Upload config.json
Browse files- config.json +356 -0
config.json
ADDED
|
@@ -0,0 +1,356 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.4986101388931274,
|
| 3 |
+
"best_model_checkpoint": "BLOOM-alpaca",
|
| 4 |
+
"epoch": 2.5628704148646486,
|
| 5 |
+
"global_step": 1000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.05,
|
| 12 |
+
"learning_rate": 5.9999999999999995e-05,
|
| 13 |
+
"loss": 3.799,
|
| 14 |
+
"step": 20
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.1,
|
| 18 |
+
"learning_rate": 0.00011999999999999999,
|
| 19 |
+
"loss": 3.1924,
|
| 20 |
+
"step": 40
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.15,
|
| 24 |
+
"learning_rate": 0.00017999999999999998,
|
| 25 |
+
"loss": 2.1651,
|
| 26 |
+
"step": 60
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.21,
|
| 30 |
+
"learning_rate": 0.00023999999999999998,
|
| 31 |
+
"loss": 1.7068,
|
| 32 |
+
"step": 80
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.26,
|
| 36 |
+
"learning_rate": 0.0003,
|
| 37 |
+
"loss": 1.6352,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.31,
|
| 42 |
+
"learning_rate": 0.00029439252336448596,
|
| 43 |
+
"loss": 1.5979,
|
| 44 |
+
"step": 120
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.36,
|
| 48 |
+
"learning_rate": 0.00028878504672897194,
|
| 49 |
+
"loss": 1.5722,
|
| 50 |
+
"step": 140
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.41,
|
| 54 |
+
"learning_rate": 0.0002831775700934579,
|
| 55 |
+
"loss": 1.5735,
|
| 56 |
+
"step": 160
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.46,
|
| 60 |
+
"learning_rate": 0.0002775700934579439,
|
| 61 |
+
"loss": 1.5481,
|
| 62 |
+
"step": 180
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.51,
|
| 66 |
+
"learning_rate": 0.0002719626168224299,
|
| 67 |
+
"loss": 1.5288,
|
| 68 |
+
"step": 200
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.51,
|
| 72 |
+
"eval_loss": 1.5642188787460327,
|
| 73 |
+
"eval_runtime": 117.512,
|
| 74 |
+
"eval_samples_per_second": 17.02,
|
| 75 |
+
"eval_steps_per_second": 2.127,
|
| 76 |
+
"step": 200
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.56,
|
| 80 |
+
"learning_rate": 0.00026635514018691586,
|
| 81 |
+
"loss": 1.5394,
|
| 82 |
+
"step": 220
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 0.62,
|
| 86 |
+
"learning_rate": 0.00026074766355140184,
|
| 87 |
+
"loss": 1.5299,
|
| 88 |
+
"step": 240
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.67,
|
| 92 |
+
"learning_rate": 0.0002551401869158878,
|
| 93 |
+
"loss": 1.5137,
|
| 94 |
+
"step": 260
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.72,
|
| 98 |
+
"learning_rate": 0.0002495327102803738,
|
| 99 |
+
"loss": 1.5199,
|
| 100 |
+
"step": 280
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.77,
|
| 104 |
+
"learning_rate": 0.0002439252336448598,
|
| 105 |
+
"loss": 1.5171,
|
| 106 |
+
"step": 300
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.82,
|
| 110 |
+
"learning_rate": 0.00023831775700934577,
|
| 111 |
+
"loss": 1.5143,
|
| 112 |
+
"step": 320
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.87,
|
| 116 |
+
"learning_rate": 0.00023271028037383175,
|
| 117 |
+
"loss": 1.5053,
|
| 118 |
+
"step": 340
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.92,
|
| 122 |
+
"learning_rate": 0.00022710280373831773,
|
| 123 |
+
"loss": 1.5078,
|
| 124 |
+
"step": 360
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.97,
|
| 128 |
+
"learning_rate": 0.0002214953271028037,
|
| 129 |
+
"loss": 1.5003,
|
| 130 |
+
"step": 380
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.03,
|
| 134 |
+
"learning_rate": 0.0002158878504672897,
|
| 135 |
+
"loss": 1.5004,
|
| 136 |
+
"step": 400
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.03,
|
| 140 |
+
"eval_loss": 1.5298347473144531,
|
| 141 |
+
"eval_runtime": 117.6931,
|
| 142 |
+
"eval_samples_per_second": 16.993,
|
| 143 |
+
"eval_steps_per_second": 2.124,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.08,
|
| 148 |
+
"learning_rate": 0.00021028037383177567,
|
| 149 |
+
"loss": 1.4853,
|
| 150 |
+
"step": 420
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.13,
|
| 154 |
+
"learning_rate": 0.00020467289719626166,
|
| 155 |
+
"loss": 1.511,
|
| 156 |
+
"step": 440
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.18,
|
| 160 |
+
"learning_rate": 0.00019906542056074764,
|
| 161 |
+
"loss": 1.4976,
|
| 162 |
+
"step": 460
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.23,
|
| 166 |
+
"learning_rate": 0.00019345794392523362,
|
| 167 |
+
"loss": 1.4804,
|
| 168 |
+
"step": 480
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.28,
|
| 172 |
+
"learning_rate": 0.0001878504672897196,
|
| 173 |
+
"loss": 1.4795,
|
| 174 |
+
"step": 500
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 1.33,
|
| 178 |
+
"learning_rate": 0.00018224299065420558,
|
| 179 |
+
"loss": 1.4876,
|
| 180 |
+
"step": 520
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 1.38,
|
| 184 |
+
"learning_rate": 0.00017663551401869156,
|
| 185 |
+
"loss": 1.4791,
|
| 186 |
+
"step": 540
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.44,
|
| 190 |
+
"learning_rate": 0.00017102803738317754,
|
| 191 |
+
"loss": 1.4903,
|
| 192 |
+
"step": 560
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.49,
|
| 196 |
+
"learning_rate": 0.00016542056074766352,
|
| 197 |
+
"loss": 1.4817,
|
| 198 |
+
"step": 580
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.54,
|
| 202 |
+
"learning_rate": 0.0001598130841121495,
|
| 203 |
+
"loss": 1.5033,
|
| 204 |
+
"step": 600
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.54,
|
| 208 |
+
"eval_loss": 1.513887882232666,
|
| 209 |
+
"eval_runtime": 118.2404,
|
| 210 |
+
"eval_samples_per_second": 16.915,
|
| 211 |
+
"eval_steps_per_second": 2.114,
|
| 212 |
+
"step": 600
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.59,
|
| 216 |
+
"learning_rate": 0.0001542056074766355,
|
| 217 |
+
"loss": 1.4738,
|
| 218 |
+
"step": 620
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.64,
|
| 222 |
+
"learning_rate": 0.00014859813084112147,
|
| 223 |
+
"loss": 1.4903,
|
| 224 |
+
"step": 640
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.69,
|
| 228 |
+
"learning_rate": 0.00014299065420560745,
|
| 229 |
+
"loss": 1.4982,
|
| 230 |
+
"step": 660
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.74,
|
| 234 |
+
"learning_rate": 0.00013738317757009343,
|
| 235 |
+
"loss": 1.4773,
|
| 236 |
+
"step": 680
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.79,
|
| 240 |
+
"learning_rate": 0.0001317757009345794,
|
| 241 |
+
"loss": 1.4663,
|
| 242 |
+
"step": 700
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.85,
|
| 246 |
+
"learning_rate": 0.0001261682242990654,
|
| 247 |
+
"loss": 1.478,
|
| 248 |
+
"step": 720
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.9,
|
| 252 |
+
"learning_rate": 0.00012056074766355139,
|
| 253 |
+
"loss": 1.4548,
|
| 254 |
+
"step": 740
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.95,
|
| 258 |
+
"learning_rate": 0.00011495327102803737,
|
| 259 |
+
"loss": 1.4928,
|
| 260 |
+
"step": 760
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.0,
|
| 264 |
+
"learning_rate": 0.00010934579439252335,
|
| 265 |
+
"loss": 1.4769,
|
| 266 |
+
"step": 780
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.05,
|
| 270 |
+
"learning_rate": 0.00010373831775700933,
|
| 271 |
+
"loss": 1.4553,
|
| 272 |
+
"step": 800
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.05,
|
| 276 |
+
"eval_loss": 1.5058649778366089,
|
| 277 |
+
"eval_runtime": 117.6034,
|
| 278 |
+
"eval_samples_per_second": 17.006,
|
| 279 |
+
"eval_steps_per_second": 2.126,
|
| 280 |
+
"step": 800
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 2.1,
|
| 284 |
+
"learning_rate": 9.813084112149531e-05,
|
| 285 |
+
"loss": 1.4699,
|
| 286 |
+
"step": 820
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 2.15,
|
| 290 |
+
"learning_rate": 9.25233644859813e-05,
|
| 291 |
+
"loss": 1.4596,
|
| 292 |
+
"step": 840
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 2.2,
|
| 296 |
+
"learning_rate": 8.691588785046728e-05,
|
| 297 |
+
"loss": 1.4786,
|
| 298 |
+
"step": 860
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 2.26,
|
| 302 |
+
"learning_rate": 8.130841121495326e-05,
|
| 303 |
+
"loss": 1.4701,
|
| 304 |
+
"step": 880
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 2.31,
|
| 308 |
+
"learning_rate": 7.570093457943924e-05,
|
| 309 |
+
"loss": 1.4608,
|
| 310 |
+
"step": 900
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 2.36,
|
| 314 |
+
"learning_rate": 7.009345794392522e-05,
|
| 315 |
+
"loss": 1.47,
|
| 316 |
+
"step": 920
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 2.41,
|
| 320 |
+
"learning_rate": 6.44859813084112e-05,
|
| 321 |
+
"loss": 1.4593,
|
| 322 |
+
"step": 940
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 2.46,
|
| 326 |
+
"learning_rate": 5.887850467289719e-05,
|
| 327 |
+
"loss": 1.4566,
|
| 328 |
+
"step": 960
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.51,
|
| 332 |
+
"learning_rate": 5.327102803738317e-05,
|
| 333 |
+
"loss": 1.4624,
|
| 334 |
+
"step": 980
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 2.56,
|
| 338 |
+
"learning_rate": 4.766355140186915e-05,
|
| 339 |
+
"loss": 1.4484,
|
| 340 |
+
"step": 1000
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.56,
|
| 344 |
+
"eval_loss": 1.4986101388931274,
|
| 345 |
+
"eval_runtime": 117.7967,
|
| 346 |
+
"eval_samples_per_second": 16.978,
|
| 347 |
+
"eval_steps_per_second": 2.122,
|
| 348 |
+
"step": 1000
|
| 349 |
+
}
|
| 350 |
+
],
|
| 351 |
+
"max_steps": 1170,
|
| 352 |
+
"num_train_epochs": 3,
|
| 353 |
+
"total_flos": 5.959000367982182e+16,
|
| 354 |
+
"trial_name": null,
|
| 355 |
+
"trial_params": null
|
| 356 |
+
}
|