Delete trainer_state.json
Browse files- trainer_state.json +0 -1282
trainer_state.json
DELETED
|
@@ -1,1282 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.9984,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 78,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.0128,
|
| 14 |
-
"grad_norm": 8.806152221916586,
|
| 15 |
-
"learning_rate": 1.25e-07,
|
| 16 |
-
"logits/chosen": -0.603271484375,
|
| 17 |
-
"logits/rejected": -0.609130859375,
|
| 18 |
-
"logps/chosen": -0.79052734375,
|
| 19 |
-
"logps/rejected": -0.80419921875,
|
| 20 |
-
"loss": 1.5099,
|
| 21 |
-
"nll_loss": 0.76611328125,
|
| 22 |
-
"rewards/accuracies": 0.5078125,
|
| 23 |
-
"rewards/chosen": -0.157958984375,
|
| 24 |
-
"rewards/margins": 0.0026302337646484375,
|
| 25 |
-
"rewards/rejected": -0.16064453125,
|
| 26 |
-
"step": 1
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.0256,
|
| 30 |
-
"grad_norm": 10.295030998935504,
|
| 31 |
-
"learning_rate": 2.5e-07,
|
| 32 |
-
"logits/chosen": -0.54736328125,
|
| 33 |
-
"logits/rejected": -0.553466796875,
|
| 34 |
-
"logps/chosen": -0.86181640625,
|
| 35 |
-
"logps/rejected": -0.8994140625,
|
| 36 |
-
"loss": 1.5817,
|
| 37 |
-
"nll_loss": 0.8330078125,
|
| 38 |
-
"rewards/accuracies": 0.5234375,
|
| 39 |
-
"rewards/chosen": -0.1724853515625,
|
| 40 |
-
"rewards/margins": 0.00754547119140625,
|
| 41 |
-
"rewards/rejected": -0.1798095703125,
|
| 42 |
-
"step": 2
|
| 43 |
-
},
|
| 44 |
-
{
|
| 45 |
-
"epoch": 0.0384,
|
| 46 |
-
"grad_norm": 9.606994036202016,
|
| 47 |
-
"learning_rate": 3.75e-07,
|
| 48 |
-
"logits/chosen": -0.573486328125,
|
| 49 |
-
"logits/rejected": -0.5751953125,
|
| 50 |
-
"logps/chosen": -0.7890625,
|
| 51 |
-
"logps/rejected": -0.85205078125,
|
| 52 |
-
"loss": 1.5178,
|
| 53 |
-
"nll_loss": 0.77392578125,
|
| 54 |
-
"rewards/accuracies": 0.5234375,
|
| 55 |
-
"rewards/chosen": -0.1578369140625,
|
| 56 |
-
"rewards/margins": 0.01264190673828125,
|
| 57 |
-
"rewards/rejected": -0.170654296875,
|
| 58 |
-
"step": 3
|
| 59 |
-
},
|
| 60 |
-
{
|
| 61 |
-
"epoch": 0.0512,
|
| 62 |
-
"grad_norm": 9.199710542852184,
|
| 63 |
-
"learning_rate": 5e-07,
|
| 64 |
-
"logits/chosen": -0.542724609375,
|
| 65 |
-
"logits/rejected": -0.5205078125,
|
| 66 |
-
"logps/chosen": -0.80712890625,
|
| 67 |
-
"logps/rejected": -0.83447265625,
|
| 68 |
-
"loss": 1.556,
|
| 69 |
-
"nll_loss": 0.80517578125,
|
| 70 |
-
"rewards/accuracies": 0.5,
|
| 71 |
-
"rewards/chosen": -0.1614990234375,
|
| 72 |
-
"rewards/margins": 0.005481719970703125,
|
| 73 |
-
"rewards/rejected": -0.1669921875,
|
| 74 |
-
"step": 4
|
| 75 |
-
},
|
| 76 |
-
{
|
| 77 |
-
"epoch": 0.064,
|
| 78 |
-
"grad_norm": 8.376911002240213,
|
| 79 |
-
"learning_rate": 6.249999999999999e-07,
|
| 80 |
-
"logits/chosen": -0.5986328125,
|
| 81 |
-
"logits/rejected": -0.60107421875,
|
| 82 |
-
"logps/chosen": -0.7783203125,
|
| 83 |
-
"logps/rejected": -0.81396484375,
|
| 84 |
-
"loss": 1.522,
|
| 85 |
-
"nll_loss": 0.77880859375,
|
| 86 |
-
"rewards/accuracies": 0.578125,
|
| 87 |
-
"rewards/chosen": -0.15576171875,
|
| 88 |
-
"rewards/margins": 0.00701141357421875,
|
| 89 |
-
"rewards/rejected": -0.16259765625,
|
| 90 |
-
"step": 5
|
| 91 |
-
},
|
| 92 |
-
{
|
| 93 |
-
"epoch": 0.0768,
|
| 94 |
-
"grad_norm": 8.510503674314895,
|
| 95 |
-
"learning_rate": 7.5e-07,
|
| 96 |
-
"logits/chosen": -0.583740234375,
|
| 97 |
-
"logits/rejected": -0.547119140625,
|
| 98 |
-
"logps/chosen": -0.80078125,
|
| 99 |
-
"logps/rejected": -0.82666015625,
|
| 100 |
-
"loss": 1.5448,
|
| 101 |
-
"nll_loss": 0.79638671875,
|
| 102 |
-
"rewards/accuracies": 0.4921875,
|
| 103 |
-
"rewards/chosen": -0.16015625,
|
| 104 |
-
"rewards/margins": 0.005126953125,
|
| 105 |
-
"rewards/rejected": -0.1650390625,
|
| 106 |
-
"step": 6
|
| 107 |
-
},
|
| 108 |
-
{
|
| 109 |
-
"epoch": 0.0896,
|
| 110 |
-
"grad_norm": 8.952623070299966,
|
| 111 |
-
"learning_rate": 8.75e-07,
|
| 112 |
-
"logits/chosen": -0.56787109375,
|
| 113 |
-
"logits/rejected": -0.58544921875,
|
| 114 |
-
"logps/chosen": -0.80615234375,
|
| 115 |
-
"logps/rejected": -0.8486328125,
|
| 116 |
-
"loss": 1.5365,
|
| 117 |
-
"nll_loss": 0.79150390625,
|
| 118 |
-
"rewards/accuracies": 0.578125,
|
| 119 |
-
"rewards/chosen": -0.1612548828125,
|
| 120 |
-
"rewards/margins": 0.008548736572265625,
|
| 121 |
-
"rewards/rejected": -0.1695556640625,
|
| 122 |
-
"step": 7
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.1024,
|
| 126 |
-
"grad_norm": 7.748818338200202,
|
| 127 |
-
"learning_rate": 1e-06,
|
| 128 |
-
"logits/chosen": -0.57763671875,
|
| 129 |
-
"logits/rejected": -0.548828125,
|
| 130 |
-
"logps/chosen": -0.77783203125,
|
| 131 |
-
"logps/rejected": -0.81201171875,
|
| 132 |
-
"loss": 1.5188,
|
| 133 |
-
"nll_loss": 0.7744140625,
|
| 134 |
-
"rewards/accuracies": 0.578125,
|
| 135 |
-
"rewards/chosen": -0.1556396484375,
|
| 136 |
-
"rewards/margins": 0.0068511962890625,
|
| 137 |
-
"rewards/rejected": -0.1622314453125,
|
| 138 |
-
"step": 8
|
| 139 |
-
},
|
| 140 |
-
{
|
| 141 |
-
"epoch": 0.1152,
|
| 142 |
-
"grad_norm": 8.056921633356916,
|
| 143 |
-
"learning_rate": 9.994965332706572e-07,
|
| 144 |
-
"logits/chosen": -0.57373046875,
|
| 145 |
-
"logits/rejected": -0.57373046875,
|
| 146 |
-
"logps/chosen": -0.8466796875,
|
| 147 |
-
"logps/rejected": -0.81640625,
|
| 148 |
-
"loss": 1.5742,
|
| 149 |
-
"nll_loss": 0.82666015625,
|
| 150 |
-
"rewards/accuracies": 0.5,
|
| 151 |
-
"rewards/chosen": -0.16943359375,
|
| 152 |
-
"rewards/margins": -0.006122589111328125,
|
| 153 |
-
"rewards/rejected": -0.163330078125,
|
| 154 |
-
"step": 9
|
| 155 |
-
},
|
| 156 |
-
{
|
| 157 |
-
"epoch": 0.128,
|
| 158 |
-
"grad_norm": 4.494952443924689,
|
| 159 |
-
"learning_rate": 9.979871469976195e-07,
|
| 160 |
-
"logits/chosen": -0.630859375,
|
| 161 |
-
"logits/rejected": -0.6259765625,
|
| 162 |
-
"logps/chosen": -0.7421875,
|
| 163 |
-
"logps/rejected": -0.73193359375,
|
| 164 |
-
"loss": 1.4967,
|
| 165 |
-
"nll_loss": 0.75048828125,
|
| 166 |
-
"rewards/accuracies": 0.453125,
|
| 167 |
-
"rewards/chosen": -0.1485595703125,
|
| 168 |
-
"rewards/margins": -0.002147674560546875,
|
| 169 |
-
"rewards/rejected": -0.1466064453125,
|
| 170 |
-
"step": 10
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.1408,
|
| 174 |
-
"grad_norm": 4.844771345193396,
|
| 175 |
-
"learning_rate": 9.954748808839674e-07,
|
| 176 |
-
"logits/chosen": -0.63134765625,
|
| 177 |
-
"logits/rejected": -0.626953125,
|
| 178 |
-
"logps/chosen": -0.78759765625,
|
| 179 |
-
"logps/rejected": -0.79541015625,
|
| 180 |
-
"loss": 1.5304,
|
| 181 |
-
"nll_loss": 0.791015625,
|
| 182 |
-
"rewards/accuracies": 0.4921875,
|
| 183 |
-
"rewards/chosen": -0.1575927734375,
|
| 184 |
-
"rewards/margins": 0.00148773193359375,
|
| 185 |
-
"rewards/rejected": -0.158935546875,
|
| 186 |
-
"step": 11
|
| 187 |
-
},
|
| 188 |
-
{
|
| 189 |
-
"epoch": 0.1536,
|
| 190 |
-
"grad_norm": 4.494757883851837,
|
| 191 |
-
"learning_rate": 9.919647942993147e-07,
|
| 192 |
-
"logits/chosen": -0.65283203125,
|
| 193 |
-
"logits/rejected": -0.64892578125,
|
| 194 |
-
"logps/chosen": -0.7685546875,
|
| 195 |
-
"logps/rejected": -0.74462890625,
|
| 196 |
-
"loss": 1.5063,
|
| 197 |
-
"nll_loss": 0.76025390625,
|
| 198 |
-
"rewards/accuracies": 0.4296875,
|
| 199 |
-
"rewards/chosen": -0.153564453125,
|
| 200 |
-
"rewards/margins": -0.0048351287841796875,
|
| 201 |
-
"rewards/rejected": -0.1488037109375,
|
| 202 |
-
"step": 12
|
| 203 |
-
},
|
| 204 |
-
{
|
| 205 |
-
"epoch": 0.1664,
|
| 206 |
-
"grad_norm": 4.166590855226656,
|
| 207 |
-
"learning_rate": 9.874639560909118e-07,
|
| 208 |
-
"logits/chosen": -0.65966796875,
|
| 209 |
-
"logits/rejected": -0.619140625,
|
| 210 |
-
"logps/chosen": -0.7275390625,
|
| 211 |
-
"logps/rejected": -0.74951171875,
|
| 212 |
-
"loss": 1.4634,
|
| 213 |
-
"nll_loss": 0.7216796875,
|
| 214 |
-
"rewards/accuracies": 0.5234375,
|
| 215 |
-
"rewards/chosen": -0.1453857421875,
|
| 216 |
-
"rewards/margins": 0.0044803619384765625,
|
| 217 |
-
"rewards/rejected": -0.14990234375,
|
| 218 |
-
"step": 13
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 0.1792,
|
| 222 |
-
"grad_norm": 3.92079299651966,
|
| 223 |
-
"learning_rate": 9.819814303479267e-07,
|
| 224 |
-
"logits/chosen": -0.78564453125,
|
| 225 |
-
"logits/rejected": -0.771484375,
|
| 226 |
-
"logps/chosen": -0.6826171875,
|
| 227 |
-
"logps/rejected": -0.7392578125,
|
| 228 |
-
"loss": 1.4136,
|
| 229 |
-
"nll_loss": 0.68017578125,
|
| 230 |
-
"rewards/accuracies": 0.4921875,
|
| 231 |
-
"rewards/chosen": -0.1365966796875,
|
| 232 |
-
"rewards/margins": 0.011296272277832031,
|
| 233 |
-
"rewards/rejected": -0.1475830078125,
|
| 234 |
-
"step": 14
|
| 235 |
-
},
|
| 236 |
-
{
|
| 237 |
-
"epoch": 0.192,
|
| 238 |
-
"grad_norm": 3.8602455794330903,
|
| 239 |
-
"learning_rate": 9.755282581475767e-07,
|
| 240 |
-
"logits/chosen": -0.81591796875,
|
| 241 |
-
"logits/rejected": -0.8095703125,
|
| 242 |
-
"logps/chosen": -0.6787109375,
|
| 243 |
-
"logps/rejected": -0.7314453125,
|
| 244 |
-
"loss": 1.4152,
|
| 245 |
-
"nll_loss": 0.67919921875,
|
| 246 |
-
"rewards/accuracies": 0.515625,
|
| 247 |
-
"rewards/chosen": -0.13555908203125,
|
| 248 |
-
"rewards/margins": 0.010591506958007812,
|
| 249 |
-
"rewards/rejected": -0.146240234375,
|
| 250 |
-
"step": 15
|
| 251 |
-
},
|
| 252 |
-
{
|
| 253 |
-
"epoch": 0.2048,
|
| 254 |
-
"grad_norm": 3.7707893130154932,
|
| 255 |
-
"learning_rate": 9.681174353198686e-07,
|
| 256 |
-
"logits/chosen": -0.833984375,
|
| 257 |
-
"logits/rejected": -0.81689453125,
|
| 258 |
-
"logps/chosen": -0.65380859375,
|
| 259 |
-
"logps/rejected": -0.71435546875,
|
| 260 |
-
"loss": 1.371,
|
| 261 |
-
"nll_loss": 0.638671875,
|
| 262 |
-
"rewards/accuracies": 0.59375,
|
| 263 |
-
"rewards/chosen": -0.1307373046875,
|
| 264 |
-
"rewards/margins": 0.012170791625976562,
|
| 265 |
-
"rewards/rejected": -0.1429443359375,
|
| 266 |
-
"step": 16
|
| 267 |
-
},
|
| 268 |
-
{
|
| 269 |
-
"epoch": 0.2176,
|
| 270 |
-
"grad_norm": 3.766035255465641,
|
| 271 |
-
"learning_rate": 9.597638862757253e-07,
|
| 272 |
-
"logits/chosen": -0.904296875,
|
| 273 |
-
"logits/rejected": -0.89111328125,
|
| 274 |
-
"logps/chosen": -0.67138671875,
|
| 275 |
-
"logps/rejected": -0.69287109375,
|
| 276 |
-
"loss": 1.4225,
|
| 277 |
-
"nll_loss": 0.68896484375,
|
| 278 |
-
"rewards/accuracies": 0.5390625,
|
| 279 |
-
"rewards/chosen": -0.13427734375,
|
| 280 |
-
"rewards/margins": 0.0042171478271484375,
|
| 281 |
-
"rewards/rejected": -0.138671875,
|
| 282 |
-
"step": 17
|
| 283 |
-
},
|
| 284 |
-
{
|
| 285 |
-
"epoch": 0.2304,
|
| 286 |
-
"grad_norm": 3.471968249039566,
|
| 287 |
-
"learning_rate": 9.504844339512094e-07,
|
| 288 |
-
"logits/chosen": -0.95751953125,
|
| 289 |
-
"logits/rejected": -0.95556640625,
|
| 290 |
-
"logps/chosen": -0.65478515625,
|
| 291 |
-
"logps/rejected": -0.67822265625,
|
| 292 |
-
"loss": 1.4049,
|
| 293 |
-
"nll_loss": 0.67041015625,
|
| 294 |
-
"rewards/accuracies": 0.5078125,
|
| 295 |
-
"rewards/chosen": -0.130859375,
|
| 296 |
-
"rewards/margins": 0.004756927490234375,
|
| 297 |
-
"rewards/rejected": -0.13555908203125,
|
| 298 |
-
"step": 18
|
| 299 |
-
},
|
| 300 |
-
{
|
| 301 |
-
"epoch": 0.2432,
|
| 302 |
-
"grad_norm": 3.624883189456873,
|
| 303 |
-
"learning_rate": 9.402977659283689e-07,
|
| 304 |
-
"logits/chosen": -1.06982421875,
|
| 305 |
-
"logits/rejected": -1.03466796875,
|
| 306 |
-
"logps/chosen": -0.658203125,
|
| 307 |
-
"logps/rejected": -0.6982421875,
|
| 308 |
-
"loss": 1.3997,
|
| 309 |
-
"nll_loss": 0.6669921875,
|
| 310 |
-
"rewards/accuracies": 0.5625,
|
| 311 |
-
"rewards/chosen": -0.131591796875,
|
| 312 |
-
"rewards/margins": 0.00794219970703125,
|
| 313 |
-
"rewards/rejected": -0.13958740234375,
|
| 314 |
-
"step": 19
|
| 315 |
-
},
|
| 316 |
-
{
|
| 317 |
-
"epoch": 0.256,
|
| 318 |
-
"grad_norm": 3.776625903303235,
|
| 319 |
-
"learning_rate": 9.29224396800933e-07,
|
| 320 |
-
"logits/chosen": -1.0859375,
|
| 321 |
-
"logits/rejected": -1.07421875,
|
| 322 |
-
"logps/chosen": -0.70361328125,
|
| 323 |
-
"logps/rejected": -0.70263671875,
|
| 324 |
-
"loss": 1.4501,
|
| 325 |
-
"nll_loss": 0.71630859375,
|
| 326 |
-
"rewards/accuracies": 0.4609375,
|
| 327 |
-
"rewards/chosen": -0.140869140625,
|
| 328 |
-
"rewards/margins": -0.0001373291015625,
|
| 329 |
-
"rewards/rejected": -0.140625,
|
| 330 |
-
"step": 20
|
| 331 |
-
},
|
| 332 |
-
{
|
| 333 |
-
"epoch": 0.2688,
|
| 334 |
-
"grad_norm": 3.9609518257799983,
|
| 335 |
-
"learning_rate": 9.172866268606513e-07,
|
| 336 |
-
"logits/chosen": -1.0908203125,
|
| 337 |
-
"logits/rejected": -1.08740234375,
|
| 338 |
-
"logps/chosen": -0.69873046875,
|
| 339 |
-
"logps/rejected": -0.72412109375,
|
| 340 |
-
"loss": 1.4291,
|
| 341 |
-
"nll_loss": 0.7021484375,
|
| 342 |
-
"rewards/accuracies": 0.46875,
|
| 343 |
-
"rewards/chosen": -0.139892578125,
|
| 344 |
-
"rewards/margins": 0.00504302978515625,
|
| 345 |
-
"rewards/rejected": -0.14483642578125,
|
| 346 |
-
"step": 21
|
| 347 |
-
},
|
| 348 |
-
{
|
| 349 |
-
"epoch": 0.2816,
|
| 350 |
-
"grad_norm": 4.302366208928238,
|
| 351 |
-
"learning_rate": 9.045084971874737e-07,
|
| 352 |
-
"logits/chosen": -1.30078125,
|
| 353 |
-
"logits/rejected": -1.294921875,
|
| 354 |
-
"logps/chosen": -0.6845703125,
|
| 355 |
-
"logps/rejected": -0.66650390625,
|
| 356 |
-
"loss": 1.4155,
|
| 357 |
-
"nll_loss": 0.677734375,
|
| 358 |
-
"rewards/accuracies": 0.5,
|
| 359 |
-
"rewards/chosen": -0.1368408203125,
|
| 360 |
-
"rewards/margins": -0.0035572052001953125,
|
| 361 |
-
"rewards/rejected": -0.13323974609375,
|
| 362 |
-
"step": 22
|
| 363 |
-
},
|
| 364 |
-
{
|
| 365 |
-
"epoch": 0.2944,
|
| 366 |
-
"grad_norm": 3.58562486403533,
|
| 367 |
-
"learning_rate": 8.909157412340149e-07,
|
| 368 |
-
"logits/chosen": -1.2666015625,
|
| 369 |
-
"logits/rejected": -1.294921875,
|
| 370 |
-
"logps/chosen": -0.63330078125,
|
| 371 |
-
"logps/rejected": -0.65185546875,
|
| 372 |
-
"loss": 1.3572,
|
| 373 |
-
"nll_loss": 0.62646484375,
|
| 374 |
-
"rewards/accuracies": 0.5390625,
|
| 375 |
-
"rewards/chosen": -0.12664794921875,
|
| 376 |
-
"rewards/margins": 0.003692626953125,
|
| 377 |
-
"rewards/rejected": -0.13043212890625,
|
| 378 |
-
"step": 23
|
| 379 |
-
},
|
| 380 |
-
{
|
| 381 |
-
"epoch": 0.3072,
|
| 382 |
-
"grad_norm": 4.211955158078266,
|
| 383 |
-
"learning_rate": 8.765357330018055e-07,
|
| 384 |
-
"logits/chosen": -1.1435546875,
|
| 385 |
-
"logits/rejected": -1.126953125,
|
| 386 |
-
"logps/chosen": -0.69140625,
|
| 387 |
-
"logps/rejected": -0.68115234375,
|
| 388 |
-
"loss": 1.4253,
|
| 389 |
-
"nll_loss": 0.69287109375,
|
| 390 |
-
"rewards/accuracies": 0.4765625,
|
| 391 |
-
"rewards/chosen": -0.13812255859375,
|
| 392 |
-
"rewards/margins": -0.0019550323486328125,
|
| 393 |
-
"rewards/rejected": -0.13629150390625,
|
| 394 |
-
"step": 24
|
| 395 |
-
},
|
| 396 |
-
{
|
| 397 |
-
"epoch": 0.32,
|
| 398 |
-
"grad_norm": 3.6426142550151868,
|
| 399 |
-
"learning_rate": 8.613974319136957e-07,
|
| 400 |
-
"logits/chosen": -1.046875,
|
| 401 |
-
"logits/rejected": -1.03515625,
|
| 402 |
-
"logps/chosen": -0.6533203125,
|
| 403 |
-
"logps/rejected": -0.62353515625,
|
| 404 |
-
"loss": 1.3998,
|
| 405 |
-
"nll_loss": 0.6640625,
|
| 406 |
-
"rewards/accuracies": 0.4609375,
|
| 407 |
-
"rewards/chosen": -0.130615234375,
|
| 408 |
-
"rewards/margins": -0.005828857421875,
|
| 409 |
-
"rewards/rejected": -0.12469482421875,
|
| 410 |
-
"step": 25
|
| 411 |
-
},
|
| 412 |
-
{
|
| 413 |
-
"epoch": 0.3328,
|
| 414 |
-
"grad_norm": 3.170105238410109,
|
| 415 |
-
"learning_rate": 8.455313244934324e-07,
|
| 416 |
-
"logits/chosen": -0.92333984375,
|
| 417 |
-
"logits/rejected": -0.90771484375,
|
| 418 |
-
"logps/chosen": -0.6552734375,
|
| 419 |
-
"logps/rejected": -0.65478515625,
|
| 420 |
-
"loss": 1.403,
|
| 421 |
-
"nll_loss": 0.6640625,
|
| 422 |
-
"rewards/accuracies": 0.4921875,
|
| 423 |
-
"rewards/chosen": -0.13092041015625,
|
| 424 |
-
"rewards/margins": -6.67572021484375e-05,
|
| 425 |
-
"rewards/rejected": -0.13104248046875,
|
| 426 |
-
"step": 26
|
| 427 |
-
},
|
| 428 |
-
{
|
| 429 |
-
"epoch": 0.3456,
|
| 430 |
-
"grad_norm": 3.525274842281527,
|
| 431 |
-
"learning_rate": 8.289693629698563e-07,
|
| 432 |
-
"logits/chosen": -0.84326171875,
|
| 433 |
-
"logits/rejected": -0.81298828125,
|
| 434 |
-
"logps/chosen": -0.63037109375,
|
| 435 |
-
"logps/rejected": -0.65234375,
|
| 436 |
-
"loss": 1.3705,
|
| 437 |
-
"nll_loss": 0.6376953125,
|
| 438 |
-
"rewards/accuracies": 0.5234375,
|
| 439 |
-
"rewards/chosen": -0.12615966796875,
|
| 440 |
-
"rewards/margins": 0.004344940185546875,
|
| 441 |
-
"rewards/rejected": -0.130615234375,
|
| 442 |
-
"step": 27
|
| 443 |
-
},
|
| 444 |
-
{
|
| 445 |
-
"epoch": 0.3584,
|
| 446 |
-
"grad_norm": 3.183918605214547,
|
| 447 |
-
"learning_rate": 8.117449009293668e-07,
|
| 448 |
-
"logits/chosen": -0.806640625,
|
| 449 |
-
"logits/rejected": -0.8115234375,
|
| 450 |
-
"logps/chosen": -0.65087890625,
|
| 451 |
-
"logps/rejected": -0.65576171875,
|
| 452 |
-
"loss": 1.3967,
|
| 453 |
-
"nll_loss": 0.6630859375,
|
| 454 |
-
"rewards/accuracies": 0.515625,
|
| 455 |
-
"rewards/chosen": -0.13018798828125,
|
| 456 |
-
"rewards/margins": 0.0011444091796875,
|
| 457 |
-
"rewards/rejected": -0.13128662109375,
|
| 458 |
-
"step": 28
|
| 459 |
-
},
|
| 460 |
-
{
|
| 461 |
-
"epoch": 0.3712,
|
| 462 |
-
"grad_norm": 3.444618483901198,
|
| 463 |
-
"learning_rate": 7.938926261462365e-07,
|
| 464 |
-
"logits/chosen": -0.76025390625,
|
| 465 |
-
"logits/rejected": -0.74072265625,
|
| 466 |
-
"logps/chosen": -0.64453125,
|
| 467 |
-
"logps/rejected": -0.67333984375,
|
| 468 |
-
"loss": 1.373,
|
| 469 |
-
"nll_loss": 0.6455078125,
|
| 470 |
-
"rewards/accuracies": 0.5234375,
|
| 471 |
-
"rewards/chosen": -0.12890625,
|
| 472 |
-
"rewards/margins": 0.005756378173828125,
|
| 473 |
-
"rewards/rejected": -0.1346435546875,
|
| 474 |
-
"step": 29
|
| 475 |
-
},
|
| 476 |
-
{
|
| 477 |
-
"epoch": 0.384,
|
| 478 |
-
"grad_norm": 2.991968211612028,
|
| 479 |
-
"learning_rate": 7.754484907260512e-07,
|
| 480 |
-
"logits/chosen": -0.81640625,
|
| 481 |
-
"logits/rejected": -0.7978515625,
|
| 482 |
-
"logps/chosen": -0.62548828125,
|
| 483 |
-
"logps/rejected": -0.64306640625,
|
| 484 |
-
"loss": 1.3764,
|
| 485 |
-
"nll_loss": 0.64501953125,
|
| 486 |
-
"rewards/accuracies": 0.5234375,
|
| 487 |
-
"rewards/chosen": -0.1251220703125,
|
| 488 |
-
"rewards/margins": 0.003475189208984375,
|
| 489 |
-
"rewards/rejected": -0.12872314453125,
|
| 490 |
-
"step": 30
|
| 491 |
-
},
|
| 492 |
-
{
|
| 493 |
-
"epoch": 0.3968,
|
| 494 |
-
"grad_norm": 3.3394459626962907,
|
| 495 |
-
"learning_rate": 7.564496387029531e-07,
|
| 496 |
-
"logits/chosen": -0.796875,
|
| 497 |
-
"logits/rejected": -0.77587890625,
|
| 498 |
-
"logps/chosen": -0.60498046875,
|
| 499 |
-
"logps/rejected": -0.63427734375,
|
| 500 |
-
"loss": 1.344,
|
| 501 |
-
"nll_loss": 0.61181640625,
|
| 502 |
-
"rewards/accuracies": 0.5390625,
|
| 503 |
-
"rewards/chosen": -0.1209716796875,
|
| 504 |
-
"rewards/margins": 0.005809783935546875,
|
| 505 |
-
"rewards/rejected": -0.12664794921875,
|
| 506 |
-
"step": 31
|
| 507 |
-
},
|
| 508 |
-
{
|
| 509 |
-
"epoch": 0.4096,
|
| 510 |
-
"grad_norm": 3.1372300962795117,
|
| 511 |
-
"learning_rate": 7.369343312364993e-07,
|
| 512 |
-
"logits/chosen": -0.79248046875,
|
| 513 |
-
"logits/rejected": -0.763671875,
|
| 514 |
-
"logps/chosen": -0.609375,
|
| 515 |
-
"logps/rejected": -0.66357421875,
|
| 516 |
-
"loss": 1.3492,
|
| 517 |
-
"nll_loss": 0.62060546875,
|
| 518 |
-
"rewards/accuracies": 0.5234375,
|
| 519 |
-
"rewards/chosen": -0.121826171875,
|
| 520 |
-
"rewards/margins": 0.01091766357421875,
|
| 521 |
-
"rewards/rejected": -0.13275146484375,
|
| 522 |
-
"step": 32
|
| 523 |
-
},
|
| 524 |
-
{
|
| 525 |
-
"epoch": 0.4224,
|
| 526 |
-
"grad_norm": 2.9015142762383306,
|
| 527 |
-
"learning_rate": 7.16941869558779e-07,
|
| 528 |
-
"logits/chosen": -0.8271484375,
|
| 529 |
-
"logits/rejected": -0.81298828125,
|
| 530 |
-
"logps/chosen": -0.6416015625,
|
| 531 |
-
"logps/rejected": -0.5986328125,
|
| 532 |
-
"loss": 1.3815,
|
| 533 |
-
"nll_loss": 0.6455078125,
|
| 534 |
-
"rewards/accuracies": 0.4296875,
|
| 535 |
-
"rewards/chosen": -0.12841796875,
|
| 536 |
-
"rewards/margins": -0.008785247802734375,
|
| 537 |
-
"rewards/rejected": -0.11944580078125,
|
| 538 |
-
"step": 33
|
| 539 |
-
},
|
| 540 |
-
{
|
| 541 |
-
"epoch": 0.4352,
|
| 542 |
-
"grad_norm": 2.8996886974175613,
|
| 543 |
-
"learning_rate": 6.965125158269618e-07,
|
| 544 |
-
"logits/chosen": -0.83349609375,
|
| 545 |
-
"logits/rejected": -0.802734375,
|
| 546 |
-
"logps/chosen": -0.62060546875,
|
| 547 |
-
"logps/rejected": -0.671875,
|
| 548 |
-
"loss": 1.3428,
|
| 549 |
-
"nll_loss": 0.61962890625,
|
| 550 |
-
"rewards/accuracies": 0.53125,
|
| 551 |
-
"rewards/chosen": -0.1240234375,
|
| 552 |
-
"rewards/margins": 0.01045989990234375,
|
| 553 |
-
"rewards/rejected": -0.1343994140625,
|
| 554 |
-
"step": 34
|
| 555 |
-
},
|
| 556 |
-
{
|
| 557 |
-
"epoch": 0.448,
|
| 558 |
-
"grad_norm": 3.2300028712965543,
|
| 559 |
-
"learning_rate": 6.756874120406714e-07,
|
| 560 |
-
"logits/chosen": -0.87744140625,
|
| 561 |
-
"logits/rejected": -0.84326171875,
|
| 562 |
-
"logps/chosen": -0.67919921875,
|
| 563 |
-
"logps/rejected": -0.64697265625,
|
| 564 |
-
"loss": 1.4155,
|
| 565 |
-
"nll_loss": 0.68017578125,
|
| 566 |
-
"rewards/accuracies": 0.4296875,
|
| 567 |
-
"rewards/chosen": -0.1356201171875,
|
| 568 |
-
"rewards/margins": -0.0063838958740234375,
|
| 569 |
-
"rewards/rejected": -0.1295166015625,
|
| 570 |
-
"step": 35
|
| 571 |
-
},
|
| 572 |
-
{
|
| 573 |
-
"epoch": 0.4608,
|
| 574 |
-
"grad_norm": 3.2182679046770506,
|
| 575 |
-
"learning_rate": 6.545084971874736e-07,
|
| 576 |
-
"logits/chosen": -0.7998046875,
|
| 577 |
-
"logits/rejected": -0.7900390625,
|
| 578 |
-
"logps/chosen": -0.69189453125,
|
| 579 |
-
"logps/rejected": -0.66015625,
|
| 580 |
-
"loss": 1.4003,
|
| 581 |
-
"nll_loss": 0.66796875,
|
| 582 |
-
"rewards/accuracies": 0.46875,
|
| 583 |
-
"rewards/chosen": -0.13861083984375,
|
| 584 |
-
"rewards/margins": -0.0065364837646484375,
|
| 585 |
-
"rewards/rejected": -0.13177490234375,
|
| 586 |
-
"step": 36
|
| 587 |
-
},
|
| 588 |
-
{
|
| 589 |
-
"epoch": 0.4736,
|
| 590 |
-
"grad_norm": 3.357656349751252,
|
| 591 |
-
"learning_rate": 6.330184227833375e-07,
|
| 592 |
-
"logits/chosen": -0.75,
|
| 593 |
-
"logits/rejected": -0.77734375,
|
| 594 |
-
"logps/chosen": -0.6796875,
|
| 595 |
-
"logps/rejected": -0.6865234375,
|
| 596 |
-
"loss": 1.3884,
|
| 597 |
-
"nll_loss": 0.6611328125,
|
| 598 |
-
"rewards/accuracies": 0.484375,
|
| 599 |
-
"rewards/chosen": -0.13580322265625,
|
| 600 |
-
"rewards/margins": 0.00138092041015625,
|
| 601 |
-
"rewards/rejected": -0.13726806640625,
|
| 602 |
-
"step": 37
|
| 603 |
-
},
|
| 604 |
-
{
|
| 605 |
-
"epoch": 0.4864,
|
| 606 |
-
"grad_norm": 2.958420783527872,
|
| 607 |
-
"learning_rate": 6.112604669781572e-07,
|
| 608 |
-
"logits/chosen": -0.80078125,
|
| 609 |
-
"logits/rejected": -0.7607421875,
|
| 610 |
-
"logps/chosen": -0.62548828125,
|
| 611 |
-
"logps/rejected": -0.62060546875,
|
| 612 |
-
"loss": 1.3707,
|
| 613 |
-
"nll_loss": 0.63818359375,
|
| 614 |
-
"rewards/accuracies": 0.5390625,
|
| 615 |
-
"rewards/chosen": -0.1251220703125,
|
| 616 |
-
"rewards/margins": -0.0009746551513671875,
|
| 617 |
-
"rewards/rejected": -0.12420654296875,
|
| 618 |
-
"step": 38
|
| 619 |
-
},
|
| 620 |
-
{
|
| 621 |
-
"epoch": 0.4992,
|
| 622 |
-
"grad_norm": 2.9518014676807196,
|
| 623 |
-
"learning_rate": 5.892784473993183e-07,
|
| 624 |
-
"logits/chosen": -0.76416015625,
|
| 625 |
-
"logits/rejected": -0.75390625,
|
| 626 |
-
"logps/chosen": -0.62060546875,
|
| 627 |
-
"logps/rejected": -0.64599609375,
|
| 628 |
-
"loss": 1.3696,
|
| 629 |
-
"nll_loss": 0.64501953125,
|
| 630 |
-
"rewards/accuracies": 0.4921875,
|
| 631 |
-
"rewards/chosen": -0.1241455078125,
|
| 632 |
-
"rewards/margins": 0.004924774169921875,
|
| 633 |
-
"rewards/rejected": -0.129150390625,
|
| 634 |
-
"step": 39
|
| 635 |
-
},
|
| 636 |
-
{
|
| 637 |
-
"epoch": 0.512,
|
| 638 |
-
"grad_norm": 3.378944917137857,
|
| 639 |
-
"learning_rate": 5.671166329088277e-07,
|
| 640 |
-
"logits/chosen": -0.7734375,
|
| 641 |
-
"logits/rejected": -0.765625,
|
| 642 |
-
"logps/chosen": -0.6513671875,
|
| 643 |
-
"logps/rejected": -0.63671875,
|
| 644 |
-
"loss": 1.3698,
|
| 645 |
-
"nll_loss": 0.63525390625,
|
| 646 |
-
"rewards/accuracies": 0.4921875,
|
| 647 |
-
"rewards/chosen": -0.13043212890625,
|
| 648 |
-
"rewards/margins": -0.0031585693359375,
|
| 649 |
-
"rewards/rejected": -0.12725830078125,
|
| 650 |
-
"step": 40
|
| 651 |
-
},
|
| 652 |
-
{
|
| 653 |
-
"epoch": 0.5248,
|
| 654 |
-
"grad_norm": 2.9085205396792935,
|
| 655 |
-
"learning_rate": 5.448196544517167e-07,
|
| 656 |
-
"logits/chosen": -0.732421875,
|
| 657 |
-
"logits/rejected": -0.75390625,
|
| 658 |
-
"logps/chosen": -0.60546875,
|
| 659 |
-
"logps/rejected": -0.61376953125,
|
| 660 |
-
"loss": 1.3409,
|
| 661 |
-
"nll_loss": 0.6103515625,
|
| 662 |
-
"rewards/accuracies": 0.5390625,
|
| 663 |
-
"rewards/chosen": -0.12115478515625,
|
| 664 |
-
"rewards/margins": 0.0018444061279296875,
|
| 665 |
-
"rewards/rejected": -0.12286376953125,
|
| 666 |
-
"step": 41
|
| 667 |
-
},
|
| 668 |
-
{
|
| 669 |
-
"epoch": 0.5376,
|
| 670 |
-
"grad_norm": 3.181717208399589,
|
| 671 |
-
"learning_rate": 5.224324151752575e-07,
|
| 672 |
-
"logits/chosen": -0.6962890625,
|
| 673 |
-
"logits/rejected": -0.6875,
|
| 674 |
-
"logps/chosen": -0.703125,
|
| 675 |
-
"logps/rejected": -0.64794921875,
|
| 676 |
-
"loss": 1.4314,
|
| 677 |
-
"nll_loss": 0.6953125,
|
| 678 |
-
"rewards/accuracies": 0.421875,
|
| 679 |
-
"rewards/chosen": -0.14068603515625,
|
| 680 |
-
"rewards/margins": -0.0110321044921875,
|
| 681 |
-
"rewards/rejected": -0.1295166015625,
|
| 682 |
-
"step": 42
|
| 683 |
-
},
|
| 684 |
-
{
|
| 685 |
-
"epoch": 0.5504,
|
| 686 |
-
"grad_norm": 2.9543674015200043,
|
| 687 |
-
"learning_rate": 5e-07,
|
| 688 |
-
"logits/chosen": -0.6279296875,
|
| 689 |
-
"logits/rejected": -0.63818359375,
|
| 690 |
-
"logps/chosen": -0.62158203125,
|
| 691 |
-
"logps/rejected": -0.68359375,
|
| 692 |
-
"loss": 1.3519,
|
| 693 |
-
"nll_loss": 0.62548828125,
|
| 694 |
-
"rewards/accuracies": 0.546875,
|
| 695 |
-
"rewards/chosen": -0.1243896484375,
|
| 696 |
-
"rewards/margins": 0.01232147216796875,
|
| 697 |
-
"rewards/rejected": -0.1365966796875,
|
| 698 |
-
"step": 43
|
| 699 |
-
},
|
| 700 |
-
{
|
| 701 |
-
"epoch": 0.5632,
|
| 702 |
-
"grad_norm": 2.886051167891686,
|
| 703 |
-
"learning_rate": 4.775675848247427e-07,
|
| 704 |
-
"logits/chosen": -0.66796875,
|
| 705 |
-
"logits/rejected": -0.6806640625,
|
| 706 |
-
"logps/chosen": -0.62890625,
|
| 707 |
-
"logps/rejected": -0.64501953125,
|
| 708 |
-
"loss": 1.3691,
|
| 709 |
-
"nll_loss": 0.64013671875,
|
| 710 |
-
"rewards/accuracies": 0.4296875,
|
| 711 |
-
"rewards/chosen": -0.12591552734375,
|
| 712 |
-
"rewards/margins": 0.003047943115234375,
|
| 713 |
-
"rewards/rejected": -0.1290283203125,
|
| 714 |
-
"step": 44
|
| 715 |
-
},
|
| 716 |
-
{
|
| 717 |
-
"epoch": 0.576,
|
| 718 |
-
"grad_norm": 2.9113924204417487,
|
| 719 |
-
"learning_rate": 4.5518034554828327e-07,
|
| 720 |
-
"logits/chosen": -0.673828125,
|
| 721 |
-
"logits/rejected": -0.65234375,
|
| 722 |
-
"logps/chosen": -0.619140625,
|
| 723 |
-
"logps/rejected": -0.629150390625,
|
| 724 |
-
"loss": 1.3697,
|
| 725 |
-
"nll_loss": 0.63916015625,
|
| 726 |
-
"rewards/accuracies": 0.4453125,
|
| 727 |
-
"rewards/chosen": -0.12384033203125,
|
| 728 |
-
"rewards/margins": 0.001987457275390625,
|
| 729 |
-
"rewards/rejected": -0.1258544921875,
|
| 730 |
-
"step": 45
|
| 731 |
-
},
|
| 732 |
-
{
|
| 733 |
-
"epoch": 0.5888,
|
| 734 |
-
"grad_norm": 2.8855468683899925,
|
| 735 |
-
"learning_rate": 4.328833670911724e-07,
|
| 736 |
-
"logits/chosen": -0.580078125,
|
| 737 |
-
"logits/rejected": -0.5810546875,
|
| 738 |
-
"logps/chosen": -0.638671875,
|
| 739 |
-
"logps/rejected": -0.6015625,
|
| 740 |
-
"loss": 1.3763,
|
| 741 |
-
"nll_loss": 0.6435546875,
|
| 742 |
-
"rewards/accuracies": 0.4453125,
|
| 743 |
-
"rewards/chosen": -0.12762451171875,
|
| 744 |
-
"rewards/margins": -0.00727081298828125,
|
| 745 |
-
"rewards/rejected": -0.120361328125,
|
| 746 |
-
"step": 46
|
| 747 |
-
},
|
| 748 |
-
{
|
| 749 |
-
"epoch": 0.6016,
|
| 750 |
-
"grad_norm": 3.0797159953277253,
|
| 751 |
-
"learning_rate": 4.107215526006817e-07,
|
| 752 |
-
"logits/chosen": -0.59326171875,
|
| 753 |
-
"logits/rejected": -0.58642578125,
|
| 754 |
-
"logps/chosen": -0.65673828125,
|
| 755 |
-
"logps/rejected": -0.6201171875,
|
| 756 |
-
"loss": 1.3972,
|
| 757 |
-
"nll_loss": 0.66162109375,
|
| 758 |
-
"rewards/accuracies": 0.484375,
|
| 759 |
-
"rewards/chosen": -0.13153076171875,
|
| 760 |
-
"rewards/margins": -0.0073795318603515625,
|
| 761 |
-
"rewards/rejected": -0.12408447265625,
|
| 762 |
-
"step": 47
|
| 763 |
-
},
|
| 764 |
-
{
|
| 765 |
-
"epoch": 0.6144,
|
| 766 |
-
"grad_norm": 2.8066569349734865,
|
| 767 |
-
"learning_rate": 3.8873953302184283e-07,
|
| 768 |
-
"logits/chosen": -0.609619140625,
|
| 769 |
-
"logits/rejected": -0.586669921875,
|
| 770 |
-
"logps/chosen": -0.65576171875,
|
| 771 |
-
"logps/rejected": -0.63525390625,
|
| 772 |
-
"loss": 1.3826,
|
| 773 |
-
"nll_loss": 0.650390625,
|
| 774 |
-
"rewards/accuracies": 0.40625,
|
| 775 |
-
"rewards/chosen": -0.1312255859375,
|
| 776 |
-
"rewards/margins": -0.0042667388916015625,
|
| 777 |
-
"rewards/rejected": -0.1270751953125,
|
| 778 |
-
"step": 48
|
| 779 |
-
},
|
| 780 |
-
{
|
| 781 |
-
"epoch": 0.6272,
|
| 782 |
-
"grad_norm": 2.8520211464952627,
|
| 783 |
-
"learning_rate": 3.669815772166625e-07,
|
| 784 |
-
"logits/chosen": -0.55810546875,
|
| 785 |
-
"logits/rejected": -0.52392578125,
|
| 786 |
-
"logps/chosen": -0.59716796875,
|
| 787 |
-
"logps/rejected": -0.61669921875,
|
| 788 |
-
"loss": 1.3468,
|
| 789 |
-
"nll_loss": 0.6171875,
|
| 790 |
-
"rewards/accuracies": 0.453125,
|
| 791 |
-
"rewards/chosen": -0.1195068359375,
|
| 792 |
-
"rewards/margins": 0.00395965576171875,
|
| 793 |
-
"rewards/rejected": -0.12322998046875,
|
| 794 |
-
"step": 49
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 0.64,
|
| 798 |
-
"grad_norm": 2.7486285344409525,
|
| 799 |
-
"learning_rate": 3.454915028125263e-07,
|
| 800 |
-
"logits/chosen": -0.614013671875,
|
| 801 |
-
"logits/rejected": -0.58740234375,
|
| 802 |
-
"logps/chosen": -0.68701171875,
|
| 803 |
-
"logps/rejected": -0.6611328125,
|
| 804 |
-
"loss": 1.4158,
|
| 805 |
-
"nll_loss": 0.6806640625,
|
| 806 |
-
"rewards/accuracies": 0.4140625,
|
| 807 |
-
"rewards/chosen": -0.1375732421875,
|
| 808 |
-
"rewards/margins": -0.0052089691162109375,
|
| 809 |
-
"rewards/rejected": -0.1324462890625,
|
| 810 |
-
"step": 50
|
| 811 |
-
},
|
| 812 |
-
{
|
| 813 |
-
"epoch": 0.6528,
|
| 814 |
-
"grad_norm": 2.703029476528635,
|
| 815 |
-
"learning_rate": 3.243125879593286e-07,
|
| 816 |
-
"logits/chosen": -0.6357421875,
|
| 817 |
-
"logits/rejected": -0.63671875,
|
| 818 |
-
"logps/chosen": -0.65234375,
|
| 819 |
-
"logps/rejected": -0.625,
|
| 820 |
-
"loss": 1.3873,
|
| 821 |
-
"nll_loss": 0.65283203125,
|
| 822 |
-
"rewards/accuracies": 0.421875,
|
| 823 |
-
"rewards/chosen": -0.13043212890625,
|
| 824 |
-
"rewards/margins": -0.0056476593017578125,
|
| 825 |
-
"rewards/rejected": -0.12469482421875,
|
| 826 |
-
"step": 51
|
| 827 |
-
},
|
| 828 |
-
{
|
| 829 |
-
"epoch": 0.6656,
|
| 830 |
-
"grad_norm": 2.913686937773188,
|
| 831 |
-
"learning_rate": 3.034874841730382e-07,
|
| 832 |
-
"logits/chosen": -0.62744140625,
|
| 833 |
-
"logits/rejected": -0.63427734375,
|
| 834 |
-
"logps/chosen": -0.6162109375,
|
| 835 |
-
"logps/rejected": -0.592041015625,
|
| 836 |
-
"loss": 1.3665,
|
| 837 |
-
"nll_loss": 0.63330078125,
|
| 838 |
-
"rewards/accuracies": 0.4765625,
|
| 839 |
-
"rewards/chosen": -0.12322998046875,
|
| 840 |
-
"rewards/margins": -0.004718780517578125,
|
| 841 |
-
"rewards/rejected": -0.11846923828125,
|
| 842 |
-
"step": 52
|
| 843 |
-
},
|
| 844 |
-
{
|
| 845 |
-
"epoch": 0.6784,
|
| 846 |
-
"grad_norm": 2.8293012838870455,
|
| 847 |
-
"learning_rate": 2.8305813044122093e-07,
|
| 848 |
-
"logits/chosen": -0.603271484375,
|
| 849 |
-
"logits/rejected": -0.560546875,
|
| 850 |
-
"logps/chosen": -0.63525390625,
|
| 851 |
-
"logps/rejected": -0.6455078125,
|
| 852 |
-
"loss": 1.3786,
|
| 853 |
-
"nll_loss": 0.6474609375,
|
| 854 |
-
"rewards/accuracies": 0.4765625,
|
| 855 |
-
"rewards/chosen": -0.1270751953125,
|
| 856 |
-
"rewards/margins": 0.002071380615234375,
|
| 857 |
-
"rewards/rejected": -0.12908935546875,
|
| 858 |
-
"step": 53
|
| 859 |
-
},
|
| 860 |
-
{
|
| 861 |
-
"epoch": 0.6912,
|
| 862 |
-
"grad_norm": 2.9232332876985856,
|
| 863 |
-
"learning_rate": 2.6306566876350067e-07,
|
| 864 |
-
"logits/chosen": -0.585205078125,
|
| 865 |
-
"logits/rejected": -0.56689453125,
|
| 866 |
-
"logps/chosen": -0.59716796875,
|
| 867 |
-
"logps/rejected": -0.619140625,
|
| 868 |
-
"loss": 1.3566,
|
| 869 |
-
"nll_loss": 0.6279296875,
|
| 870 |
-
"rewards/accuracies": 0.5703125,
|
| 871 |
-
"rewards/chosen": -0.11944580078125,
|
| 872 |
-
"rewards/margins": 0.0044689178466796875,
|
| 873 |
-
"rewards/rejected": -0.12384033203125,
|
| 874 |
-
"step": 54
|
| 875 |
-
},
|
| 876 |
-
{
|
| 877 |
-
"epoch": 0.704,
|
| 878 |
-
"grad_norm": 2.896454906763871,
|
| 879 |
-
"learning_rate": 2.4355036129704696e-07,
|
| 880 |
-
"logits/chosen": -0.61572265625,
|
| 881 |
-
"logits/rejected": -0.5947265625,
|
| 882 |
-
"logps/chosen": -0.69970703125,
|
| 883 |
-
"logps/rejected": -0.66015625,
|
| 884 |
-
"loss": 1.4196,
|
| 885 |
-
"nll_loss": 0.68603515625,
|
| 886 |
-
"rewards/accuracies": 0.4453125,
|
| 887 |
-
"rewards/chosen": -0.1400146484375,
|
| 888 |
-
"rewards/margins": -0.00769805908203125,
|
| 889 |
-
"rewards/rejected": -0.13226318359375,
|
| 890 |
-
"step": 55
|
| 891 |
-
},
|
| 892 |
-
{
|
| 893 |
-
"epoch": 0.7168,
|
| 894 |
-
"grad_norm": 3.1419979908995637,
|
| 895 |
-
"learning_rate": 2.2455150927394878e-07,
|
| 896 |
-
"logits/chosen": -0.609375,
|
| 897 |
-
"logits/rejected": -0.6103515625,
|
| 898 |
-
"logps/chosen": -0.669921875,
|
| 899 |
-
"logps/rejected": -0.66162109375,
|
| 900 |
-
"loss": 1.4077,
|
| 901 |
-
"nll_loss": 0.6787109375,
|
| 902 |
-
"rewards/accuracies": 0.4375,
|
| 903 |
-
"rewards/chosen": -0.1339111328125,
|
| 904 |
-
"rewards/margins": -0.001544952392578125,
|
| 905 |
-
"rewards/rejected": -0.1322021484375,
|
| 906 |
-
"step": 56
|
| 907 |
-
},
|
| 908 |
-
{
|
| 909 |
-
"epoch": 0.7296,
|
| 910 |
-
"grad_norm": 2.6977667650287436,
|
| 911 |
-
"learning_rate": 2.0610737385376348e-07,
|
| 912 |
-
"logits/chosen": -0.626953125,
|
| 913 |
-
"logits/rejected": -0.64013671875,
|
| 914 |
-
"logps/chosen": -0.6337890625,
|
| 915 |
-
"logps/rejected": -0.65771484375,
|
| 916 |
-
"loss": 1.3617,
|
| 917 |
-
"nll_loss": 0.62744140625,
|
| 918 |
-
"rewards/accuracies": 0.4921875,
|
| 919 |
-
"rewards/chosen": -0.1268310546875,
|
| 920 |
-
"rewards/margins": 0.0049648284912109375,
|
| 921 |
-
"rewards/rejected": -0.13165283203125,
|
| 922 |
-
"step": 57
|
| 923 |
-
},
|
| 924 |
-
{
|
| 925 |
-
"epoch": 0.7424,
|
| 926 |
-
"grad_norm": 2.6962689015906887,
|
| 927 |
-
"learning_rate": 1.8825509907063326e-07,
|
| 928 |
-
"logits/chosen": -0.60107421875,
|
| 929 |
-
"logits/rejected": -0.595703125,
|
| 930 |
-
"logps/chosen": -0.67333984375,
|
| 931 |
-
"logps/rejected": -0.65576171875,
|
| 932 |
-
"loss": 1.4066,
|
| 933 |
-
"nll_loss": 0.67236328125,
|
| 934 |
-
"rewards/accuracies": 0.46875,
|
| 935 |
-
"rewards/chosen": -0.13458251953125,
|
| 936 |
-
"rewards/margins": -0.0033740997314453125,
|
| 937 |
-
"rewards/rejected": -0.13134765625,
|
| 938 |
-
"step": 58
|
| 939 |
-
},
|
| 940 |
-
{
|
| 941 |
-
"epoch": 0.7552,
|
| 942 |
-
"grad_norm": 2.831274649742223,
|
| 943 |
-
"learning_rate": 1.710306370301437e-07,
|
| 944 |
-
"logits/chosen": -0.606689453125,
|
| 945 |
-
"logits/rejected": -0.6005859375,
|
| 946 |
-
"logps/chosen": -0.6484375,
|
| 947 |
-
"logps/rejected": -0.63916015625,
|
| 948 |
-
"loss": 1.387,
|
| 949 |
-
"nll_loss": 0.65234375,
|
| 950 |
-
"rewards/accuracies": 0.4375,
|
| 951 |
-
"rewards/chosen": -0.129638671875,
|
| 952 |
-
"rewards/margins": -0.0019359588623046875,
|
| 953 |
-
"rewards/rejected": -0.12774658203125,
|
| 954 |
-
"step": 59
|
| 955 |
-
},
|
| 956 |
-
{
|
| 957 |
-
"epoch": 0.768,
|
| 958 |
-
"grad_norm": 2.8001049294565368,
|
| 959 |
-
"learning_rate": 1.5446867550656767e-07,
|
| 960 |
-
"logits/chosen": -0.61376953125,
|
| 961 |
-
"logits/rejected": -0.606689453125,
|
| 962 |
-
"logps/chosen": -0.57080078125,
|
| 963 |
-
"logps/rejected": -0.60791015625,
|
| 964 |
-
"loss": 1.3087,
|
| 965 |
-
"nll_loss": 0.5830078125,
|
| 966 |
-
"rewards/accuracies": 0.5234375,
|
| 967 |
-
"rewards/chosen": -0.114013671875,
|
| 968 |
-
"rewards/margins": 0.007526397705078125,
|
| 969 |
-
"rewards/rejected": -0.12158203125,
|
| 970 |
-
"step": 60
|
| 971 |
-
},
|
| 972 |
-
{
|
| 973 |
-
"epoch": 0.7808,
|
| 974 |
-
"grad_norm": 2.6480528965763517,
|
| 975 |
-
"learning_rate": 1.3860256808630427e-07,
|
| 976 |
-
"logits/chosen": -0.648193359375,
|
| 977 |
-
"logits/rejected": -0.64013671875,
|
| 978 |
-
"logps/chosen": -0.593994140625,
|
| 979 |
-
"logps/rejected": -0.638671875,
|
| 980 |
-
"loss": 1.3475,
|
| 981 |
-
"nll_loss": 0.62060546875,
|
| 982 |
-
"rewards/accuracies": 0.53125,
|
| 983 |
-
"rewards/chosen": -0.1187744140625,
|
| 984 |
-
"rewards/margins": 0.009082794189453125,
|
| 985 |
-
"rewards/rejected": -0.1278076171875,
|
| 986 |
-
"step": 61
|
| 987 |
-
},
|
| 988 |
-
{
|
| 989 |
-
"epoch": 0.7936,
|
| 990 |
-
"grad_norm": 2.8416205177773515,
|
| 991 |
-
"learning_rate": 1.2346426699819456e-07,
|
| 992 |
-
"logits/chosen": -0.5224609375,
|
| 993 |
-
"logits/rejected": -0.52880859375,
|
| 994 |
-
"logps/chosen": -0.60302734375,
|
| 995 |
-
"logps/rejected": -0.620361328125,
|
| 996 |
-
"loss": 1.3488,
|
| 997 |
-
"nll_loss": 0.62255859375,
|
| 998 |
-
"rewards/accuracies": 0.5078125,
|
| 999 |
-
"rewards/chosen": -0.12066650390625,
|
| 1000 |
-
"rewards/margins": 0.00335693359375,
|
| 1001 |
-
"rewards/rejected": -0.1240234375,
|
| 1002 |
-
"step": 62
|
| 1003 |
-
},
|
| 1004 |
-
{
|
| 1005 |
-
"epoch": 0.8064,
|
| 1006 |
-
"grad_norm": 2.991877049209725,
|
| 1007 |
-
"learning_rate": 1.090842587659851e-07,
|
| 1008 |
-
"logits/chosen": -0.56982421875,
|
| 1009 |
-
"logits/rejected": -0.55224609375,
|
| 1010 |
-
"logps/chosen": -0.65283203125,
|
| 1011 |
-
"logps/rejected": -0.61962890625,
|
| 1012 |
-
"loss": 1.3911,
|
| 1013 |
-
"nll_loss": 0.65869140625,
|
| 1014 |
-
"rewards/accuracies": 0.453125,
|
| 1015 |
-
"rewards/chosen": -0.13055419921875,
|
| 1016 |
-
"rewards/margins": -0.0066547393798828125,
|
| 1017 |
-
"rewards/rejected": -0.1239013671875,
|
| 1018 |
-
"step": 63
|
| 1019 |
-
},
|
| 1020 |
-
{
|
| 1021 |
-
"epoch": 0.8192,
|
| 1022 |
-
"grad_norm": 2.9240870950358966,
|
| 1023 |
-
"learning_rate": 9.549150281252632e-08,
|
| 1024 |
-
"logits/chosen": -0.64208984375,
|
| 1025 |
-
"logits/rejected": -0.62939453125,
|
| 1026 |
-
"logps/chosen": -0.6103515625,
|
| 1027 |
-
"logps/rejected": -0.68408203125,
|
| 1028 |
-
"loss": 1.3376,
|
| 1029 |
-
"nll_loss": 0.61083984375,
|
| 1030 |
-
"rewards/accuracies": 0.53125,
|
| 1031 |
-
"rewards/chosen": -0.1221923828125,
|
| 1032 |
-
"rewards/margins": 0.014739990234375,
|
| 1033 |
-
"rewards/rejected": -0.13702392578125,
|
| 1034 |
-
"step": 64
|
| 1035 |
-
},
|
| 1036 |
-
{
|
| 1037 |
-
"epoch": 0.832,
|
| 1038 |
-
"grad_norm": 2.8243780447328737,
|
| 1039 |
-
"learning_rate": 8.271337313934867e-08,
|
| 1040 |
-
"logits/chosen": -0.61376953125,
|
| 1041 |
-
"logits/rejected": -0.61181640625,
|
| 1042 |
-
"logps/chosen": -0.62109375,
|
| 1043 |
-
"logps/rejected": -0.63037109375,
|
| 1044 |
-
"loss": 1.366,
|
| 1045 |
-
"nll_loss": 0.6328125,
|
| 1046 |
-
"rewards/accuracies": 0.53125,
|
| 1047 |
-
"rewards/chosen": -0.12420654296875,
|
| 1048 |
-
"rewards/margins": 0.001834869384765625,
|
| 1049 |
-
"rewards/rejected": -0.1260986328125,
|
| 1050 |
-
"step": 65
|
| 1051 |
-
},
|
| 1052 |
-
{
|
| 1053 |
-
"epoch": 0.8448,
|
| 1054 |
-
"grad_norm": 2.7696317108399424,
|
| 1055 |
-
"learning_rate": 7.077560319906694e-08,
|
| 1056 |
-
"logits/chosen": -0.611328125,
|
| 1057 |
-
"logits/rejected": -0.583984375,
|
| 1058 |
-
"logps/chosen": -0.62158203125,
|
| 1059 |
-
"logps/rejected": -0.6455078125,
|
| 1060 |
-
"loss": 1.3609,
|
| 1061 |
-
"nll_loss": 0.63330078125,
|
| 1062 |
-
"rewards/accuracies": 0.5078125,
|
| 1063 |
-
"rewards/chosen": -0.124267578125,
|
| 1064 |
-
"rewards/margins": 0.00481414794921875,
|
| 1065 |
-
"rewards/rejected": -0.129150390625,
|
| 1066 |
-
"step": 66
|
| 1067 |
-
},
|
| 1068 |
-
{
|
| 1069 |
-
"epoch": 0.8576,
|
| 1070 |
-
"grad_norm": 2.830373976297677,
|
| 1071 |
-
"learning_rate": 5.9702234071631e-08,
|
| 1072 |
-
"logits/chosen": -0.65771484375,
|
| 1073 |
-
"logits/rejected": -0.63037109375,
|
| 1074 |
-
"logps/chosen": -0.619140625,
|
| 1075 |
-
"logps/rejected": -0.60791015625,
|
| 1076 |
-
"loss": 1.3408,
|
| 1077 |
-
"nll_loss": 0.611328125,
|
| 1078 |
-
"rewards/accuracies": 0.484375,
|
| 1079 |
-
"rewards/chosen": -0.12371826171875,
|
| 1080 |
-
"rewards/margins": -0.0019969940185546875,
|
| 1081 |
-
"rewards/rejected": -0.1217041015625,
|
| 1082 |
-
"step": 67
|
| 1083 |
-
},
|
| 1084 |
-
{
|
| 1085 |
-
"epoch": 0.8704,
|
| 1086 |
-
"grad_norm": 3.1263626179626285,
|
| 1087 |
-
"learning_rate": 4.951556604879048e-08,
|
| 1088 |
-
"logits/chosen": -0.590087890625,
|
| 1089 |
-
"logits/rejected": -0.60205078125,
|
| 1090 |
-
"logps/chosen": -0.6533203125,
|
| 1091 |
-
"logps/rejected": -0.68310546875,
|
| 1092 |
-
"loss": 1.3989,
|
| 1093 |
-
"nll_loss": 0.6640625,
|
| 1094 |
-
"rewards/accuracies": 0.53125,
|
| 1095 |
-
"rewards/chosen": -0.13067626953125,
|
| 1096 |
-
"rewards/margins": 0.005911827087402344,
|
| 1097 |
-
"rewards/rejected": -0.13671875,
|
| 1098 |
-
"step": 68
|
| 1099 |
-
},
|
| 1100 |
-
{
|
| 1101 |
-
"epoch": 0.8832,
|
| 1102 |
-
"grad_norm": 2.6445234966305287,
|
| 1103 |
-
"learning_rate": 4.023611372427471e-08,
|
| 1104 |
-
"logits/chosen": -0.63037109375,
|
| 1105 |
-
"logits/rejected": -0.60400390625,
|
| 1106 |
-
"logps/chosen": -0.6259765625,
|
| 1107 |
-
"logps/rejected": -0.63330078125,
|
| 1108 |
-
"loss": 1.3585,
|
| 1109 |
-
"nll_loss": 0.626953125,
|
| 1110 |
-
"rewards/accuracies": 0.4453125,
|
| 1111 |
-
"rewards/chosen": -0.12506103515625,
|
| 1112 |
-
"rewards/margins": 0.0015201568603515625,
|
| 1113 |
-
"rewards/rejected": -0.1265869140625,
|
| 1114 |
-
"step": 69
|
| 1115 |
-
},
|
| 1116 |
-
{
|
| 1117 |
-
"epoch": 0.896,
|
| 1118 |
-
"grad_norm": 2.7600272041859806,
|
| 1119 |
-
"learning_rate": 3.188256468013139e-08,
|
| 1120 |
-
"logits/chosen": -0.606689453125,
|
| 1121 |
-
"logits/rejected": -0.603515625,
|
| 1122 |
-
"logps/chosen": -0.62939453125,
|
| 1123 |
-
"logps/rejected": -0.64794921875,
|
| 1124 |
-
"loss": 1.3466,
|
| 1125 |
-
"nll_loss": 0.61669921875,
|
| 1126 |
-
"rewards/accuracies": 0.5234375,
|
| 1127 |
-
"rewards/chosen": -0.1260986328125,
|
| 1128 |
-
"rewards/margins": 0.00376129150390625,
|
| 1129 |
-
"rewards/rejected": -0.129638671875,
|
| 1130 |
-
"step": 70
|
| 1131 |
-
},
|
| 1132 |
-
{
|
| 1133 |
-
"epoch": 0.9088,
|
| 1134 |
-
"grad_norm": 2.8920578163456443,
|
| 1135 |
-
"learning_rate": 2.4471741852423233e-08,
|
| 1136 |
-
"logits/chosen": -0.602294921875,
|
| 1137 |
-
"logits/rejected": -0.62451171875,
|
| 1138 |
-
"logps/chosen": -0.6181640625,
|
| 1139 |
-
"logps/rejected": -0.638671875,
|
| 1140 |
-
"loss": 1.3366,
|
| 1141 |
-
"nll_loss": 0.60791015625,
|
| 1142 |
-
"rewards/accuracies": 0.4921875,
|
| 1143 |
-
"rewards/chosen": -0.12371826171875,
|
| 1144 |
-
"rewards/margins": 0.0041656494140625,
|
| 1145 |
-
"rewards/rejected": -0.12786865234375,
|
| 1146 |
-
"step": 71
|
| 1147 |
-
},
|
| 1148 |
-
{
|
| 1149 |
-
"epoch": 0.9216,
|
| 1150 |
-
"grad_norm": 2.824155688076418,
|
| 1151 |
-
"learning_rate": 1.8018569652073378e-08,
|
| 1152 |
-
"logits/chosen": -0.599609375,
|
| 1153 |
-
"logits/rejected": -0.596923828125,
|
| 1154 |
-
"logps/chosen": -0.65673828125,
|
| 1155 |
-
"logps/rejected": -0.62646484375,
|
| 1156 |
-
"loss": 1.3907,
|
| 1157 |
-
"nll_loss": 0.65576171875,
|
| 1158 |
-
"rewards/accuracies": 0.421875,
|
| 1159 |
-
"rewards/chosen": -0.13116455078125,
|
| 1160 |
-
"rewards/margins": -0.005977630615234375,
|
| 1161 |
-
"rewards/rejected": -0.1251220703125,
|
| 1162 |
-
"step": 72
|
| 1163 |
-
},
|
| 1164 |
-
{
|
| 1165 |
-
"epoch": 0.9344,
|
| 1166 |
-
"grad_norm": 2.853212647277363,
|
| 1167 |
-
"learning_rate": 1.253604390908819e-08,
|
| 1168 |
-
"logits/chosen": -0.614990234375,
|
| 1169 |
-
"logits/rejected": -0.59326171875,
|
| 1170 |
-
"logps/chosen": -0.6591796875,
|
| 1171 |
-
"logps/rejected": -0.64306640625,
|
| 1172 |
-
"loss": 1.4049,
|
| 1173 |
-
"nll_loss": 0.671875,
|
| 1174 |
-
"rewards/accuracies": 0.4453125,
|
| 1175 |
-
"rewards/chosen": -0.13177490234375,
|
| 1176 |
-
"rewards/margins": -0.00328826904296875,
|
| 1177 |
-
"rewards/rejected": -0.128662109375,
|
| 1178 |
-
"step": 73
|
| 1179 |
-
},
|
| 1180 |
-
{
|
| 1181 |
-
"epoch": 0.9472,
|
| 1182 |
-
"grad_norm": 2.7951321419475836,
|
| 1183 |
-
"learning_rate": 8.035205700685165e-09,
|
| 1184 |
-
"logits/chosen": -0.595703125,
|
| 1185 |
-
"logits/rejected": -0.59765625,
|
| 1186 |
-
"logps/chosen": -0.669921875,
|
| 1187 |
-
"logps/rejected": -0.63427734375,
|
| 1188 |
-
"loss": 1.3859,
|
| 1189 |
-
"nll_loss": 0.65185546875,
|
| 1190 |
-
"rewards/accuracies": 0.4765625,
|
| 1191 |
-
"rewards/chosen": -0.13385009765625,
|
| 1192 |
-
"rewards/margins": -0.0070209503173828125,
|
| 1193 |
-
"rewards/rejected": -0.126953125,
|
| 1194 |
-
"step": 74
|
| 1195 |
-
},
|
| 1196 |
-
{
|
| 1197 |
-
"epoch": 0.96,
|
| 1198 |
-
"grad_norm": 3.0464830977927186,
|
| 1199 |
-
"learning_rate": 4.5251191160326495e-09,
|
| 1200 |
-
"logits/chosen": -0.63232421875,
|
| 1201 |
-
"logits/rejected": -0.66015625,
|
| 1202 |
-
"logps/chosen": -0.64501953125,
|
| 1203 |
-
"logps/rejected": -0.64453125,
|
| 1204 |
-
"loss": 1.3751,
|
| 1205 |
-
"nll_loss": 0.6435546875,
|
| 1206 |
-
"rewards/accuracies": 0.5,
|
| 1207 |
-
"rewards/chosen": -0.1290283203125,
|
| 1208 |
-
"rewards/margins": -9.1552734375e-05,
|
| 1209 |
-
"rewards/rejected": -0.12890625,
|
| 1210 |
-
"step": 75
|
| 1211 |
-
},
|
| 1212 |
-
{
|
| 1213 |
-
"epoch": 0.9728,
|
| 1214 |
-
"grad_norm": 2.6891038234268576,
|
| 1215 |
-
"learning_rate": 2.0128530023804656e-09,
|
| 1216 |
-
"logits/chosen": -0.65966796875,
|
| 1217 |
-
"logits/rejected": -0.6416015625,
|
| 1218 |
-
"logps/chosen": -0.61376953125,
|
| 1219 |
-
"logps/rejected": -0.62841796875,
|
| 1220 |
-
"loss": 1.3403,
|
| 1221 |
-
"nll_loss": 0.60986328125,
|
| 1222 |
-
"rewards/accuracies": 0.5,
|
| 1223 |
-
"rewards/chosen": -0.1226806640625,
|
| 1224 |
-
"rewards/margins": 0.00304412841796875,
|
| 1225 |
-
"rewards/rejected": -0.12579345703125,
|
| 1226 |
-
"step": 76
|
| 1227 |
-
},
|
| 1228 |
-
{
|
| 1229 |
-
"epoch": 0.9856,
|
| 1230 |
-
"grad_norm": 2.8901246333719484,
|
| 1231 |
-
"learning_rate": 5.034667293427053e-10,
|
| 1232 |
-
"logits/chosen": -0.557373046875,
|
| 1233 |
-
"logits/rejected": -0.55078125,
|
| 1234 |
-
"logps/chosen": -0.62939453125,
|
| 1235 |
-
"logps/rejected": -0.62890625,
|
| 1236 |
-
"loss": 1.3691,
|
| 1237 |
-
"nll_loss": 0.6416015625,
|
| 1238 |
-
"rewards/accuracies": 0.4609375,
|
| 1239 |
-
"rewards/chosen": -0.1258544921875,
|
| 1240 |
-
"rewards/margins": -0.000141143798828125,
|
| 1241 |
-
"rewards/rejected": -0.1256103515625,
|
| 1242 |
-
"step": 77
|
| 1243 |
-
},
|
| 1244 |
-
{
|
| 1245 |
-
"epoch": 0.9984,
|
| 1246 |
-
"grad_norm": 2.938601997312395,
|
| 1247 |
-
"learning_rate": 0.0,
|
| 1248 |
-
"logits/chosen": -0.515625,
|
| 1249 |
-
"logits/rejected": -0.54541015625,
|
| 1250 |
-
"logps/chosen": -0.6240234375,
|
| 1251 |
-
"logps/rejected": -0.6455078125,
|
| 1252 |
-
"loss": 1.3417,
|
| 1253 |
-
"nll_loss": 0.61376953125,
|
| 1254 |
-
"rewards/accuracies": 0.5078125,
|
| 1255 |
-
"rewards/chosen": -0.124755859375,
|
| 1256 |
-
"rewards/margins": 0.004207611083984375,
|
| 1257 |
-
"rewards/rejected": -0.1290283203125,
|
| 1258 |
-
"step": 78
|
| 1259 |
-
}
|
| 1260 |
-
],
|
| 1261 |
-
"logging_steps": 1.0,
|
| 1262 |
-
"max_steps": 78,
|
| 1263 |
-
"num_input_tokens_seen": 0,
|
| 1264 |
-
"num_train_epochs": 1,
|
| 1265 |
-
"save_steps": 500,
|
| 1266 |
-
"stateful_callbacks": {
|
| 1267 |
-
"TrainerControl": {
|
| 1268 |
-
"args": {
|
| 1269 |
-
"should_epoch_stop": false,
|
| 1270 |
-
"should_evaluate": false,
|
| 1271 |
-
"should_log": false,
|
| 1272 |
-
"should_save": true,
|
| 1273 |
-
"should_training_stop": true
|
| 1274 |
-
},
|
| 1275 |
-
"attributes": {}
|
| 1276 |
-
}
|
| 1277 |
-
},
|
| 1278 |
-
"total_flos": 0.0,
|
| 1279 |
-
"train_batch_size": 2,
|
| 1280 |
-
"trial_name": null,
|
| 1281 |
-
"trial_params": null
|
| 1282 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|