Upload ./trainer_state.json with huggingface_hub
#7
by
datastreams
- opened
- trainer_state.json +319 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,319 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 6.756756756756757,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.14,
|
| 13 |
+
"learning_rate": 4.996861268047709e-05,
|
| 14 |
+
"loss": 2.8343,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.27,
|
| 19 |
+
"learning_rate": 4.9937225360954175e-05,
|
| 20 |
+
"loss": 2.6082,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.41,
|
| 25 |
+
"learning_rate": 4.9905838041431265e-05,
|
| 26 |
+
"loss": 2.7027,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.54,
|
| 31 |
+
"learning_rate": 4.987445072190835e-05,
|
| 32 |
+
"loss": 2.5134,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.68,
|
| 37 |
+
"learning_rate": 4.984306340238544e-05,
|
| 38 |
+
"loss": 2.413,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.81,
|
| 43 |
+
"learning_rate": 4.981167608286253e-05,
|
| 44 |
+
"loss": 2.4923,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.95,
|
| 49 |
+
"learning_rate": 4.978028876333961e-05,
|
| 50 |
+
"loss": 2.5383,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 1.08,
|
| 55 |
+
"learning_rate": 4.97489014438167e-05,
|
| 56 |
+
"loss": 2.1952,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 1.22,
|
| 61 |
+
"learning_rate": 4.971751412429379e-05,
|
| 62 |
+
"loss": 2.3365,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.35,
|
| 67 |
+
"learning_rate": 4.968612680477088e-05,
|
| 68 |
+
"loss": 2.2645,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 1.49,
|
| 73 |
+
"learning_rate": 4.965473948524796e-05,
|
| 74 |
+
"loss": 2.2224,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 1.62,
|
| 79 |
+
"learning_rate": 4.962335216572505e-05,
|
| 80 |
+
"loss": 2.1858,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.76,
|
| 85 |
+
"learning_rate": 4.959196484620214e-05,
|
| 86 |
+
"loss": 2.01,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.89,
|
| 91 |
+
"learning_rate": 4.956057752667922e-05,
|
| 92 |
+
"loss": 2.0707,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 2.03,
|
| 97 |
+
"learning_rate": 4.952919020715631e-05,
|
| 98 |
+
"loss": 2.1271,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 2.16,
|
| 103 |
+
"learning_rate": 4.9497802887633396e-05,
|
| 104 |
+
"loss": 2.1912,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 2.3,
|
| 109 |
+
"learning_rate": 4.9466415568110485e-05,
|
| 110 |
+
"loss": 1.9693,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 2.43,
|
| 115 |
+
"learning_rate": 4.9435028248587575e-05,
|
| 116 |
+
"loss": 1.7561,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 2.57,
|
| 121 |
+
"learning_rate": 4.940364092906466e-05,
|
| 122 |
+
"loss": 2.1115,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.7,
|
| 127 |
+
"learning_rate": 4.937225360954175e-05,
|
| 128 |
+
"loss": 1.9374,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 2.84,
|
| 133 |
+
"learning_rate": 4.934086629001883e-05,
|
| 134 |
+
"loss": 1.8973,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 2.97,
|
| 139 |
+
"learning_rate": 4.930947897049592e-05,
|
| 140 |
+
"loss": 1.9069,
|
| 141 |
+
"step": 220
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 3.11,
|
| 145 |
+
"learning_rate": 4.927809165097301e-05,
|
| 146 |
+
"loss": 1.8768,
|
| 147 |
+
"step": 230
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 3.24,
|
| 151 |
+
"learning_rate": 4.924670433145009e-05,
|
| 152 |
+
"loss": 1.7559,
|
| 153 |
+
"step": 240
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 3.38,
|
| 157 |
+
"learning_rate": 4.921531701192718e-05,
|
| 158 |
+
"loss": 1.6955,
|
| 159 |
+
"step": 250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 3.51,
|
| 163 |
+
"learning_rate": 4.918392969240427e-05,
|
| 164 |
+
"loss": 1.4821,
|
| 165 |
+
"step": 260
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.65,
|
| 169 |
+
"learning_rate": 4.915254237288136e-05,
|
| 170 |
+
"loss": 1.7175,
|
| 171 |
+
"step": 270
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 3.78,
|
| 175 |
+
"learning_rate": 4.9121155053358444e-05,
|
| 176 |
+
"loss": 1.6685,
|
| 177 |
+
"step": 280
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 3.92,
|
| 181 |
+
"learning_rate": 4.9089767733835534e-05,
|
| 182 |
+
"loss": 1.6313,
|
| 183 |
+
"step": 290
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 4.05,
|
| 187 |
+
"learning_rate": 4.9058380414312623e-05,
|
| 188 |
+
"loss": 1.6231,
|
| 189 |
+
"step": 300
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 4.19,
|
| 193 |
+
"learning_rate": 4.9026993094789706e-05,
|
| 194 |
+
"loss": 1.4119,
|
| 195 |
+
"step": 310
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 4.32,
|
| 199 |
+
"learning_rate": 4.8995605775266796e-05,
|
| 200 |
+
"loss": 1.2547,
|
| 201 |
+
"step": 320
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 4.46,
|
| 205 |
+
"learning_rate": 4.896421845574388e-05,
|
| 206 |
+
"loss": 1.3748,
|
| 207 |
+
"step": 330
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 4.59,
|
| 211 |
+
"learning_rate": 4.893283113622097e-05,
|
| 212 |
+
"loss": 1.3336,
|
| 213 |
+
"step": 340
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 4.73,
|
| 217 |
+
"learning_rate": 4.890144381669806e-05,
|
| 218 |
+
"loss": 1.4334,
|
| 219 |
+
"step": 350
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 4.86,
|
| 223 |
+
"learning_rate": 4.887005649717514e-05,
|
| 224 |
+
"loss": 1.4155,
|
| 225 |
+
"step": 360
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 5.0,
|
| 229 |
+
"learning_rate": 4.883866917765223e-05,
|
| 230 |
+
"loss": 1.4832,
|
| 231 |
+
"step": 370
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 5.14,
|
| 235 |
+
"learning_rate": 4.8807281858129313e-05,
|
| 236 |
+
"loss": 0.9712,
|
| 237 |
+
"step": 380
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 5.27,
|
| 241 |
+
"learning_rate": 4.87758945386064e-05,
|
| 242 |
+
"loss": 1.1672,
|
| 243 |
+
"step": 390
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 5.41,
|
| 247 |
+
"learning_rate": 4.874450721908349e-05,
|
| 248 |
+
"loss": 1.338,
|
| 249 |
+
"step": 400
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 5.54,
|
| 253 |
+
"learning_rate": 4.8713119899560576e-05,
|
| 254 |
+
"loss": 1.0129,
|
| 255 |
+
"step": 410
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 5.68,
|
| 259 |
+
"learning_rate": 4.8681732580037665e-05,
|
| 260 |
+
"loss": 1.1344,
|
| 261 |
+
"step": 420
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 5.81,
|
| 265 |
+
"learning_rate": 4.8650345260514755e-05,
|
| 266 |
+
"loss": 1.139,
|
| 267 |
+
"step": 430
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 5.95,
|
| 271 |
+
"learning_rate": 4.8618957940991844e-05,
|
| 272 |
+
"loss": 1.1943,
|
| 273 |
+
"step": 440
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 6.08,
|
| 277 |
+
"learning_rate": 4.8587570621468934e-05,
|
| 278 |
+
"loss": 0.9281,
|
| 279 |
+
"step": 450
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 6.22,
|
| 283 |
+
"learning_rate": 4.855618330194602e-05,
|
| 284 |
+
"loss": 0.9121,
|
| 285 |
+
"step": 460
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 6.35,
|
| 289 |
+
"learning_rate": 4.8524795982423107e-05,
|
| 290 |
+
"loss": 1.0089,
|
| 291 |
+
"step": 470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 6.49,
|
| 295 |
+
"learning_rate": 4.849340866290019e-05,
|
| 296 |
+
"loss": 0.9739,
|
| 297 |
+
"step": 480
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 6.62,
|
| 301 |
+
"learning_rate": 4.846202134337728e-05,
|
| 302 |
+
"loss": 0.958,
|
| 303 |
+
"step": 490
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 6.76,
|
| 307 |
+
"learning_rate": 4.843063402385437e-05,
|
| 308 |
+
"loss": 0.7523,
|
| 309 |
+
"step": 500
|
| 310 |
+
}
|
| 311 |
+
],
|
| 312 |
+
"logging_steps": 10,
|
| 313 |
+
"max_steps": 15930,
|
| 314 |
+
"num_train_epochs": 216,
|
| 315 |
+
"save_steps": 500,
|
| 316 |
+
"total_flos": 2.821509238259712e+16,
|
| 317 |
+
"trial_name": null,
|
| 318 |
+
"trial_params": null
|
| 319 |
+
}
|