File size: 13,559 Bytes
330df1b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 |
{
"best_metric": 0.3756321966648102,
"best_model_checkpoint": "./mistral/20-04-24-Weni-WeniGPT-Agents-Mistral-1.0.6-SFT-1.0.7-DPO_Experiment on DPO with other hyperparameters and best SFT model of WeniGPT-2_max_steps-1470_batch_1_2024-04-20_ppid_9/checkpoint-180",
"epoch": 0.7346938775510204,
"eval_steps": 30,
"global_step": 180,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"grad_norm": 83.44550323486328,
"learning_rate": 6.666666666666667e-07,
"logits/chosen": -1.7825796604156494,
"logits/rejected": -1.7886196374893188,
"logps/chosen": -169.71481323242188,
"logps/rejected": -167.6764373779297,
"loss": 0.6937,
"rewards/accuracies": 0.20000000298023224,
"rewards/chosen": -0.0002815247280523181,
"rewards/margins": -0.0010146332206204534,
"rewards/rejected": 0.0007331084925681353,
"step": 10
},
{
"epoch": 0.08,
"grad_norm": 87.56982421875,
"learning_rate": 1.777777777777778e-06,
"logits/chosen": -1.623265027999878,
"logits/rejected": -1.6968040466308594,
"logps/chosen": -128.7312469482422,
"logps/rejected": -184.20187377929688,
"loss": 0.6912,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0030528828501701355,
"rewards/margins": 0.004004458896815777,
"rewards/rejected": -0.0009515761630609632,
"step": 20
},
{
"epoch": 0.12,
"grad_norm": 113.4741439819336,
"learning_rate": 2.888888888888889e-06,
"logits/chosen": -1.7152118682861328,
"logits/rejected": -1.758681058883667,
"logps/chosen": -146.91494750976562,
"logps/rejected": -213.5966339111328,
"loss": 0.6781,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.03251449763774872,
"rewards/margins": 0.030697176232933998,
"rewards/rejected": 0.0018173219868913293,
"step": 30
},
{
"epoch": 0.12,
"eval_logits/chosen": -1.795080304145813,
"eval_logits/rejected": -1.8307874202728271,
"eval_logps/chosen": -194.3749542236328,
"eval_logps/rejected": -243.1332244873047,
"eval_loss": 0.6762310266494751,
"eval_rewards/accuracies": 0.75,
"eval_rewards/chosen": 0.05039369314908981,
"eval_rewards/margins": 0.03467999026179314,
"eval_rewards/rejected": 0.01571369729936123,
"eval_runtime": 26.2365,
"eval_samples_per_second": 1.067,
"eval_steps_per_second": 1.067,
"step": 30
},
{
"epoch": 0.16,
"grad_norm": 50.695308685302734,
"learning_rate": 4.000000000000001e-06,
"logits/chosen": -1.9357620477676392,
"logits/rejected": -1.9375683069229126,
"logps/chosen": -183.8218231201172,
"logps/rejected": -115.80632019042969,
"loss": 0.6768,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.056698836386203766,
"rewards/margins": 0.03413978964090347,
"rewards/rejected": 0.02255905233323574,
"step": 40
},
{
"epoch": 0.2,
"grad_norm": 71.6834716796875,
"learning_rate": 4.996491228070176e-06,
"logits/chosen": -1.7773185968399048,
"logits/rejected": -1.8331743478775024,
"logps/chosen": -167.78533935546875,
"logps/rejected": -227.0582733154297,
"loss": 0.6205,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.1449195146560669,
"rewards/margins": 0.15578031539916992,
"rewards/rejected": -0.010860783979296684,
"step": 50
},
{
"epoch": 0.24,
"grad_norm": 93.83502197265625,
"learning_rate": 4.96140350877193e-06,
"logits/chosen": -1.7739536762237549,
"logits/rejected": -1.8414790630340576,
"logps/chosen": -153.57481384277344,
"logps/rejected": -194.3485870361328,
"loss": 0.5918,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.2464807778596878,
"rewards/margins": 0.22958576679229736,
"rewards/rejected": 0.016895027831196785,
"step": 60
},
{
"epoch": 0.24,
"eval_logits/chosen": -1.7974774837493896,
"eval_logits/rejected": -1.8332849740982056,
"eval_logps/chosen": -193.71743774414062,
"eval_logps/rejected": -243.05783081054688,
"eval_loss": 0.5998325943946838,
"eval_rewards/accuracies": 0.7857142686843872,
"eval_rewards/chosen": 0.24764502048492432,
"eval_rewards/margins": 0.20930658280849457,
"eval_rewards/rejected": 0.03833846375346184,
"eval_runtime": 26.2349,
"eval_samples_per_second": 1.067,
"eval_steps_per_second": 1.067,
"step": 60
},
{
"epoch": 0.29,
"grad_norm": 109.71602630615234,
"learning_rate": 4.926315789473685e-06,
"logits/chosen": -1.8950660228729248,
"logits/rejected": -1.9148075580596924,
"logps/chosen": -169.89028930664062,
"logps/rejected": -177.87301635742188,
"loss": 0.5992,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.15271134674549103,
"rewards/margins": 0.21031668782234192,
"rewards/rejected": -0.0576053187251091,
"step": 70
},
{
"epoch": 0.33,
"grad_norm": 90.3881607055664,
"learning_rate": 4.8912280701754386e-06,
"logits/chosen": -1.798600196838379,
"logits/rejected": -1.855521559715271,
"logps/chosen": -160.3217315673828,
"logps/rejected": -247.6388397216797,
"loss": 0.5551,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.29141485691070557,
"rewards/margins": 0.3369868993759155,
"rewards/rejected": -0.04557197913527489,
"step": 80
},
{
"epoch": 0.37,
"grad_norm": 147.27755737304688,
"learning_rate": 4.856140350877193e-06,
"logits/chosen": -1.8193047046661377,
"logits/rejected": -1.8773219585418701,
"logps/chosen": -155.37123107910156,
"logps/rejected": -245.6790008544922,
"loss": 0.4932,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 0.5351110696792603,
"rewards/margins": 0.5927553176879883,
"rewards/rejected": -0.05764435604214668,
"step": 90
},
{
"epoch": 0.37,
"eval_logits/chosen": -1.8004387617111206,
"eval_logits/rejected": -1.836446762084961,
"eval_logps/chosen": -192.66908264160156,
"eval_logps/rejected": -242.95899963378906,
"eval_loss": 0.5072147846221924,
"eval_rewards/accuracies": 0.8214285969734192,
"eval_rewards/chosen": 0.5621528029441833,
"eval_rewards/margins": 0.4941706657409668,
"eval_rewards/rejected": 0.06798211485147476,
"eval_runtime": 26.2422,
"eval_samples_per_second": 1.067,
"eval_steps_per_second": 1.067,
"step": 90
},
{
"epoch": 0.41,
"grad_norm": 62.75928497314453,
"learning_rate": 4.821052631578948e-06,
"logits/chosen": -1.7835988998413086,
"logits/rejected": -1.8885362148284912,
"logps/chosen": -109.94624328613281,
"logps/rejected": -232.5241241455078,
"loss": 0.4252,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5394761562347412,
"rewards/margins": 0.787738025188446,
"rewards/rejected": -0.248261958360672,
"step": 100
},
{
"epoch": 0.45,
"grad_norm": 84.42039489746094,
"learning_rate": 4.785964912280702e-06,
"logits/chosen": -1.7081111669540405,
"logits/rejected": -1.7244077920913696,
"logps/chosen": -92.51998138427734,
"logps/rejected": -135.5617218017578,
"loss": 0.5077,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.6028972864151001,
"rewards/margins": 0.511993408203125,
"rewards/rejected": 0.09090389311313629,
"step": 110
},
{
"epoch": 0.49,
"grad_norm": 26.61499786376953,
"learning_rate": 4.754385964912281e-06,
"logits/chosen": -1.9250166416168213,
"logits/rejected": -1.9426968097686768,
"logps/chosen": -152.0164337158203,
"logps/rejected": -154.2359161376953,
"loss": 0.4391,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 0.7938548922538757,
"rewards/margins": 0.8368040919303894,
"rewards/rejected": -0.042949117720127106,
"step": 120
},
{
"epoch": 0.49,
"eval_logits/chosen": -1.8051280975341797,
"eval_logits/rejected": -1.8413046598434448,
"eval_logps/chosen": -191.2982177734375,
"eval_logps/rejected": -242.81198120117188,
"eval_loss": 0.4335751235485077,
"eval_rewards/accuracies": 0.7857142686843872,
"eval_rewards/chosen": 0.9734136462211609,
"eval_rewards/margins": 0.8613229990005493,
"eval_rewards/rejected": 0.11209066957235336,
"eval_runtime": 26.2345,
"eval_samples_per_second": 1.067,
"eval_steps_per_second": 1.067,
"step": 120
},
{
"epoch": 0.53,
"grad_norm": 90.6368637084961,
"learning_rate": 4.7192982456140355e-06,
"logits/chosen": -1.865144968032837,
"logits/rejected": -1.9209120273590088,
"logps/chosen": -149.78990173339844,
"logps/rejected": -293.8749084472656,
"loss": 0.4133,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 0.4554510712623596,
"rewards/margins": 0.9285024404525757,
"rewards/rejected": -0.47305139899253845,
"step": 130
},
{
"epoch": 0.57,
"grad_norm": 6.8597941398620605,
"learning_rate": 4.68421052631579e-06,
"logits/chosen": -1.8089487552642822,
"logits/rejected": -1.8955650329589844,
"logps/chosen": -100.26250457763672,
"logps/rejected": -209.1724090576172,
"loss": 0.3414,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9389357566833496,
"rewards/margins": 1.521246314048767,
"rewards/rejected": -0.5823107361793518,
"step": 140
},
{
"epoch": 0.61,
"grad_norm": 79.9189453125,
"learning_rate": 4.652631578947368e-06,
"logits/chosen": -1.760371446609497,
"logits/rejected": -1.8096182346343994,
"logps/chosen": -193.62579345703125,
"logps/rejected": -166.35800170898438,
"loss": 0.3208,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 1.2600924968719482,
"rewards/margins": 1.5845091342926025,
"rewards/rejected": -0.32441645860671997,
"step": 150
},
{
"epoch": 0.61,
"eval_logits/chosen": -1.8129692077636719,
"eval_logits/rejected": -1.849239706993103,
"eval_logps/chosen": -189.88929748535156,
"eval_logps/rejected": -242.91104125976562,
"eval_loss": 0.39326220750808716,
"eval_rewards/accuracies": 0.7857142686843872,
"eval_rewards/chosen": 1.3960883617401123,
"eval_rewards/margins": 1.3137177228927612,
"eval_rewards/rejected": 0.08237060159444809,
"eval_runtime": 26.3372,
"eval_samples_per_second": 1.063,
"eval_steps_per_second": 1.063,
"step": 150
},
{
"epoch": 0.65,
"grad_norm": 3.151426315307617,
"learning_rate": 4.617543859649123e-06,
"logits/chosen": -1.7185112237930298,
"logits/rejected": -1.749427080154419,
"logps/chosen": -178.8656463623047,
"logps/rejected": -201.32858276367188,
"loss": 0.4418,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": 1.3884060382843018,
"rewards/margins": 1.4542597532272339,
"rewards/rejected": -0.06585375964641571,
"step": 160
},
{
"epoch": 0.69,
"grad_norm": 37.79099655151367,
"learning_rate": 4.582456140350878e-06,
"logits/chosen": -1.730787992477417,
"logits/rejected": -1.8099721670150757,
"logps/chosen": -158.50277709960938,
"logps/rejected": -185.48129272460938,
"loss": 0.3912,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": 1.4253065586090088,
"rewards/margins": 1.2418479919433594,
"rewards/rejected": 0.18345877528190613,
"step": 170
},
{
"epoch": 0.73,
"grad_norm": 18.49593734741211,
"learning_rate": 4.547368421052632e-06,
"logits/chosen": -1.842390775680542,
"logits/rejected": -1.8834030628204346,
"logps/chosen": -125.84483337402344,
"logps/rejected": -182.78854370117188,
"loss": 0.3215,
"rewards/accuracies": 0.8999999761581421,
"rewards/chosen": 1.5408955812454224,
"rewards/margins": 1.4677695035934448,
"rewards/rejected": 0.07312598079442978,
"step": 180
},
{
"epoch": 0.73,
"eval_logits/chosen": -1.8194406032562256,
"eval_logits/rejected": -1.8561830520629883,
"eval_logps/chosen": -188.3820037841797,
"eval_logps/rejected": -243.13536071777344,
"eval_loss": 0.3756321966648102,
"eval_rewards/accuracies": 0.7857142686843872,
"eval_rewards/chosen": 1.8482797145843506,
"eval_rewards/margins": 1.8331985473632812,
"eval_rewards/rejected": 0.015081183984875679,
"eval_runtime": 26.2381,
"eval_samples_per_second": 1.067,
"eval_steps_per_second": 1.067,
"step": 180
}
],
"logging_steps": 10,
"max_steps": 1470,
"num_input_tokens_seen": 0,
"num_train_epochs": 6,
"save_steps": 90,
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}
|