 |
-
},
|
| 378 |
-
{
|
| 379 |
-
"epoch": 2.0,
|
| 380 |
-
"eval_accuracy": 0.7416287722199256,
|
| 381 |
-
"eval_loss": 1.5747343301773071,
|
| 382 |
-
"eval_runtime": 9.2952,
|
| 383 |
-
"eval_samples_per_second": 260.242,
|
| 384 |
-
"eval_steps_per_second": 16.353,
|
| 385 |
-
"step": 2576
|
| 386 |
-
},
|
| 387 |
-
{
|
| 388 |
-
"epoch": 2.018633540372671,
|
| 389 |
-
"grad_norm": 9.695998191833496,
|
| 390 |
-
"learning_rate": 6.34561374525029e-06,
|
| 391 |
-
"loss": 1.7679312133789062,
|
| 392 |
-
"step": 2600
|
| 393 |
-
},
|
| 394 |
-
{
|
| 395 |
-
"epoch": 2.0574534161490683,
|
| 396 |
-
"grad_norm": 8.576300621032715,
|
| 397 |
-
"learning_rate": 6.2630100776474475e-06,
|
| 398 |
-
"loss": 1.699483642578125,
|
| 399 |
-
"step": 2650
|
| 400 |
-
},
|
| 401 |
-
{
|
| 402 |
-
"epoch": 2.0962732919254656,
|
| 403 |
-
"grad_norm": 9.594664573669434,
|
| 404 |
-
"learning_rate": 6.180406410044606e-06,
|
| 405 |
-
"loss": 1.6502584838867187,
|
| 406 |
-
"step": 2700
|
| 407 |
-
},
|
| 408 |
-
{
|
| 409 |
-
"epoch": 2.1350931677018634,
|
| 410 |
-
"grad_norm": 9.277363777160645,
|
| 411 |
-
"learning_rate": 6.097802742441765e-06,
|
| 412 |
-
"loss": 1.7002778625488282,
|
| 413 |
-
"step": 2750
|
| 414 |
-
},
|
| 415 |
-
{
|
| 416 |
-
"epoch": 2.1739130434782608,
|
| 417 |
-
"grad_norm": 9.334654808044434,
|
| 418 |
-
"learning_rate": 6.015199074838923e-06,
|
| 419 |
-
"loss": 1.6473841857910156,
|
| 420 |
-
"step": 2800
|
| 421 |
-
},
|
| 422 |
-
{
|
| 423 |
-
"epoch": 2.2127329192546585,
|
| 424 |
-
"grad_norm": 10.770955085754395,
|
| 425 |
-
"learning_rate": 5.932595407236081e-06,
|
| 426 |
-
"loss": 1.7070599365234376,
|
| 427 |
-
"step": 2850
|
| 428 |
-
},
|
| 429 |
-
{
|
| 430 |
-
"epoch": 2.251552795031056,
|
| 431 |
-
"grad_norm": 9.412900924682617,
|
| 432 |
-
"learning_rate": 5.849991739633241e-06,
|
| 433 |
-
"loss": 1.5200923156738282,
|
| 434 |
-
"step": 2900
|
| 435 |
-
},
|
| 436 |
-
{
|
| 437 |
-
"epoch": 2.2903726708074537,
|
| 438 |
-
"grad_norm": 8.708724021911621,
|
| 439 |
-
"learning_rate": 5.767388072030399e-06,
|
| 440 |
-
"loss": 1.5532386779785157,
|
| 441 |
-
"step": 2950
|
| 442 |
-
},
|
| 443 |
-
{
|
| 444 |
-
"epoch": 2.329192546583851,
|
| 445 |
-
"grad_norm": 16.545541763305664,
|
| 446 |
-
"learning_rate": 5.684784404427557e-06,
|
| 447 |
-
"loss": 1.4987814331054687,
|
| 448 |
-
"step": 3000
|
| 449 |
-
},
|
| 450 |
-
{
|
| 451 |
-
"epoch": 2.3680124223602483,
|
| 452 |
-
"grad_norm": 10.015238761901855,
|
| 453 |
-
"learning_rate": 5.602180736824715e-06,
|
| 454 |
-
"loss": 1.5647314453125,
|
| 455 |
-
"step": 3050
|
| 456 |
-
},
|
| 457 |
-
{
|
| 458 |
-
"epoch": 2.406832298136646,
|
| 459 |
-
"grad_norm": 12.274828910827637,
|
| 460 |
-
"learning_rate": 5.5195770692218746e-06,
|
| 461 |
-
"loss": 1.4848635864257813,
|
| 462 |
-
"step": 3100
|
| 463 |
-
},
|
| 464 |
-
{
|
| 465 |
-
"epoch": 2.4456521739130435,
|
| 466 |
-
"grad_norm": 10.715838432312012,
|
| 467 |
-
"learning_rate": 5.436973401619032e-06,
|
| 468 |
-
"loss": 1.4095025634765626,
|
| 469 |
-
"step": 3150
|
| 470 |
-
},
|
| 471 |
-
{
|
| 472 |
-
"epoch": 2.4844720496894412,
|
| 473 |
-
"grad_norm": 11.896187782287598,
|
| 474 |
-
"learning_rate": 5.354369734016191e-06,
|
| 475 |
-
"loss": 1.3431706237792969,
|
| 476 |
-
"step": 3200
|
| 477 |
-
},
|
| 478 |
-
{
|
| 479 |
-
"epoch": 2.5232919254658386,
|
| 480 |
-
"grad_norm": 12.155756950378418,
|
| 481 |
-
"learning_rate": 5.27176606641335e-06,
|
| 482 |
-
"loss": 1.3861366271972657,
|
| 483 |
-
"step": 3250
|
| 484 |
-
},
|
| 485 |
-
{
|
| 486 |
-
"epoch": 2.562111801242236,
|
| 487 |
-
"grad_norm": 8.469603538513184,
|
| 488 |
-
"learning_rate": 5.189162398810508e-06,
|
| 489 |
-
"loss": 1.3476237487792968,
|
| 490 |
-
"step": 3300
|
| 491 |
-
},
|
| 492 |
-
{
|
| 493 |
-
"epoch": 2.6009316770186337,
|
| 494 |
-
"grad_norm": 8.618110656738281,
|
| 495 |
-
"learning_rate": 5.106558731207666e-06,
|
| 496 |
-
"loss": 1.3636611938476562,
|
| 497 |
-
"step": 3350
|
| 498 |
-
},
|
| 499 |
-
{
|
| 500 |
-
"epoch": 2.639751552795031,
|
| 501 |
-
"grad_norm": 16.85852813720703,
|
| 502 |
-
"learning_rate": 5.023955063604824e-06,
|
| 503 |
-
"loss": 1.2232911682128906,
|
| 504 |
-
"step": 3400
|
| 505 |
-
},
|
| 506 |
-
{
|
| 507 |
-
"epoch": 2.678571428571429,
|
| 508 |
-
"grad_norm": 13.419601440429688,
|
| 509 |
-
"learning_rate": 4.941351396001983e-06,
|
| 510 |
-
"loss": 1.280044403076172,
|
| 511 |
-
"step": 3450
|
| 512 |
-
},
|
| 513 |
-
{
|
| 514 |
-
"epoch": 2.717391304347826,
|
| 515 |
-
"grad_norm": 9.274198532104492,
|
| 516 |
-
"learning_rate": 4.8587477283991415e-06,
|
| 517 |
-
"loss": 1.3282122802734375,
|
| 518 |
-
"step": 3500
|
| 519 |
-
},
|
| 520 |
-
{
|
| 521 |
-
"epoch": 2.7562111801242235,
|
| 522 |
-
"grad_norm": 10.566126823425293,
|
| 523 |
-
"learning_rate": 4.7761440607963e-06,
|
| 524 |
-
"loss": 1.2765910339355468,
|
| 525 |
-
"step": 3550
|
| 526 |
-
},
|
| 527 |
-
{
|
| 528 |
-
"epoch": 2.795031055900621,
|
| 529 |
-
"grad_norm": 12.02775764465332,
|
| 530 |
-
"learning_rate": 4.6935403931934584e-06,
|
| 531 |
-
"loss": 1.2579024505615235,
|
| 532 |
-
"step": 3600
|
| 533 |
-
},
|
| 534 |
-
{
|
| 535 |
-
"epoch": 2.8338509316770186,
|
| 536 |
-
"grad_norm": 6.950809478759766,
|
| 537 |
-
"learning_rate": 4.610936725590617e-06,
|
| 538 |
-
"loss": 1.1432655334472657,
|
| 539 |
-
"step": 3650
|
| 540 |
-
},
|
| 541 |
-
{
|
| 542 |
-
"epoch": 2.8726708074534164,
|
| 543 |
-
"grad_norm": 9.714532852172852,
|
| 544 |
-
"learning_rate": 4.5283330579877745e-06,
|
| 545 |
-
"loss": 1.243320541381836,
|
| 546 |
-
"step": 3700
|
| 547 |
-
},
|
| 548 |
-
{
|
| 549 |
-
"epoch": 2.9114906832298137,
|
| 550 |
-
"grad_norm": 8.740928649902344,
|
| 551 |
-
"learning_rate": 4.445729390384934e-06,
|
| 552 |
-
"loss": 1.2830023193359374,
|
| 553 |
-
"step": 3750
|
| 554 |
-
},
|
| 555 |
-
{
|
| 556 |
-
"epoch": 2.950310559006211,
|
| 557 |
-
"grad_norm": 11.397253036499023,
|
| 558 |
-
"learning_rate": 4.3631257227820915e-06,
|
| 559 |
-
"loss": 1.225728530883789,
|
| 560 |
-
"step": 3800
|
| 561 |
-
},
|
| 562 |
-
{
|
| 563 |
-
"epoch": 2.9891304347826084,
|
| 564 |
-
"grad_norm": 8.867121696472168,
|
| 565 |
-
"learning_rate": 4.28052205517925e-06,
|
| 566 |
-
"loss": 1.2054911041259766,
|
| 567 |
-
"step": 3850
|
| 568 |
-
},
|
| 569 |
-
{
|
| 570 |
-
"epoch": 3.0,
|
| 571 |
-
"eval_accuracy": 0.8048780487804879,
|
| 572 |
-
"eval_loss": 1.1265352964401245,
|
| 573 |
-
"eval_runtime": 9.2925,
|
| 574 |
-
"eval_samples_per_second": 260.317,
|
| 575 |
-
"eval_steps_per_second": 16.357,
|
| 576 |
-
"step": 3864
|
| 577 |
-
},
|
| 578 |
-
{
|
| 579 |
-
"epoch": 3.027950310559006,
|
| 580 |
-
"grad_norm": 15.09593677520752,
|
| 581 |
-
"learning_rate": 4.1979183875764084e-06,
|
| 582 |
-
"loss": 1.0494297790527343,
|
| 583 |
-
"step": 3900
|
| 584 |
-
},
|
| 585 |
-
{
|
| 586 |
-
"epoch": 3.0667701863354035,
|
| 587 |
-
"grad_norm": 7.0367751121521,
|
| 588 |
-
"learning_rate": 4.115314719973567e-06,
|
| 589 |
-
"loss": 1.195704803466797,
|
| 590 |
-
"step": 3950
|
| 591 |
-
},
|
| 592 |
-
{
|
| 593 |
-
"epoch": 3.1055900621118013,
|
| 594 |
-
"grad_norm": 6.594099521636963,
|
| 595 |
-
"learning_rate": 4.032711052370726e-06,
|
| 596 |
-
"loss": 1.1527919006347656,
|
| 597 |
-
"step": 4000
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 3.1444099378881987,
|
| 601 |
-
"grad_norm": 11.393077850341797,
|
| 602 |
-
"learning_rate": 3.950107384767884e-06,
|
| 603 |
-
"loss": 1.009382858276367,
|
| 604 |
-
"step": 4050
|
| 605 |
-
},
|
| 606 |
-
{
|
| 607 |
-
"epoch": 3.1832298136645965,
|
| 608 |
-
"grad_norm": 10.961431503295898,
|
| 609 |
-
"learning_rate": 3.867503717165042e-06,
|
| 610 |
-
"loss": 1.045629119873047,
|
| 611 |
-
"step": 4100
|
| 612 |
-
},
|
| 613 |
-
{
|
| 614 |
-
"epoch": 3.222049689440994,
|
| 615 |
-
"grad_norm": 7.3205180168151855,
|
| 616 |
-
"learning_rate": 3.784900049562201e-06,
|
| 617 |
-
"loss": 1.0478419494628906,
|
| 618 |
-
"step": 4150
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 3.260869565217391,
|
| 622 |
-
"grad_norm": 8.874663352966309,
|
| 623 |
-
"learning_rate": 3.7022963819593593e-06,
|
| 624 |
-
"loss": 1.0866982269287109,
|
| 625 |
-
"step": 4200
|
| 626 |
-
},
|
| 627 |
-
{
|
| 628 |
-
"epoch": 3.299689440993789,
|
| 629 |
-
"grad_norm": 7.505184650421143,
|
| 630 |
-
"learning_rate": 3.6196927143565173e-06,
|
| 631 |
-
"loss": 0.9710902404785157,
|
| 632 |
-
"step": 4250
|
| 633 |
-
},
|
| 634 |
-
{
|
| 635 |
-
"epoch": 3.3385093167701863,
|
| 636 |
-
"grad_norm": 5.981348514556885,
|
| 637 |
-
"learning_rate": 3.537089046753676e-06,
|
| 638 |
-
"loss": 1.0303248596191406,
|
| 639 |
-
"step": 4300
|
| 640 |
-
},
|
| 641 |
-
{
|
| 642 |
-
"epoch": 3.377329192546584,
|
| 643 |
-
"grad_norm": 9.041892051696777,
|
| 644 |
-
"learning_rate": 3.4544853791508347e-06,
|
| 645 |
-
"loss": 0.9202809143066406,
|
| 646 |
-
"step": 4350
|
| 647 |
-
},
|
| 648 |
-
{
|
| 649 |
-
"epoch": 3.4161490683229814,
|
| 650 |
-
"grad_norm": 7.664228916168213,
|
| 651 |
-
"learning_rate": 3.3718817115479927e-06,
|
| 652 |
-
"loss": 0.9286480712890625,
|
| 653 |
-
"step": 4400
|
| 654 |
-
},
|
| 655 |
-
{
|
| 656 |
-
"epoch": 3.4549689440993787,
|
| 657 |
-
"grad_norm": 10.143519401550293,
|
| 658 |
-
"learning_rate": 3.2892780439451516e-06,
|
| 659 |
-
"loss": 1.0028343963623048,
|
| 660 |
-
"step": 4450
|
| 661 |
-
},
|
| 662 |
-
{
|
| 663 |
-
"epoch": 3.4937888198757765,
|
| 664 |
-
"grad_norm": 13.038041114807129,
|
| 665 |
-
"learning_rate": 3.2066743763423097e-06,
|
| 666 |
-
"loss": 1.0670313262939453,
|
| 667 |
-
"step": 4500
|
| 668 |
-
},
|
| 669 |
-
{
|
| 670 |
-
"epoch": 3.532608695652174,
|
| 671 |
-
"grad_norm": 9.287400245666504,
|
| 672 |
-
"learning_rate": 3.1240707087394686e-06,
|
| 673 |
-
"loss": 0.8689695739746094,
|
| 674 |
-
"step": 4550
|
| 675 |
-
},
|
| 676 |
-
{
|
| 677 |
-
"epoch": 3.571428571428571,
|
| 678 |
-
"grad_norm": 7.978451728820801,
|
| 679 |
-
"learning_rate": 3.0414670411366266e-06,
|
| 680 |
-
"loss": 1.0953305053710938,
|
| 681 |
-
"step": 4600
|
| 682 |
-
},
|
| 683 |
-
{
|
| 684 |
-
"epoch": 3.610248447204969,
|
| 685 |
-
"grad_norm": 8.50739860534668,
|
| 686 |
-
"learning_rate": 2.958863373533785e-06,
|
| 687 |
-
"loss": 1.0060964965820312,
|
| 688 |
-
"step": 4650
|
| 689 |
-
},
|
| 690 |
-
{
|
| 691 |
-
"epoch": 3.6490683229813663,
|
| 692 |
-
"grad_norm": 8.6639986038208,
|
| 693 |
-
"learning_rate": 2.876259705930944e-06,
|
| 694 |
-
"loss": 0.9362821960449219,
|
| 695 |
-
"step": 4700
|
| 696 |
-
},
|
| 697 |
-
{
|
| 698 |
-
"epoch": 3.687888198757764,
|
| 699 |
-
"grad_norm": 9.980217933654785,
|
| 700 |
-
"learning_rate": 2.793656038328102e-06,
|
| 701 |
-
"loss": 0.9748807525634766,
|
| 702 |
-
"step": 4750
|
| 703 |
-
},
|
| 704 |
-
{
|
| 705 |
-
"epoch": 3.7267080745341614,
|
| 706 |
-
"grad_norm": 8.53567123413086,
|
| 707 |
-
"learning_rate": 2.7110523707252605e-06,
|
| 708 |
-
"loss": 0.9750958251953125,
|
| 709 |
-
"step": 4800
|
| 710 |
-
},
|
| 711 |
-
{
|
| 712 |
-
"epoch": 3.7655279503105588,
|
| 713 |
-
"grad_norm": 9.582347869873047,
|
| 714 |
-
"learning_rate": 2.6284487031224186e-06,
|
| 715 |
-
"loss": 0.9185539245605469,
|
| 716 |
-
"step": 4850
|
| 717 |
-
},
|
| 718 |
-
{
|
| 719 |
-
"epoch": 3.8043478260869565,
|
| 720 |
-
"grad_norm": 12.004874229431152,
|
| 721 |
-
"learning_rate": 2.5458450355195775e-06,
|
| 722 |
-
"loss": 0.9237863159179688,
|
| 723 |
-
"step": 4900
|
| 724 |
-
},
|
| 725 |
-
{
|
| 726 |
-
"epoch": 3.843167701863354,
|
| 727 |
-
"grad_norm": 13.144355773925781,
|
| 728 |
-
"learning_rate": 2.463241367916736e-06,
|
| 729 |
-
"loss": 0.9831585693359375,
|
| 730 |
-
"step": 4950
|
| 731 |
-
},
|
| 732 |
-
{
|
| 733 |
-
"epoch": 3.8819875776397517,
|
| 734 |
-
"grad_norm": 8.973755836486816,
|
| 735 |
-
"learning_rate": 2.380637700313894e-06,
|
| 736 |
-
"loss": 1.0324047088623047,
|
| 737 |
-
"step": 5000
|
| 738 |
-
},
|
| 739 |
-
{
|
| 740 |
-
"epoch": 3.920807453416149,
|
| 741 |
-
"grad_norm": 9.93669319152832,
|
| 742 |
-
"learning_rate": 2.2980340327110525e-06,
|
| 743 |
-
"loss": 0.9055207061767578,
|
| 744 |
-
"step": 5050
|
| 745 |
-
},
|
| 746 |
-
{
|
| 747 |
-
"epoch": 3.9596273291925463,
|
| 748 |
-
"grad_norm": 4.113827705383301,
|
| 749 |
-
"learning_rate": 2.215430365108211e-06,
|
| 750 |
-
"loss": 0.8888311004638672,
|
| 751 |
-
"step": 5100
|
| 752 |
-
},
|
| 753 |
-
{
|
| 754 |
-
"epoch": 3.998447204968944,
|
| 755 |
-
"grad_norm": 14.113999366760254,
|
| 756 |
-
"learning_rate": 2.1328266975053694e-06,
|
| 757 |
-
"loss": 0.9598866271972656,
|
| 758 |
-
"step": 5150
|
| 759 |
-
},
|
| 760 |
-
{
|
| 761 |
-
"epoch": 4.0,
|
| 762 |
-
"eval_accuracy": 0.8267879288962381,
|
| 763 |
-
"eval_loss": 0.9564170241355896,
|
| 764 |
-
"eval_runtime": 9.318,
|
| 765 |
-
"eval_samples_per_second": 259.605,
|
| 766 |
-
"eval_steps_per_second": 16.312,
|
| 767 |
-
"step": 5152
|
| 768 |
-
},
|
| 769 |
-
{
|
| 770 |
-
"epoch": 4.037267080745342,
|
| 771 |
-
"grad_norm": 9.910421371459961,
|
| 772 |
-
"learning_rate": 2.050223029902528e-06,
|
| 773 |
-
"loss": 0.8574532318115234,
|
| 774 |
-
"step": 5200
|
| 775 |
-
},
|
| 776 |
-
{
|
| 777 |
-
"epoch": 4.076086956521739,
|
| 778 |
-
"grad_norm": 13.372318267822266,
|
| 779 |
-
"learning_rate": 1.9676193622996863e-06,
|
| 780 |
-
"loss": 0.86721923828125,
|
| 781 |
-
"step": 5250
|
| 782 |
-
},
|
| 783 |
-
{
|
| 784 |
-
"epoch": 4.114906832298137,
|
| 785 |
-
"grad_norm": 4.227599620819092,
|
| 786 |
-
"learning_rate": 1.8850156946968448e-06,
|
| 787 |
-
"loss": 0.7417903137207031,
|
| 788 |
-
"step": 5300
|
| 789 |
-
},
|
| 790 |
-
{
|
| 791 |
-
"epoch": 4.153726708074534,
|
| 792 |
-
"grad_norm": 7.521543979644775,
|
| 793 |
-
"learning_rate": 1.8024120270940033e-06,
|
| 794 |
-
"loss": 0.7657796478271485,
|
| 795 |
-
"step": 5350
|
| 796 |
-
},
|
| 797 |
-
{
|
| 798 |
-
"epoch": 4.192546583850931,
|
| 799 |
-
"grad_norm": 8.862835884094238,
|
| 800 |
-
"learning_rate": 1.7198083594911616e-06,
|
| 801 |
-
"loss": 0.7435103607177734,
|
| 802 |
-
"step": 5400
|
| 803 |
-
},
|
| 804 |
-
{
|
| 805 |
-
"epoch": 4.2313664596273295,
|
| 806 |
-
"grad_norm": 7.360652923583984,
|
| 807 |
-
"learning_rate": 1.63720469188832e-06,
|
| 808 |
-
"loss": 0.8516057586669922,
|
| 809 |
-
"step": 5450
|
| 810 |
-
},
|
| 811 |
-
{
|
| 812 |
-
"epoch": 4.270186335403727,
|
| 813 |
-
"grad_norm": 7.56041955947876,
|
| 814 |
-
"learning_rate": 1.5546010242854783e-06,
|
| 815 |
-
"loss": 0.8586593627929687,
|
| 816 |
-
"step": 5500
|
| 817 |
-
},
|
| 818 |
-
{
|
| 819 |
-
"epoch": 4.309006211180124,
|
| 820 |
-
"grad_norm": 10.998615264892578,
|
| 821 |
-
"learning_rate": 1.4719973566826368e-06,
|
| 822 |
-
"loss": 0.871690902709961,
|
| 823 |
-
"step": 5550
|
| 824 |
-
},
|
| 825 |
-
{
|
| 826 |
-
"epoch": 4.3478260869565215,
|
| 827 |
-
"grad_norm": 7.930166244506836,
|
| 828 |
-
"learning_rate": 1.3893936890797952e-06,
|
| 829 |
-
"loss": 0.8314385986328126,
|
| 830 |
-
"step": 5600
|
| 831 |
-
},
|
| 832 |
-
{
|
| 833 |
-
"epoch": 4.386645962732919,
|
| 834 |
-
"grad_norm": 12.592620849609375,
|
| 835 |
-
"learning_rate": 1.3067900214769535e-06,
|
| 836 |
-
"loss": 0.8933805847167968,
|
| 837 |
-
"step": 5650
|
| 838 |
-
},
|
| 839 |
-
{
|
| 840 |
-
"epoch": 4.425465838509317,
|
| 841 |
-
"grad_norm": 9.266133308410645,
|
| 842 |
-
"learning_rate": 1.2241863538741122e-06,
|
| 843 |
-
"loss": 0.8784015655517579,
|
| 844 |
-
"step": 5700
|
| 845 |
-
},
|
| 846 |
-
{
|
| 847 |
-
"epoch": 4.464285714285714,
|
| 848 |
-
"grad_norm": 11.813312530517578,
|
| 849 |
-
"learning_rate": 1.1415826862712704e-06,
|
| 850 |
-
"loss": 0.8659978485107422,
|
| 851 |
-
"step": 5750
|
| 852 |
-
},
|
| 853 |
-
{
|
| 854 |
-
"epoch": 4.503105590062112,
|
| 855 |
-
"grad_norm": 7.200871467590332,
|
| 856 |
-
"learning_rate": 1.058979018668429e-06,
|
| 857 |
-
"loss": 0.7724305725097657,
|
| 858 |
-
"step": 5800
|
| 859 |
-
},
|
| 860 |
-
{
|
| 861 |
-
"epoch": 4.541925465838509,
|
| 862 |
-
"grad_norm": 6.725843906402588,
|
| 863 |
-
"learning_rate": 9.763753510655874e-07,
|
| 864 |
-
"loss": 0.74711669921875,
|
| 865 |
-
"step": 5850
|
| 866 |
-
},
|
| 867 |
-
{
|
| 868 |
-
"epoch": 4.580745341614907,
|
| 869 |
-
"grad_norm": 6.0406694412231445,
|
| 870 |
-
"learning_rate": 8.937716834627458e-07,
|
| 871 |
-
"loss": 0.8589053344726563,
|
| 872 |
-
"step": 5900
|
| 873 |
-
},
|
| 874 |
-
{
|
| 875 |
-
"epoch": 4.619565217391305,
|
| 876 |
-
"grad_norm": 6.727048397064209,
|
| 877 |
-
"learning_rate": 8.111680158599042e-07,
|
| 878 |
-
"loss": 0.8026133728027344,
|
| 879 |
-
"step": 5950
|
| 880 |
-
},
|
| 881 |
-
{
|
| 882 |
-
"epoch": 4.658385093167702,
|
| 883 |
-
"grad_norm": 5.949310302734375,
|
| 884 |
-
"learning_rate": 7.285643482570627e-07,
|
| 885 |
-
"loss": 0.7926577758789063,
|
| 886 |
-
"step": 6000
|
| 887 |
-
},
|
| 888 |
-
{
|
| 889 |
-
"epoch": 4.697204968944099,
|
| 890 |
-
"grad_norm": 14.107566833496094,
|
| 891 |
-
"learning_rate": 6.459606806542211e-07,
|
| 892 |
-
"loss": 0.9208444213867187,
|
| 893 |
-
"step": 6050
|
| 894 |
-
},
|
| 895 |
-
{
|
| 896 |
-
"epoch": 4.736024844720497,
|
| 897 |
-
"grad_norm": 7.512182712554932,
|
| 898 |
-
"learning_rate": 5.633570130513795e-07,
|
| 899 |
-
"loss": 0.740212173461914,
|
| 900 |
-
"step": 6100
|
| 901 |
-
},
|
| 902 |
-
{
|
| 903 |
-
"epoch": 4.774844720496894,
|
| 904 |
-
"grad_norm": 5.291878700256348,
|
| 905 |
-
"learning_rate": 4.807533454485379e-07,
|
| 906 |
-
"loss": 0.8455804443359375,
|
| 907 |
-
"step": 6150
|
| 908 |
-
},
|
| 909 |
-
{
|
| 910 |
-
"epoch": 4.813664596273292,
|
| 911 |
-
"grad_norm": 11.112459182739258,
|
| 912 |
-
"learning_rate": 3.9814967784569643e-07,
|
| 913 |
-
"loss": 0.77959716796875,
|
| 914 |
-
"step": 6200
|
| 915 |
-
},
|
| 916 |
-
{
|
| 917 |
-
"epoch": 4.85248447204969,
|
| 918 |
-
"grad_norm": 2.925304651260376,
|
| 919 |
-
"learning_rate": 3.155460102428548e-07,
|
| 920 |
-
"loss": 0.7745262145996094,
|
| 921 |
-
"step": 6250
|
| 922 |
-
},
|
| 923 |
-
{
|
| 924 |
-
"epoch": 4.891304347826087,
|
| 925 |
-
"grad_norm": 10.317737579345703,
|
| 926 |
-
"learning_rate": 2.3294234264001324e-07,
|
| 927 |
-
"loss": 0.9354353332519532,
|
| 928 |
-
"step": 6300
|
| 929 |
-
},
|
| 930 |
-
{
|
| 931 |
-
"epoch": 4.930124223602484,
|
| 932 |
-
"grad_norm": 12.665268898010254,
|
| 933 |
-
"learning_rate": 1.5033867503717166e-07,
|
| 934 |
-
"loss": 0.7373665618896484,
|
| 935 |
-
"step": 6350
|
| 936 |
-
},
|
| 937 |
-
{
|
| 938 |
-
"epoch": 4.9689440993788825,
|
| 939 |
-
"grad_norm": 12.559833526611328,
|
| 940 |
-
"learning_rate": 6.77350074343301e-08,
|
| 941 |
-
"loss": 0.8966941833496094,
|
| 942 |
-
"step": 6400
|
| 943 |
-
},
|
| 944 |
-
{
|
| 945 |
-
"epoch": 5.0,
|
| 946 |
-
"eval_accuracy": 0.836709384042993,
|
| 947 |
-
"eval_loss": 0.9067481756210327,
|
| 948 |
-
"eval_runtime": 9.2389,
|
| 949 |
-
"eval_samples_per_second": 261.828,
|
| 950 |
-
"eval_steps_per_second": 16.452,
|
| 951 |
-
"step": 6440
|
| 952 |
-
}
|
| 953 |
-
],
|
| 954 |
-
"logging_steps": 50,
|
| 955 |
-
"max_steps": 6440,
|
| 956 |
-
"num_input_tokens_seen": 0,
|
| 957 |
-
"num_train_epochs": 5,
|
| 958 |
-
"save_steps": 500,
|
| 959 |
-
"stateful_callbacks": {
|
| 960 |
-
"TrainerControl": {
|
| 961 |
-
"args": {
|
| 962 |
-
"should_epoch_stop": false,
|
| 963 |
-
"should_evaluate": false,
|
| 964 |
-
"should_log": false,
|
| 965 |
-
"should_save": true,
|
| 966 |
-
"should_training_stop": true
|
| 967 |
-
},
|
| 968 |
-
"attributes": {}
|
| 969 |
-
}
|
| 970 |
-
},
|
| 971 |
-
"total_flos": 0.0,
|
| 972 |
-
"train_batch_size": 16,
|
| 973 |
-
"trial_name": null,
|
| 974 |
-
"trial_params": null
|
| 975 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-6440/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:74437768892f28df5eb9cd67dc81c630774091f4cf430ac604b093bf8b442dbd
|
| 3 |
-
size 5201
|
|
|
|
|
|
|
|
|
|
|
|