Model save
Browse files- README.md +1 -1
- adapter_model.safetensors +1 -1
- all_results.json +4 -4
- runs/May09_02-26-36_sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute/events.out.tfevents.1746757671.sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute.113817.0 +3 -0
- train_results.json +4 -4
- trainer_state.json +1069 -1069
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -27,7 +27,7 @@ print(output["generated_text"])
|
|
| 27 |
|
| 28 |
## Training procedure
|
| 29 |
|
| 30 |
-
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/polyagent/MamayLM-ORPO/runs/
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with ORPO, a method introduced in [ORPO: Monolithic Preference Optimization without Reference Model](https://huggingface.co/papers/2403.07691).
|
|
|
|
| 27 |
|
| 28 |
## Training procedure
|
| 29 |
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/polyagent/MamayLM-ORPO/runs/zb7rmm5o)
|
| 31 |
|
| 32 |
|
| 33 |
This model was trained with ORPO, a method introduced in [ORPO: Monolithic Preference Optimization without Reference Model](https://huggingface.co/papers/2403.07691).
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 8967968
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e96fe9529b4ab46fe5f50ef2921914d16fd9382afd6cf24459607e2d666285f
|
| 3 |
size 8967968
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
-
"train_loss": 1.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 12400,
|
| 7 |
-
"train_samples_per_second": 8.
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.1359047379032259,
|
| 5 |
+
"train_runtime": 1458.2503,
|
| 6 |
"train_samples": 12400,
|
| 7 |
+
"train_samples_per_second": 8.503,
|
| 8 |
+
"train_steps_per_second": 0.531
|
| 9 |
}
|
runs/May09_02-26-36_sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute/events.out.tfevents.1746757671.sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute.113817.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b1bc3e7fb3887ebd0d383e91cc63543b2f16e50b38426f64d2c1435c0d6d5fb8
|
| 3 |
+
size 72143
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
-
"train_loss": 1.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 12400,
|
| 7 |
-
"train_samples_per_second": 8.
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.1359047379032259,
|
| 5 |
+
"train_runtime": 1458.2503,
|
| 6 |
"train_samples": 12400,
|
| 7 |
+
"train_samples_per_second": 8.503,
|
| 8 |
+
"train_steps_per_second": 0.531
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -11,1398 +11,1398 @@
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.012903225806451613,
|
| 14 |
-
"grad_norm": 9.
|
| 15 |
-
"learning_rate":
|
| 16 |
-
"log_odds_chosen": 1.
|
| 17 |
-
"log_odds_ratio": -0.
|
| 18 |
-
"logits/chosen": -1.
|
| 19 |
"logits/rejected": 15.675000190734863,
|
| 20 |
-
"logps/chosen": -1.
|
| 21 |
-
"logps/rejected": -2.
|
| 22 |
-
"loss": 2.
|
| 23 |
-
"nll_loss":
|
| 24 |
"rewards/accuracies": 0.949999988079071,
|
| 25 |
-
"rewards/chosen": -0.
|
| 26 |
-
"rewards/margins": 0.
|
| 27 |
-
"rewards/rejected": -0.
|
| 28 |
"step": 10
|
| 29 |
},
|
| 30 |
{
|
| 31 |
"epoch": 0.025806451612903226,
|
| 32 |
-
"grad_norm": 7.
|
| 33 |
-
"learning_rate":
|
| 34 |
-
"log_odds_chosen": 1.
|
| 35 |
-
"log_odds_ratio": -0.
|
| 36 |
-
"logits/chosen": -1.
|
| 37 |
-
"logits/rejected": 16.
|
| 38 |
-
"logps/chosen": -1.
|
| 39 |
-
"logps/rejected": -2.
|
| 40 |
-
"loss": 2.
|
| 41 |
-
"nll_loss":
|
| 42 |
"rewards/accuracies": 0.9750000238418579,
|
| 43 |
-
"rewards/chosen": -0.
|
| 44 |
-
"rewards/margins": 0.
|
| 45 |
-
"rewards/rejected": -0.
|
| 46 |
"step": 20
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.03870967741935484,
|
| 50 |
-
"grad_norm": 7.
|
| 51 |
-
"learning_rate":
|
| 52 |
-
"log_odds_chosen":
|
| 53 |
-
"log_odds_ratio": -0.
|
| 54 |
-
"logits/chosen": -1.
|
| 55 |
-
"logits/rejected": 16.
|
| 56 |
-
"logps/chosen": -1.
|
| 57 |
-
"logps/rejected": -2.
|
| 58 |
-
"loss":
|
| 59 |
-
"nll_loss":
|
| 60 |
"rewards/accuracies": 0.9375,
|
| 61 |
-
"rewards/chosen": -0.
|
| 62 |
-
"rewards/margins": 0.
|
| 63 |
-
"rewards/rejected": -0.
|
| 64 |
"step": 30
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"epoch": 0.05161290322580645,
|
| 68 |
-
"grad_norm":
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"log_odds_chosen":
|
| 71 |
-
"log_odds_ratio": -0.
|
| 72 |
-
"logits/chosen": -0.
|
| 73 |
-
"logits/rejected": 16.
|
| 74 |
-
"logps/chosen": -1.
|
| 75 |
-
"logps/rejected": -2.
|
| 76 |
-
"loss": 1.
|
| 77 |
-
"nll_loss": 1.
|
| 78 |
"rewards/accuracies": 0.9624999761581421,
|
| 79 |
-
"rewards/chosen": -0.
|
| 80 |
-
"rewards/margins": 0.
|
| 81 |
-
"rewards/rejected": -0.
|
| 82 |
"step": 40
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"epoch": 0.06451612903225806,
|
| 86 |
-
"grad_norm":
|
| 87 |
-
"learning_rate":
|
| 88 |
-
"log_odds_chosen": 0.
|
| 89 |
-
"log_odds_ratio": -0.
|
| 90 |
-
"logits/chosen":
|
| 91 |
-
"logits/rejected": 16.
|
| 92 |
-
"logps/chosen": -1.
|
| 93 |
-
"logps/rejected": -2.
|
| 94 |
-
"loss": 1.
|
| 95 |
-
"nll_loss": 1.
|
| 96 |
-
"rewards/accuracies": 0.
|
| 97 |
-
"rewards/chosen": -0.
|
| 98 |
-
"rewards/margins": 0.
|
| 99 |
-
"rewards/rejected": -0.
|
| 100 |
"step": 50
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.07741935483870968,
|
| 104 |
-
"grad_norm":
|
| 105 |
-
"learning_rate":
|
| 106 |
-
"log_odds_chosen": 0.
|
| 107 |
-
"log_odds_ratio": -0.
|
| 108 |
-
"logits/chosen": 0.
|
| 109 |
-
"logits/rejected": 16.
|
| 110 |
-
"logps/chosen": -1.
|
| 111 |
-
"logps/rejected": -
|
| 112 |
-
"loss": 1.
|
| 113 |
-
"nll_loss": 1.
|
| 114 |
-
"rewards/accuracies": 0.
|
| 115 |
-
"rewards/chosen": -0.
|
| 116 |
-
"rewards/margins": 0.
|
| 117 |
-
"rewards/rejected": -0.
|
| 118 |
"step": 60
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 0.09032258064516129,
|
| 122 |
-
"grad_norm":
|
| 123 |
-
"learning_rate":
|
| 124 |
-
"log_odds_chosen": 0.
|
| 125 |
-
"log_odds_ratio": -0.
|
| 126 |
-
"logits/chosen":
|
| 127 |
-
"logits/rejected": 17.
|
| 128 |
-
"logps/chosen": -1.
|
| 129 |
-
"logps/rejected": -
|
| 130 |
-
"loss": 1.
|
| 131 |
-
"nll_loss": 1.
|
| 132 |
-
"rewards/accuracies": 0.
|
| 133 |
-
"rewards/chosen": -0.
|
| 134 |
-
"rewards/margins": 0.
|
| 135 |
-
"rewards/rejected": -0.
|
| 136 |
"step": 70
|
| 137 |
},
|
| 138 |
{
|
| 139 |
"epoch": 0.1032258064516129,
|
| 140 |
-
"grad_norm":
|
| 141 |
-
"learning_rate":
|
| 142 |
-
"log_odds_chosen": 0.
|
| 143 |
-
"log_odds_ratio": -0.
|
| 144 |
-
"logits/chosen":
|
| 145 |
-
"logits/rejected": 17.
|
| 146 |
-
"logps/chosen": -1.
|
| 147 |
-
"logps/rejected": -1.
|
| 148 |
-
"loss": 1.
|
| 149 |
-
"nll_loss": 1.
|
| 150 |
-
"rewards/accuracies": 0.
|
| 151 |
-
"rewards/chosen": -0.
|
| 152 |
-
"rewards/margins": 0.
|
| 153 |
-
"rewards/rejected": -0.
|
| 154 |
"step": 80
|
| 155 |
},
|
| 156 |
{
|
| 157 |
"epoch": 0.11612903225806452,
|
| 158 |
-
"grad_norm":
|
| 159 |
-
"learning_rate":
|
| 160 |
-
"log_odds_chosen": 0.
|
| 161 |
-
"log_odds_ratio": -0.
|
| 162 |
-
"logits/chosen": 0.
|
| 163 |
-
"logits/rejected": 17.
|
| 164 |
-
"logps/chosen": -1.
|
| 165 |
-
"logps/rejected": -1.
|
| 166 |
-
"loss": 1.
|
| 167 |
-
"nll_loss": 1.
|
| 168 |
-
"rewards/accuracies": 0.
|
| 169 |
-
"rewards/chosen": -0.
|
| 170 |
-
"rewards/margins": 0.
|
| 171 |
-
"rewards/rejected": -0.
|
| 172 |
"step": 90
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"epoch": 0.12903225806451613,
|
| 176 |
-
"grad_norm":
|
| 177 |
-
"learning_rate":
|
| 178 |
-
"log_odds_chosen": 0.
|
| 179 |
-
"log_odds_ratio": -0.
|
| 180 |
-
"logits/chosen": 0.
|
| 181 |
-
"logits/rejected": 16.
|
| 182 |
-
"logps/chosen": -1.
|
| 183 |
-
"logps/rejected": -1.
|
| 184 |
-
"loss": 1.
|
| 185 |
-
"nll_loss": 1.
|
| 186 |
-
"rewards/accuracies": 0.
|
| 187 |
-
"rewards/chosen": -0.
|
| 188 |
-
"rewards/margins": 0.
|
| 189 |
-
"rewards/rejected": -0.
|
| 190 |
"step": 100
|
| 191 |
},
|
| 192 |
{
|
| 193 |
"epoch": 0.14193548387096774,
|
| 194 |
-
"grad_norm": 2.
|
| 195 |
-
"learning_rate":
|
| 196 |
-
"log_odds_chosen": 0.
|
| 197 |
-
"log_odds_ratio": -0.
|
| 198 |
-
"logits/chosen":
|
| 199 |
-
"logits/rejected":
|
| 200 |
-
"logps/chosen": -1.
|
| 201 |
-
"logps/rejected": -1.
|
| 202 |
-
"loss": 1.
|
| 203 |
-
"nll_loss": 1.
|
| 204 |
-
"rewards/accuracies": 0.
|
| 205 |
-
"rewards/chosen": -0.
|
| 206 |
-
"rewards/margins": 0.
|
| 207 |
"rewards/rejected": -0.07038573920726776,
|
| 208 |
"step": 110
|
| 209 |
},
|
| 210 |
{
|
| 211 |
"epoch": 0.15483870967741936,
|
| 212 |
-
"grad_norm":
|
| 213 |
-
"learning_rate":
|
| 214 |
-
"log_odds_chosen": 0.
|
| 215 |
-
"log_odds_ratio": -0.
|
| 216 |
-
"logits/chosen": 0.
|
| 217 |
-
"logits/rejected":
|
| 218 |
-
"logps/chosen": -1.
|
| 219 |
-
"logps/rejected": -1.
|
| 220 |
-
"loss": 1.
|
| 221 |
-
"nll_loss": 1.
|
| 222 |
-
"rewards/accuracies": 0.
|
| 223 |
-
"rewards/chosen": -0.
|
| 224 |
-
"rewards/margins": 0.
|
| 225 |
-
"rewards/rejected": -0.
|
| 226 |
"step": 120
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.16774193548387098,
|
| 230 |
-
"grad_norm": 2.
|
| 231 |
-
"learning_rate":
|
| 232 |
-
"log_odds_chosen":
|
| 233 |
-
"log_odds_ratio": -0.
|
| 234 |
-
"logits/chosen":
|
| 235 |
-
"logits/rejected":
|
| 236 |
-
"logps/chosen": -1.
|
| 237 |
-
"logps/rejected": -
|
| 238 |
-
"loss": 1.
|
| 239 |
-
"nll_loss": 1.
|
| 240 |
-
"rewards/accuracies": 0.
|
| 241 |
-
"rewards/chosen": -0.
|
| 242 |
-
"rewards/margins": 0.
|
| 243 |
-
"rewards/rejected": -0.
|
| 244 |
"step": 130
|
| 245 |
},
|
| 246 |
{
|
| 247 |
"epoch": 0.18064516129032257,
|
| 248 |
-
"grad_norm": 2.
|
| 249 |
-
"learning_rate":
|
| 250 |
-
"log_odds_chosen":
|
| 251 |
-
"log_odds_ratio": -0.
|
| 252 |
-
"logits/chosen": 0.
|
| 253 |
-
"logits/rejected":
|
| 254 |
-
"logps/chosen": -1.
|
| 255 |
-
"logps/rejected": -
|
| 256 |
-
"loss": 1.
|
| 257 |
-
"nll_loss": 1.
|
| 258 |
-
"rewards/accuracies": 0.
|
| 259 |
-
"rewards/chosen": -0.
|
| 260 |
-
"rewards/margins": 0.
|
| 261 |
-
"rewards/rejected": -0.
|
| 262 |
"step": 140
|
| 263 |
},
|
| 264 |
{
|
| 265 |
"epoch": 0.1935483870967742,
|
| 266 |
-
"grad_norm": 2.
|
| 267 |
-
"learning_rate":
|
| 268 |
-
"log_odds_chosen":
|
| 269 |
-
"log_odds_ratio": -0.
|
| 270 |
-
"logits/chosen": -
|
| 271 |
-
"logits/rejected":
|
| 272 |
-
"logps/chosen": -1.
|
| 273 |
-
"logps/rejected": -
|
| 274 |
-
"loss": 1.
|
| 275 |
-
"nll_loss": 1.
|
| 276 |
-
"rewards/accuracies": 0.
|
| 277 |
-
"rewards/chosen": -0.
|
| 278 |
-
"rewards/margins": 0.
|
| 279 |
-
"rewards/rejected": -0.
|
| 280 |
"step": 150
|
| 281 |
},
|
| 282 |
{
|
| 283 |
"epoch": 0.2064516129032258,
|
| 284 |
-
"grad_norm": 3.
|
| 285 |
-
"learning_rate":
|
| 286 |
-
"log_odds_chosen":
|
| 287 |
-
"log_odds_ratio": -0.
|
| 288 |
-
"logits/chosen": -1.
|
| 289 |
-
"logits/rejected": 15.
|
| 290 |
-
"logps/chosen": -1.
|
| 291 |
-
"logps/rejected": -
|
| 292 |
-
"loss": 1.
|
| 293 |
-
"nll_loss": 1.
|
| 294 |
-
"rewards/accuracies": 0.
|
| 295 |
-
"rewards/chosen": -0.
|
| 296 |
-
"rewards/margins": 0.
|
| 297 |
-
"rewards/rejected": -0.
|
| 298 |
"step": 160
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 0.21935483870967742,
|
| 302 |
-
"grad_norm": 2.
|
| 303 |
-
"learning_rate":
|
| 304 |
-
"log_odds_chosen":
|
| 305 |
-
"log_odds_ratio": -0.
|
| 306 |
-
"logits/chosen": -1.
|
| 307 |
-
"logits/rejected": 16.
|
| 308 |
-
"logps/chosen": -1.
|
| 309 |
-
"logps/rejected": -
|
| 310 |
-
"loss": 1.
|
| 311 |
-
"nll_loss": 1.
|
| 312 |
-
"rewards/accuracies": 0.
|
| 313 |
-
"rewards/chosen": -0.
|
| 314 |
-
"rewards/margins": 0.
|
| 315 |
-
"rewards/rejected": -0.
|
| 316 |
"step": 170
|
| 317 |
},
|
| 318 |
{
|
| 319 |
"epoch": 0.23225806451612904,
|
| 320 |
-
"grad_norm":
|
| 321 |
-
"learning_rate":
|
| 322 |
-
"log_odds_chosen":
|
| 323 |
-
"log_odds_ratio": -0.
|
| 324 |
-
"logits/chosen": -1.
|
| 325 |
-
"logits/rejected": 15.
|
| 326 |
-
"logps/chosen": -1.
|
| 327 |
-
"logps/rejected": -
|
| 328 |
-
"loss": 1.
|
| 329 |
-
"nll_loss": 1.
|
| 330 |
-
"rewards/accuracies": 0.
|
| 331 |
-
"rewards/chosen": -0.
|
| 332 |
-
"rewards/margins": 0.
|
| 333 |
-
"rewards/rejected": -0.
|
| 334 |
"step": 180
|
| 335 |
},
|
| 336 |
{
|
| 337 |
"epoch": 0.24516129032258063,
|
| 338 |
-
"grad_norm":
|
| 339 |
-
"learning_rate":
|
| 340 |
-
"log_odds_chosen":
|
| 341 |
-
"log_odds_ratio": -0.
|
| 342 |
-
"logits/chosen": -0.
|
| 343 |
-
"logits/rejected": 16.
|
| 344 |
-
"logps/chosen": -1.
|
| 345 |
-
"logps/rejected": -
|
| 346 |
-
"loss": 1.
|
| 347 |
-
"nll_loss": 1.
|
| 348 |
-
"rewards/accuracies": 0.
|
| 349 |
-
"rewards/chosen": -0.
|
| 350 |
-
"rewards/margins": 0.
|
| 351 |
-
"rewards/rejected": -0.
|
| 352 |
"step": 190
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.25806451612903225,
|
| 356 |
-
"grad_norm": 2.
|
| 357 |
-
"learning_rate":
|
| 358 |
-
"log_odds_chosen": 1.
|
| 359 |
-
"log_odds_ratio": -0.
|
| 360 |
-
"logits/chosen": -1.
|
| 361 |
-
"logits/rejected": 16.
|
| 362 |
-
"logps/chosen": -1.
|
| 363 |
-
"logps/rejected": -2.
|
| 364 |
-
"loss": 1.
|
| 365 |
-
"nll_loss": 1.
|
| 366 |
-
"rewards/accuracies": 0.
|
| 367 |
-
"rewards/chosen": -0.
|
| 368 |
-
"rewards/margins": 0.
|
| 369 |
-
"rewards/rejected": -0.
|
| 370 |
"step": 200
|
| 371 |
},
|
| 372 |
{
|
| 373 |
"epoch": 0.2709677419354839,
|
| 374 |
-
"grad_norm": 2.
|
| 375 |
-
"learning_rate":
|
| 376 |
-
"log_odds_chosen": 1.
|
| 377 |
-
"log_odds_ratio": -0.
|
| 378 |
-
"logits/chosen": -1.
|
| 379 |
-
"logits/rejected": 15.
|
| 380 |
-
"logps/chosen": -1.
|
| 381 |
-
"logps/rejected": -2.
|
| 382 |
-
"loss": 1.
|
| 383 |
-
"nll_loss": 1.
|
| 384 |
-
"rewards/accuracies": 0.
|
| 385 |
-
"rewards/chosen": -0.
|
| 386 |
-
"rewards/margins": 0.
|
| 387 |
-
"rewards/rejected": -0.
|
| 388 |
"step": 210
|
| 389 |
},
|
| 390 |
{
|
| 391 |
"epoch": 0.2838709677419355,
|
| 392 |
-
"grad_norm": 2.
|
| 393 |
-
"learning_rate":
|
| 394 |
-
"log_odds_chosen": 1.
|
| 395 |
-
"log_odds_ratio": -0.
|
| 396 |
-
"logits/chosen": -0.
|
| 397 |
-
"logits/rejected": 16.
|
| 398 |
-
"logps/chosen": -1.
|
| 399 |
-
"logps/rejected": -2.
|
| 400 |
-
"loss": 1.
|
| 401 |
-
"nll_loss": 1.
|
| 402 |
-
"rewards/accuracies": 0.
|
| 403 |
-
"rewards/chosen": -0.
|
| 404 |
-
"rewards/margins": 0.
|
| 405 |
-
"rewards/rejected": -0.
|
| 406 |
"step": 220
|
| 407 |
},
|
| 408 |
{
|
| 409 |
"epoch": 0.2967741935483871,
|
| 410 |
-
"grad_norm": 2.
|
| 411 |
-
"learning_rate":
|
| 412 |
-
"log_odds_chosen": 1.
|
| 413 |
-
"log_odds_ratio": -0.
|
| 414 |
-
"logits/chosen": -0.
|
| 415 |
-
"logits/rejected": 15.
|
| 416 |
-
"logps/chosen": -1.
|
| 417 |
-
"logps/rejected": -2.
|
| 418 |
-
"loss": 1.
|
| 419 |
-
"nll_loss": 1.
|
| 420 |
"rewards/accuracies": 0.9750000238418579,
|
| 421 |
-
"rewards/chosen": -0.
|
| 422 |
-
"rewards/margins": 0.
|
| 423 |
-
"rewards/rejected": -0.
|
| 424 |
"step": 230
|
| 425 |
},
|
| 426 |
{
|
| 427 |
"epoch": 0.3096774193548387,
|
| 428 |
-
"grad_norm": 2.
|
| 429 |
-
"learning_rate":
|
| 430 |
-
"log_odds_chosen": 1.
|
| 431 |
-
"log_odds_ratio": -0.
|
| 432 |
-
"logits/chosen": -1.
|
| 433 |
-
"logits/rejected": 14.
|
| 434 |
-
"logps/chosen": -1.
|
| 435 |
-
"logps/rejected": -2.
|
| 436 |
-
"loss": 1.
|
| 437 |
-
"nll_loss": 1.
|
| 438 |
-
"rewards/accuracies": 0.
|
| 439 |
-
"rewards/chosen": -0.
|
| 440 |
-
"rewards/margins": 0.
|
| 441 |
-
"rewards/rejected": -0.
|
| 442 |
"step": 240
|
| 443 |
},
|
| 444 |
{
|
| 445 |
"epoch": 0.3225806451612903,
|
| 446 |
-
"grad_norm": 2.
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"log_odds_chosen":
|
| 449 |
-
"log_odds_ratio": -0.
|
| 450 |
-
"logits/chosen": -2.
|
| 451 |
-
"logits/rejected": 15.
|
| 452 |
-
"logps/chosen": -1.
|
| 453 |
-
"logps/rejected": -2.
|
| 454 |
-
"loss": 1.
|
| 455 |
-
"nll_loss": 1.
|
| 456 |
-
"rewards/accuracies": 0.
|
| 457 |
-
"rewards/chosen": -0.
|
| 458 |
-
"rewards/margins": 0.
|
| 459 |
-
"rewards/rejected": -0.
|
| 460 |
"step": 250
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 0.33548387096774196,
|
| 464 |
-
"grad_norm": 2.
|
| 465 |
-
"learning_rate":
|
| 466 |
-
"log_odds_chosen": 1.
|
| 467 |
-
"log_odds_ratio": -0.
|
| 468 |
-
"logits/chosen": -
|
| 469 |
-
"logits/rejected": 15.
|
| 470 |
-
"logps/chosen": -1.
|
| 471 |
-
"logps/rejected": -2.
|
| 472 |
-
"loss": 1.
|
| 473 |
-
"nll_loss": 1.
|
| 474 |
-
"rewards/accuracies": 0.
|
| 475 |
-
"rewards/chosen": -0.
|
| 476 |
-
"rewards/margins": 0.
|
| 477 |
-
"rewards/rejected": -0.
|
| 478 |
"step": 260
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.34838709677419355,
|
| 482 |
-
"grad_norm": 2.
|
| 483 |
-
"learning_rate":
|
| 484 |
-
"log_odds_chosen":
|
| 485 |
-
"log_odds_ratio": -0.
|
| 486 |
-
"logits/chosen": -1.
|
| 487 |
-
"logits/rejected": 15.
|
| 488 |
-
"logps/chosen": -1.
|
| 489 |
-
"logps/rejected": -2.
|
| 490 |
-
"loss": 1.
|
| 491 |
-
"nll_loss": 1.
|
| 492 |
-
"rewards/accuracies": 0.
|
| 493 |
-
"rewards/chosen": -0.
|
| 494 |
-
"rewards/margins": 0.
|
| 495 |
-
"rewards/rejected": -0.
|
| 496 |
"step": 270
|
| 497 |
},
|
| 498 |
{
|
| 499 |
"epoch": 0.36129032258064514,
|
| 500 |
-
"grad_norm": 2.
|
| 501 |
-
"learning_rate":
|
| 502 |
-
"log_odds_chosen": 2.
|
| 503 |
-
"log_odds_ratio": -0.
|
| 504 |
-
"logits/chosen": -1.
|
| 505 |
-
"logits/rejected": 15.
|
| 506 |
-
"logps/chosen": -1.
|
| 507 |
-
"logps/rejected": -2.
|
| 508 |
-
"loss": 1.
|
| 509 |
-
"nll_loss": 1.
|
| 510 |
-
"rewards/accuracies": 0.
|
| 511 |
-
"rewards/chosen": -0.
|
| 512 |
-
"rewards/margins": 0.
|
| 513 |
-
"rewards/rejected": -0.
|
| 514 |
"step": 280
|
| 515 |
},
|
| 516 |
{
|
| 517 |
"epoch": 0.3741935483870968,
|
| 518 |
-
"grad_norm":
|
| 519 |
-
"learning_rate":
|
| 520 |
-
"log_odds_chosen": 2.
|
| 521 |
-
"log_odds_ratio": -0.
|
| 522 |
-
"logits/chosen": -1.
|
| 523 |
-
"logits/rejected": 15.
|
| 524 |
-
"logps/chosen": -1.
|
| 525 |
-
"logps/rejected": -2.
|
| 526 |
-
"loss": 1.
|
| 527 |
-
"nll_loss": 1.
|
| 528 |
-
"rewards/accuracies": 0.
|
| 529 |
-
"rewards/chosen": -0.
|
| 530 |
-
"rewards/margins": 0.
|
| 531 |
-
"rewards/rejected": -0.
|
| 532 |
"step": 290
|
| 533 |
},
|
| 534 |
{
|
| 535 |
"epoch": 0.3870967741935484,
|
| 536 |
-
"grad_norm": 3.
|
| 537 |
-
"learning_rate":
|
| 538 |
-
"log_odds_chosen": 2.
|
| 539 |
-
"log_odds_ratio": -0.
|
| 540 |
-
"logits/chosen": -1.
|
| 541 |
-
"logits/rejected": 15.
|
| 542 |
-
"logps/chosen": -1.
|
| 543 |
-
"logps/rejected": -2.
|
| 544 |
-
"loss": 1.
|
| 545 |
-
"nll_loss": 1.
|
| 546 |
-
"rewards/accuracies": 0.
|
| 547 |
-
"rewards/chosen": -0.
|
| 548 |
-
"rewards/margins": 0.
|
| 549 |
-
"rewards/rejected": -0.
|
| 550 |
"step": 300
|
| 551 |
},
|
| 552 |
{
|
| 553 |
"epoch": 0.4,
|
| 554 |
-
"grad_norm": 2.
|
| 555 |
-
"learning_rate":
|
| 556 |
-
"log_odds_chosen": 2.
|
| 557 |
-
"log_odds_ratio": -0.
|
| 558 |
-
"logits/chosen": -1.
|
| 559 |
-
"logits/rejected": 15.
|
| 560 |
-
"logps/chosen": -1.
|
| 561 |
-
"logps/rejected": -2.
|
| 562 |
-
"loss": 1.
|
| 563 |
-
"nll_loss": 1.
|
| 564 |
-
"rewards/accuracies": 0.
|
| 565 |
-
"rewards/chosen": -0.
|
| 566 |
-
"rewards/margins": 0.
|
| 567 |
-
"rewards/rejected": -0.
|
| 568 |
"step": 310
|
| 569 |
},
|
| 570 |
{
|
| 571 |
"epoch": 0.4129032258064516,
|
| 572 |
-
"grad_norm": 2.
|
| 573 |
-
"learning_rate":
|
| 574 |
-
"log_odds_chosen": 2.
|
| 575 |
-
"log_odds_ratio": -0.
|
| 576 |
-
"logits/chosen": -1.
|
| 577 |
-
"logits/rejected": 15.
|
| 578 |
-
"logps/chosen": -1.
|
| 579 |
-
"logps/rejected": -2.
|
| 580 |
-
"loss": 1.
|
| 581 |
-
"nll_loss": 1.
|
| 582 |
-
"rewards/accuracies": 0.
|
| 583 |
-
"rewards/chosen": -0.
|
| 584 |
-
"rewards/margins": 0.
|
| 585 |
-
"rewards/rejected": -0.
|
| 586 |
"step": 320
|
| 587 |
},
|
| 588 |
{
|
| 589 |
"epoch": 0.4258064516129032,
|
| 590 |
-
"grad_norm": 3.
|
| 591 |
-
"learning_rate":
|
| 592 |
-
"log_odds_chosen": 2.
|
| 593 |
-
"log_odds_ratio": -0.
|
| 594 |
-
"logits/chosen": -1.
|
| 595 |
-
"logits/rejected": 15.
|
| 596 |
-
"logps/chosen": -1.
|
| 597 |
-
"logps/rejected": -3.
|
| 598 |
-
"loss": 1.
|
| 599 |
-
"nll_loss": 1.
|
| 600 |
-
"rewards/accuracies": 0.
|
| 601 |
-
"rewards/chosen": -0.
|
| 602 |
-
"rewards/margins": 0.
|
| 603 |
-
"rewards/rejected": -0.
|
| 604 |
"step": 330
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.43870967741935485,
|
| 608 |
-
"grad_norm":
|
| 609 |
-
"learning_rate":
|
| 610 |
-
"log_odds_chosen": 2.
|
| 611 |
-
"log_odds_ratio": -0.
|
| 612 |
-
"logits/chosen": -1.
|
| 613 |
-
"logits/rejected": 15.
|
| 614 |
-
"logps/chosen": -1.
|
| 615 |
-
"logps/rejected": -3.
|
| 616 |
-
"loss": 1.
|
| 617 |
-
"nll_loss": 1.
|
| 618 |
-
"rewards/accuracies": 0.
|
| 619 |
-
"rewards/chosen": -0.
|
| 620 |
-
"rewards/margins": 0.
|
| 621 |
-
"rewards/rejected": -0.
|
| 622 |
"step": 340
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 0.45161290322580644,
|
| 626 |
-
"grad_norm": 2.
|
| 627 |
-
"learning_rate":
|
| 628 |
-
"log_odds_chosen": 2.
|
| 629 |
-
"log_odds_ratio": -0.
|
| 630 |
-
"logits/chosen": -1.
|
| 631 |
-
"logits/rejected": 15.
|
| 632 |
-
"logps/chosen": -1.
|
| 633 |
-
"logps/rejected": -3.
|
| 634 |
-
"loss": 1.
|
| 635 |
-
"nll_loss": 1.
|
| 636 |
-
"rewards/accuracies": 0.
|
| 637 |
-
"rewards/chosen": -0.
|
| 638 |
-
"rewards/margins": 0.
|
| 639 |
-
"rewards/rejected": -0.
|
| 640 |
"step": 350
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 0.4645161290322581,
|
| 644 |
-
"grad_norm": 2.
|
| 645 |
-
"learning_rate":
|
| 646 |
-
"log_odds_chosen": 2.
|
| 647 |
-
"log_odds_ratio": -0.
|
| 648 |
-
"logits/chosen": -1.
|
| 649 |
-
"logits/rejected": 15.
|
| 650 |
-
"logps/chosen": -1.
|
| 651 |
-
"logps/rejected": -3.
|
| 652 |
-
"loss": 1.
|
| 653 |
-
"nll_loss": 1.
|
| 654 |
-
"rewards/accuracies": 0.
|
| 655 |
-
"rewards/chosen": -0.
|
| 656 |
-
"rewards/margins": 0.
|
| 657 |
-
"rewards/rejected": -0.
|
| 658 |
"step": 360
|
| 659 |
},
|
| 660 |
{
|
| 661 |
"epoch": 0.4774193548387097,
|
| 662 |
-
"grad_norm":
|
| 663 |
-
"learning_rate":
|
| 664 |
-
"log_odds_chosen": 2.
|
| 665 |
-
"log_odds_ratio": -0.
|
| 666 |
-
"logits/chosen": -1.
|
| 667 |
-
"logits/rejected": 15.
|
| 668 |
-
"logps/chosen": -1.
|
| 669 |
-
"logps/rejected": -3.
|
| 670 |
-
"loss": 1.
|
| 671 |
-
"nll_loss": 1.
|
| 672 |
-
"rewards/accuracies": 0.
|
| 673 |
-
"rewards/chosen": -0.
|
| 674 |
-
"rewards/margins": 0.
|
| 675 |
-
"rewards/rejected": -0.
|
| 676 |
"step": 370
|
| 677 |
},
|
| 678 |
{
|
| 679 |
"epoch": 0.49032258064516127,
|
| 680 |
-
"grad_norm": 2.
|
| 681 |
-
"learning_rate":
|
| 682 |
-
"log_odds_chosen": 2.
|
| 683 |
-
"log_odds_ratio": -0.
|
| 684 |
-
"logits/chosen": -1.
|
| 685 |
-
"logits/rejected": 15.
|
| 686 |
-
"logps/chosen": -1.
|
| 687 |
-
"logps/rejected": -2.
|
| 688 |
-
"loss": 1.
|
| 689 |
-
"nll_loss": 1.
|
| 690 |
-
"rewards/accuracies": 0.
|
| 691 |
-
"rewards/chosen": -0.
|
| 692 |
-
"rewards/margins": 0.
|
| 693 |
-
"rewards/rejected": -0.
|
| 694 |
"step": 380
|
| 695 |
},
|
| 696 |
{
|
| 697 |
"epoch": 0.5032258064516129,
|
| 698 |
-
"grad_norm":
|
| 699 |
-
"learning_rate":
|
| 700 |
-
"log_odds_chosen": 2.
|
| 701 |
-
"log_odds_ratio": -0.
|
| 702 |
-
"logits/chosen": -1.
|
| 703 |
-
"logits/rejected": 15.
|
| 704 |
-
"logps/chosen": -1.
|
| 705 |
-
"logps/rejected": -
|
| 706 |
-
"loss": 1.
|
| 707 |
-
"nll_loss": 1.
|
| 708 |
-
"rewards/accuracies": 0.
|
| 709 |
-
"rewards/chosen": -0.
|
| 710 |
-
"rewards/margins": 0.
|
| 711 |
-
"rewards/rejected": -0.
|
| 712 |
"step": 390
|
| 713 |
},
|
| 714 |
{
|
| 715 |
"epoch": 0.5161290322580645,
|
| 716 |
-
"grad_norm": 2.
|
| 717 |
-
"learning_rate":
|
| 718 |
-
"log_odds_chosen": 2.
|
| 719 |
-
"log_odds_ratio": -0.
|
| 720 |
-
"logits/chosen": -
|
| 721 |
-
"logits/rejected": 15.
|
| 722 |
-
"logps/chosen": -1.
|
| 723 |
-
"logps/rejected": -
|
| 724 |
-
"loss": 1.
|
| 725 |
-
"nll_loss": 1.
|
| 726 |
-
"rewards/accuracies": 0.
|
| 727 |
-
"rewards/chosen": -0.
|
| 728 |
-
"rewards/margins": 0.
|
| 729 |
-
"rewards/rejected": -0.
|
| 730 |
"step": 400
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.5290322580645161,
|
| 734 |
-
"grad_norm": 2.
|
| 735 |
-
"learning_rate":
|
| 736 |
-
"log_odds_chosen": 2.
|
| 737 |
-
"log_odds_ratio": -0.
|
| 738 |
-
"logits/chosen": -1.
|
| 739 |
-
"logits/rejected": 15.
|
| 740 |
-
"logps/chosen": -1.
|
| 741 |
-
"logps/rejected": -3.
|
| 742 |
-
"loss": 1.
|
| 743 |
-
"nll_loss": 1.
|
| 744 |
-
"rewards/accuracies": 0.
|
| 745 |
-
"rewards/chosen": -0.
|
| 746 |
-
"rewards/margins": 0.
|
| 747 |
-
"rewards/rejected": -0.
|
| 748 |
"step": 410
|
| 749 |
},
|
| 750 |
{
|
| 751 |
"epoch": 0.5419354838709678,
|
| 752 |
-
"grad_norm":
|
| 753 |
-
"learning_rate":
|
| 754 |
-
"log_odds_chosen": 2.
|
| 755 |
-
"log_odds_ratio": -0.
|
| 756 |
-
"logits/chosen": -1.
|
| 757 |
-
"logits/rejected": 15.
|
| 758 |
-
"logps/chosen": -1.
|
| 759 |
-
"logps/rejected": -3.
|
| 760 |
-
"loss": 1.
|
| 761 |
-
"nll_loss": 1.
|
| 762 |
-
"rewards/accuracies": 0.
|
| 763 |
-
"rewards/chosen": -0.
|
| 764 |
-
"rewards/margins": 0.
|
| 765 |
-
"rewards/rejected": -0.
|
| 766 |
"step": 420
|
| 767 |
},
|
| 768 |
{
|
| 769 |
"epoch": 0.5548387096774193,
|
| 770 |
-
"grad_norm":
|
| 771 |
-
"learning_rate":
|
| 772 |
-
"log_odds_chosen": 2.
|
| 773 |
-
"log_odds_ratio": -0.
|
| 774 |
-
"logits/chosen": -0.
|
| 775 |
-
"logits/rejected": 16.
|
| 776 |
-
"logps/chosen": -1.
|
| 777 |
-
"logps/rejected": -3.
|
| 778 |
-
"loss": 1.
|
| 779 |
-
"nll_loss": 1.
|
| 780 |
-
"rewards/accuracies": 0.
|
| 781 |
-
"rewards/chosen": -0.
|
| 782 |
-
"rewards/margins": 0.
|
| 783 |
-
"rewards/rejected": -0.
|
| 784 |
"step": 430
|
| 785 |
},
|
| 786 |
{
|
| 787 |
"epoch": 0.567741935483871,
|
| 788 |
-
"grad_norm": 3.
|
| 789 |
-
"learning_rate":
|
| 790 |
-
"log_odds_chosen": 2.
|
| 791 |
-
"log_odds_ratio": -0.
|
| 792 |
-
"logits/chosen": -1.
|
| 793 |
-
"logits/rejected": 15.
|
| 794 |
-
"logps/chosen": -1.
|
| 795 |
-
"logps/rejected": -
|
| 796 |
-
"loss": 1.
|
| 797 |
-
"nll_loss": 1.
|
| 798 |
-
"rewards/accuracies": 0.
|
| 799 |
-
"rewards/chosen": -0.
|
| 800 |
-
"rewards/margins": 0.
|
| 801 |
-
"rewards/rejected": -0.
|
| 802 |
"step": 440
|
| 803 |
},
|
| 804 |
{
|
| 805 |
"epoch": 0.5806451612903226,
|
| 806 |
-
"grad_norm": 2.
|
| 807 |
-
"learning_rate":
|
| 808 |
-
"log_odds_chosen": 2.
|
| 809 |
-
"log_odds_ratio": -0.
|
| 810 |
-
"logits/chosen": -1.
|
| 811 |
-
"logits/rejected": 15.
|
| 812 |
-
"logps/chosen": -1.
|
| 813 |
-
"logps/rejected": -
|
| 814 |
-
"loss": 1.
|
| 815 |
-
"nll_loss": 1.
|
| 816 |
-
"rewards/accuracies": 0.
|
| 817 |
-
"rewards/chosen": -0.
|
| 818 |
-
"rewards/margins": 0.
|
| 819 |
-
"rewards/rejected": -0.
|
| 820 |
"step": 450
|
| 821 |
},
|
| 822 |
{
|
| 823 |
"epoch": 0.5935483870967742,
|
| 824 |
-
"grad_norm":
|
| 825 |
-
"learning_rate":
|
| 826 |
-
"log_odds_chosen":
|
| 827 |
-
"log_odds_ratio": -0.
|
| 828 |
-
"logits/chosen": -1.
|
| 829 |
-
"logits/rejected": 15.
|
| 830 |
-
"logps/chosen": -1.
|
| 831 |
-
"logps/rejected": -3.
|
| 832 |
-
"loss": 1.
|
| 833 |
-
"nll_loss": 1.
|
| 834 |
-
"rewards/accuracies": 0.
|
| 835 |
-
"rewards/chosen": -0.
|
| 836 |
-
"rewards/margins": 0.
|
| 837 |
-
"rewards/rejected": -0.
|
| 838 |
"step": 460
|
| 839 |
},
|
| 840 |
{
|
| 841 |
"epoch": 0.6064516129032258,
|
| 842 |
-
"grad_norm": 3.
|
| 843 |
-
"learning_rate":
|
| 844 |
-
"log_odds_chosen": 2.
|
| 845 |
-
"log_odds_ratio": -0.
|
| 846 |
-
"logits/chosen": -1.
|
| 847 |
-
"logits/rejected": 15.
|
| 848 |
-
"logps/chosen": -1.
|
| 849 |
-
"logps/rejected": -3.
|
| 850 |
-
"loss": 1.
|
| 851 |
-
"nll_loss": 1.
|
| 852 |
-
"rewards/accuracies": 0.
|
| 853 |
-
"rewards/chosen": -0.
|
| 854 |
-
"rewards/margins": 0.
|
| 855 |
-
"rewards/rejected": -0.
|
| 856 |
"step": 470
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.6193548387096774,
|
| 860 |
-
"grad_norm":
|
| 861 |
-
"learning_rate":
|
| 862 |
-
"log_odds_chosen":
|
| 863 |
-
"log_odds_ratio": -0.
|
| 864 |
-
"logits/chosen": -1.
|
| 865 |
-
"logits/rejected": 15.
|
| 866 |
-
"logps/chosen": -0.
|
| 867 |
-
"logps/rejected": -3.
|
| 868 |
-
"loss":
|
| 869 |
-
"nll_loss":
|
| 870 |
-
"rewards/accuracies": 0.
|
| 871 |
-
"rewards/chosen": -0.
|
| 872 |
-
"rewards/margins": 0.
|
| 873 |
-
"rewards/rejected": -0.
|
| 874 |
"step": 480
|
| 875 |
},
|
| 876 |
{
|
| 877 |
"epoch": 0.632258064516129,
|
| 878 |
-
"grad_norm": 3.
|
| 879 |
-
"learning_rate":
|
| 880 |
-
"log_odds_chosen": 2.
|
| 881 |
-
"log_odds_ratio": -0.
|
| 882 |
-
"logits/chosen": -1.
|
| 883 |
-
"logits/rejected": 16.
|
| 884 |
-
"logps/chosen": -1.
|
| 885 |
-
"logps/rejected": -3.
|
| 886 |
-
"loss": 1.
|
| 887 |
-
"nll_loss": 1.
|
| 888 |
-
"rewards/accuracies": 0.
|
| 889 |
-
"rewards/chosen": -0.
|
| 890 |
-
"rewards/margins": 0.
|
| 891 |
-
"rewards/rejected": -0.
|
| 892 |
"step": 490
|
| 893 |
},
|
| 894 |
{
|
| 895 |
"epoch": 0.6451612903225806,
|
| 896 |
-
"grad_norm":
|
| 897 |
-
"learning_rate":
|
| 898 |
-
"log_odds_chosen": 2.
|
| 899 |
-
"log_odds_ratio": -0.
|
| 900 |
-
"logits/chosen": -1.
|
| 901 |
-
"logits/rejected": 15.
|
| 902 |
-
"logps/chosen": -1.
|
| 903 |
-
"logps/rejected": -3.
|
| 904 |
-
"loss": 1.
|
| 905 |
-
"nll_loss": 1.
|
| 906 |
-
"rewards/accuracies": 0.
|
| 907 |
-
"rewards/chosen": -0.
|
| 908 |
-
"rewards/margins": 0.
|
| 909 |
-
"rewards/rejected": -0.
|
| 910 |
"step": 500
|
| 911 |
},
|
| 912 |
{
|
| 913 |
"epoch": 0.6580645161290323,
|
| 914 |
-
"grad_norm": 2.
|
| 915 |
-
"learning_rate":
|
| 916 |
-
"log_odds_chosen":
|
| 917 |
-
"log_odds_ratio": -0.
|
| 918 |
-
"logits/chosen": -
|
| 919 |
-
"logits/rejected": 15.
|
| 920 |
-
"logps/chosen": -
|
| 921 |
-
"logps/rejected": -3.
|
| 922 |
-
"loss": 1.
|
| 923 |
-
"nll_loss": 1.
|
| 924 |
-
"rewards/accuracies": 0.
|
| 925 |
-
"rewards/chosen": -0.
|
| 926 |
-
"rewards/margins": 0.
|
| 927 |
-
"rewards/rejected": -0.
|
| 928 |
"step": 510
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 0.6709677419354839,
|
| 932 |
-
"grad_norm":
|
| 933 |
-
"learning_rate":
|
| 934 |
-
"log_odds_chosen":
|
| 935 |
-
"log_odds_ratio": -0.
|
| 936 |
-
"logits/chosen": -
|
| 937 |
-
"logits/rejected": 15.
|
| 938 |
-
"logps/chosen": -1.
|
| 939 |
-
"logps/rejected": -3.
|
| 940 |
-
"loss": 1.
|
| 941 |
-
"nll_loss": 1.
|
| 942 |
-
"rewards/accuracies": 0.
|
| 943 |
-
"rewards/chosen": -0.
|
| 944 |
-
"rewards/margins": 0.
|
| 945 |
-
"rewards/rejected": -0.
|
| 946 |
"step": 520
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 0.6838709677419355,
|
| 950 |
-
"grad_norm": 2.
|
| 951 |
-
"learning_rate":
|
| 952 |
-
"log_odds_chosen":
|
| 953 |
-
"log_odds_ratio": -0.
|
| 954 |
-
"logits/chosen": -0.
|
| 955 |
-
"logits/rejected": 15.
|
| 956 |
-
"logps/chosen": -
|
| 957 |
-
"logps/rejected": -3.
|
| 958 |
-
"loss": 1.
|
| 959 |
-
"nll_loss": 1.
|
| 960 |
-
"rewards/accuracies": 0.
|
| 961 |
-
"rewards/chosen": -0.
|
| 962 |
-
"rewards/margins": 0.
|
| 963 |
-
"rewards/rejected": -0.
|
| 964 |
"step": 530
|
| 965 |
},
|
| 966 |
{
|
| 967 |
"epoch": 0.6967741935483871,
|
| 968 |
-
"grad_norm": 2.
|
| 969 |
-
"learning_rate":
|
| 970 |
-
"log_odds_chosen":
|
| 971 |
-
"log_odds_ratio": -0.
|
| 972 |
-
"logits/chosen": -
|
| 973 |
-
"logits/rejected": 16.
|
| 974 |
-
"logps/chosen": -1.
|
| 975 |
-
"logps/rejected": -3.
|
| 976 |
-
"loss": 1.
|
| 977 |
-
"nll_loss": 1.
|
| 978 |
-
"rewards/accuracies": 0.
|
| 979 |
-
"rewards/chosen": -0.
|
| 980 |
-
"rewards/margins": 0.
|
| 981 |
-
"rewards/rejected": -0.
|
| 982 |
"step": 540
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.7096774193548387,
|
| 986 |
-
"grad_norm": 3.
|
| 987 |
-
"learning_rate":
|
| 988 |
-
"log_odds_chosen":
|
| 989 |
-
"log_odds_ratio": -0.
|
| 990 |
-
"logits/chosen": -1.
|
| 991 |
-
"logits/rejected": 15.
|
| 992 |
-
"logps/chosen": -1.
|
| 993 |
-
"logps/rejected": -3.
|
| 994 |
-
"loss": 1.
|
| 995 |
-
"nll_loss": 1.
|
| 996 |
"rewards/accuracies": 0.987500011920929,
|
| 997 |
-
"rewards/chosen": -0.
|
| 998 |
-
"rewards/margins": 0.
|
| 999 |
-
"rewards/rejected": -0.
|
| 1000 |
"step": 550
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 0.7225806451612903,
|
| 1004 |
-
"grad_norm": 2.
|
| 1005 |
-
"learning_rate":
|
| 1006 |
-
"log_odds_chosen":
|
| 1007 |
-
"log_odds_ratio": -0.
|
| 1008 |
-
"logits/chosen": -1.
|
| 1009 |
-
"logits/rejected": 15.
|
| 1010 |
-
"logps/chosen": -1.
|
| 1011 |
-
"logps/rejected": -3.
|
| 1012 |
-
"loss": 1.
|
| 1013 |
-
"nll_loss": 1.
|
| 1014 |
-
"rewards/accuracies": 0.
|
| 1015 |
-
"rewards/chosen": -0.
|
| 1016 |
-
"rewards/margins": 0.
|
| 1017 |
-
"rewards/rejected": -0.
|
| 1018 |
"step": 560
|
| 1019 |
},
|
| 1020 |
{
|
| 1021 |
"epoch": 0.7354838709677419,
|
| 1022 |
-
"grad_norm":
|
| 1023 |
-
"learning_rate":
|
| 1024 |
-
"log_odds_chosen":
|
| 1025 |
-
"log_odds_ratio": -0.
|
| 1026 |
-
"logits/chosen": -1.
|
| 1027 |
-
"logits/rejected": 15.
|
| 1028 |
-
"logps/chosen": -
|
| 1029 |
-
"logps/rejected": -
|
| 1030 |
-
"loss": 1.
|
| 1031 |
-
"nll_loss": 1.
|
| 1032 |
-
"rewards/accuracies": 0.
|
| 1033 |
-
"rewards/chosen": -0.
|
| 1034 |
-
"rewards/margins": 0.
|
| 1035 |
-
"rewards/rejected": -0.
|
| 1036 |
"step": 570
|
| 1037 |
},
|
| 1038 |
{
|
| 1039 |
"epoch": 0.7483870967741936,
|
| 1040 |
-
"grad_norm": 2.
|
| 1041 |
-
"learning_rate":
|
| 1042 |
-
"log_odds_chosen":
|
| 1043 |
-
"log_odds_ratio": -0.
|
| 1044 |
-
"logits/chosen": -
|
| 1045 |
-
"logits/rejected": 15.
|
| 1046 |
-
"logps/chosen": -0.
|
| 1047 |
-
"logps/rejected": -3.
|
| 1048 |
-
"loss": 1.
|
| 1049 |
-
"nll_loss":
|
| 1050 |
-
"rewards/accuracies": 0.
|
| 1051 |
-
"rewards/chosen": -0.
|
| 1052 |
-
"rewards/margins": 0.
|
| 1053 |
-
"rewards/rejected": -0.
|
| 1054 |
"step": 580
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
"epoch": 0.7612903225806451,
|
| 1058 |
-
"grad_norm":
|
| 1059 |
-
"learning_rate":
|
| 1060 |
-
"log_odds_chosen":
|
| 1061 |
-
"log_odds_ratio": -0.
|
| 1062 |
-
"logits/chosen": -
|
| 1063 |
-
"logits/rejected": 15.
|
| 1064 |
-
"logps/chosen": -0.
|
| 1065 |
-
"logps/rejected": -3.
|
| 1066 |
-
"loss": 1.
|
| 1067 |
-
"nll_loss": 1.
|
| 1068 |
-
"rewards/accuracies": 0.
|
| 1069 |
-
"rewards/chosen": -0.
|
| 1070 |
-
"rewards/margins": 0.
|
| 1071 |
-
"rewards/rejected": -0.
|
| 1072 |
"step": 590
|
| 1073 |
},
|
| 1074 |
{
|
| 1075 |
"epoch": 0.7741935483870968,
|
| 1076 |
-
"grad_norm": 2.
|
| 1077 |
-
"learning_rate":
|
| 1078 |
-
"log_odds_chosen":
|
| 1079 |
-
"log_odds_ratio": -0.
|
| 1080 |
-
"logits/chosen": -1.
|
| 1081 |
-
"logits/rejected": 15.
|
| 1082 |
-
"logps/chosen": -1.
|
| 1083 |
-
"logps/rejected": -
|
| 1084 |
-
"loss": 1.
|
| 1085 |
-
"nll_loss": 1.
|
| 1086 |
-
"rewards/accuracies": 0.
|
| 1087 |
-
"rewards/chosen": -0.
|
| 1088 |
-
"rewards/margins": 0.
|
| 1089 |
-
"rewards/rejected": -0.
|
| 1090 |
"step": 600
|
| 1091 |
},
|
| 1092 |
{
|
| 1093 |
"epoch": 0.7870967741935484,
|
| 1094 |
-
"grad_norm":
|
| 1095 |
-
"learning_rate":
|
| 1096 |
-
"log_odds_chosen":
|
| 1097 |
-
"log_odds_ratio": -0.
|
| 1098 |
-
"logits/chosen": -2.
|
| 1099 |
-
"logits/rejected": 15.
|
| 1100 |
-
"logps/chosen": -1.
|
| 1101 |
-
"logps/rejected": -
|
| 1102 |
-
"loss": 1.
|
| 1103 |
-
"nll_loss": 1.
|
| 1104 |
-
"rewards/accuracies": 0.
|
| 1105 |
-
"rewards/chosen": -0.
|
| 1106 |
-
"rewards/margins": 0.
|
| 1107 |
-
"rewards/rejected": -0.
|
| 1108 |
"step": 610
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.8,
|
| 1112 |
-
"grad_norm": 3.
|
| 1113 |
-
"learning_rate":
|
| 1114 |
-
"log_odds_chosen":
|
| 1115 |
-
"log_odds_ratio": -0.
|
| 1116 |
-
"logits/chosen": -
|
| 1117 |
-
"logits/rejected": 15.
|
| 1118 |
-
"logps/chosen": -1.
|
| 1119 |
-
"logps/rejected": -
|
| 1120 |
-
"loss": 1.
|
| 1121 |
-
"nll_loss": 1.
|
| 1122 |
-
"rewards/accuracies": 0.
|
| 1123 |
-
"rewards/chosen": -0.
|
| 1124 |
-
"rewards/margins": 0.
|
| 1125 |
-
"rewards/rejected": -0.
|
| 1126 |
"step": 620
|
| 1127 |
},
|
| 1128 |
{
|
| 1129 |
"epoch": 0.8129032258064516,
|
| 1130 |
-
"grad_norm":
|
| 1131 |
-
"learning_rate":
|
| 1132 |
-
"log_odds_chosen":
|
| 1133 |
-
"log_odds_ratio": -0.
|
| 1134 |
-
"logits/chosen": -
|
| 1135 |
-
"logits/rejected": 15.
|
| 1136 |
-
"logps/chosen": -
|
| 1137 |
-
"logps/rejected": -3.
|
| 1138 |
-
"loss":
|
| 1139 |
-
"nll_loss":
|
| 1140 |
-
"rewards/accuracies": 0
|
| 1141 |
-
"rewards/chosen": -0.
|
| 1142 |
-
"rewards/margins": 0.
|
| 1143 |
-
"rewards/rejected": -0.
|
| 1144 |
"step": 630
|
| 1145 |
},
|
| 1146 |
{
|
| 1147 |
"epoch": 0.8258064516129032,
|
| 1148 |
-
"grad_norm": 3.
|
| 1149 |
-
"learning_rate":
|
| 1150 |
-
"log_odds_chosen":
|
| 1151 |
-
"log_odds_ratio": -0.
|
| 1152 |
-
"logits/chosen": -1.
|
| 1153 |
-
"logits/rejected": 15.
|
| 1154 |
-
"logps/chosen": -1.
|
| 1155 |
-
"logps/rejected": -
|
| 1156 |
-
"loss": 1.
|
| 1157 |
-
"nll_loss": 1.
|
| 1158 |
-
"rewards/accuracies": 0.
|
| 1159 |
-
"rewards/chosen": -0.
|
| 1160 |
-
"rewards/margins": 0.
|
| 1161 |
-
"rewards/rejected": -0.
|
| 1162 |
"step": 640
|
| 1163 |
},
|
| 1164 |
{
|
| 1165 |
"epoch": 0.8387096774193549,
|
| 1166 |
-
"grad_norm":
|
| 1167 |
-
"learning_rate":
|
| 1168 |
-
"log_odds_chosen":
|
| 1169 |
-
"log_odds_ratio": -0.
|
| 1170 |
-
"logits/chosen": -1.
|
| 1171 |
-
"logits/rejected": 16.
|
| 1172 |
-
"logps/chosen": -1.
|
| 1173 |
-
"logps/rejected": -
|
| 1174 |
-
"loss": 1.
|
| 1175 |
-
"nll_loss": 1.
|
| 1176 |
"rewards/accuracies": 0.987500011920929,
|
| 1177 |
-
"rewards/chosen": -0.
|
| 1178 |
-
"rewards/margins": 0.
|
| 1179 |
-
"rewards/rejected": -0.
|
| 1180 |
"step": 650
|
| 1181 |
},
|
| 1182 |
{
|
| 1183 |
"epoch": 0.8516129032258064,
|
| 1184 |
-
"grad_norm": 3.
|
| 1185 |
-
"learning_rate":
|
| 1186 |
-
"log_odds_chosen":
|
| 1187 |
-
"log_odds_ratio": -0.
|
| 1188 |
-
"logits/chosen": -1.
|
| 1189 |
-
"logits/rejected": 15.
|
| 1190 |
-
"logps/chosen": -1.
|
| 1191 |
-
"logps/rejected": -
|
| 1192 |
-
"loss": 1.
|
| 1193 |
-
"nll_loss": 1.
|
| 1194 |
-
"rewards/accuracies": 0.
|
| 1195 |
-
"rewards/chosen": -0.
|
| 1196 |
-
"rewards/margins": 0.
|
| 1197 |
-
"rewards/rejected": -0.
|
| 1198 |
"step": 660
|
| 1199 |
},
|
| 1200 |
{
|
| 1201 |
"epoch": 0.864516129032258,
|
| 1202 |
-
"grad_norm": 2.
|
| 1203 |
-
"learning_rate":
|
| 1204 |
-
"log_odds_chosen":
|
| 1205 |
-
"log_odds_ratio": -0.
|
| 1206 |
-
"logits/chosen": -
|
| 1207 |
-
"logits/rejected": 15.
|
| 1208 |
-
"logps/chosen": -1.
|
| 1209 |
-
"logps/rejected": -
|
| 1210 |
-
"loss": 1.
|
| 1211 |
-
"nll_loss": 1.
|
| 1212 |
"rewards/accuracies": 0.981249988079071,
|
| 1213 |
-
"rewards/chosen": -0.
|
| 1214 |
-
"rewards/margins": 0.
|
| 1215 |
-
"rewards/rejected": -0.
|
| 1216 |
"step": 670
|
| 1217 |
},
|
| 1218 |
{
|
| 1219 |
"epoch": 0.8774193548387097,
|
| 1220 |
-
"grad_norm": 3.
|
| 1221 |
-
"learning_rate":
|
| 1222 |
-
"log_odds_chosen":
|
| 1223 |
-
"log_odds_ratio": -0.
|
| 1224 |
-
"logits/chosen": -1.
|
| 1225 |
-
"logits/rejected": 15.
|
| 1226 |
-
"logps/chosen": -1.
|
| 1227 |
-
"logps/rejected": -
|
| 1228 |
-
"loss": 1.
|
| 1229 |
-
"nll_loss": 1.
|
| 1230 |
-
"rewards/accuracies": 0.
|
| 1231 |
-
"rewards/chosen": -0.
|
| 1232 |
-
"rewards/margins": 0.
|
| 1233 |
-
"rewards/rejected": -0.
|
| 1234 |
"step": 680
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8903225806451613,
|
| 1238 |
-
"grad_norm":
|
| 1239 |
-
"learning_rate":
|
| 1240 |
-
"log_odds_chosen":
|
| 1241 |
-
"log_odds_ratio": -0.
|
| 1242 |
-
"logits/chosen": -1.
|
| 1243 |
-
"logits/rejected": 15.
|
| 1244 |
-
"logps/chosen": -1.
|
| 1245 |
-
"logps/rejected": -
|
| 1246 |
-
"loss": 1.
|
| 1247 |
-
"nll_loss": 1.
|
| 1248 |
"rewards/accuracies": 0.9750000238418579,
|
| 1249 |
-
"rewards/chosen": -0.
|
| 1250 |
-
"rewards/margins": 0.
|
| 1251 |
-
"rewards/rejected": -0.
|
| 1252 |
"step": 690
|
| 1253 |
},
|
| 1254 |
{
|
| 1255 |
"epoch": 0.9032258064516129,
|
| 1256 |
-
"grad_norm": 3.
|
| 1257 |
-
"learning_rate":
|
| 1258 |
-
"log_odds_chosen":
|
| 1259 |
-
"log_odds_ratio": -0.
|
| 1260 |
-
"logits/chosen": -1.
|
| 1261 |
-
"logits/rejected": 15.
|
| 1262 |
-
"logps/chosen": -
|
| 1263 |
-
"logps/rejected": -
|
| 1264 |
-
"loss":
|
| 1265 |
-
"nll_loss":
|
| 1266 |
-
"rewards/accuracies": 0.
|
| 1267 |
-
"rewards/chosen": -0.
|
| 1268 |
-
"rewards/margins": 0.
|
| 1269 |
-
"rewards/rejected": -0.
|
| 1270 |
"step": 700
|
| 1271 |
},
|
| 1272 |
{
|
| 1273 |
"epoch": 0.9161290322580645,
|
| 1274 |
-
"grad_norm":
|
| 1275 |
-
"learning_rate":
|
| 1276 |
-
"log_odds_chosen":
|
| 1277 |
-
"log_odds_ratio": -0.
|
| 1278 |
-
"logits/chosen": -
|
| 1279 |
-
"logits/rejected": 15.
|
| 1280 |
-
"logps/chosen": -1.
|
| 1281 |
-
"logps/rejected": -
|
| 1282 |
-
"loss": 1.
|
| 1283 |
-
"nll_loss": 1.
|
| 1284 |
-
"rewards/accuracies": 0.
|
| 1285 |
-
"rewards/chosen": -0.
|
| 1286 |
-
"rewards/margins": 0.
|
| 1287 |
-
"rewards/rejected": -0.
|
| 1288 |
"step": 710
|
| 1289 |
},
|
| 1290 |
{
|
| 1291 |
"epoch": 0.9290322580645162,
|
| 1292 |
-
"grad_norm": 2.
|
| 1293 |
-
"learning_rate":
|
| 1294 |
-
"log_odds_chosen":
|
| 1295 |
-
"log_odds_ratio": -0.
|
| 1296 |
-
"logits/chosen": -1.
|
| 1297 |
-
"logits/rejected": 15.
|
| 1298 |
-
"logps/chosen": -
|
| 1299 |
-
"logps/rejected": -
|
| 1300 |
-
"loss": 1.
|
| 1301 |
-
"nll_loss": 1.
|
| 1302 |
-
"rewards/accuracies": 0
|
| 1303 |
-
"rewards/chosen": -0.
|
| 1304 |
-
"rewards/margins": 0.
|
| 1305 |
-
"rewards/rejected": -0.
|
| 1306 |
"step": 720
|
| 1307 |
},
|
| 1308 |
{
|
| 1309 |
"epoch": 0.9419354838709677,
|
| 1310 |
-
"grad_norm": 3.
|
| 1311 |
-
"learning_rate":
|
| 1312 |
-
"log_odds_chosen":
|
| 1313 |
-
"log_odds_ratio": -0.
|
| 1314 |
-
"logits/chosen": -
|
| 1315 |
-
"logits/rejected": 15.
|
| 1316 |
-
"logps/chosen": -1.
|
| 1317 |
-
"logps/rejected": -
|
| 1318 |
-
"loss": 1.
|
| 1319 |
-
"nll_loss": 1.
|
| 1320 |
"rewards/accuracies": 0.987500011920929,
|
| 1321 |
-
"rewards/chosen": -0.
|
| 1322 |
-
"rewards/margins": 0.
|
| 1323 |
-
"rewards/rejected": -0.
|
| 1324 |
"step": 730
|
| 1325 |
},
|
| 1326 |
{
|
| 1327 |
"epoch": 0.9548387096774194,
|
| 1328 |
-
"grad_norm":
|
| 1329 |
-
"learning_rate":
|
| 1330 |
-
"log_odds_chosen":
|
| 1331 |
-
"log_odds_ratio": -0.
|
| 1332 |
-
"logits/chosen": -2.
|
| 1333 |
-
"logits/rejected": 15.
|
| 1334 |
-
"logps/chosen": -
|
| 1335 |
-
"logps/rejected": -3.
|
| 1336 |
-
"loss":
|
| 1337 |
-
"nll_loss":
|
| 1338 |
"rewards/accuracies": 0.987500011920929,
|
| 1339 |
-
"rewards/chosen": -0.
|
| 1340 |
-
"rewards/margins": 0.
|
| 1341 |
-
"rewards/rejected": -0.
|
| 1342 |
"step": 740
|
| 1343 |
},
|
| 1344 |
{
|
| 1345 |
"epoch": 0.967741935483871,
|
| 1346 |
-
"grad_norm": 3.
|
| 1347 |
-
"learning_rate":
|
| 1348 |
-
"log_odds_chosen":
|
| 1349 |
-
"log_odds_ratio": -0.
|
| 1350 |
-
"logits/chosen": -
|
| 1351 |
-
"logits/rejected": 15.
|
| 1352 |
-
"logps/chosen": -1.
|
| 1353 |
-
"logps/rejected": -
|
| 1354 |
-
"loss": 1.
|
| 1355 |
-
"nll_loss": 1.
|
| 1356 |
-
"rewards/accuracies": 0.
|
| 1357 |
-
"rewards/chosen": -0.
|
| 1358 |
-
"rewards/margins": 0.
|
| 1359 |
-
"rewards/rejected": -0.
|
| 1360 |
"step": 750
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.9806451612903225,
|
| 1364 |
-
"grad_norm":
|
| 1365 |
-
"learning_rate":
|
| 1366 |
-
"log_odds_chosen":
|
| 1367 |
-
"log_odds_ratio": -0.
|
| 1368 |
-
"logits/chosen": -1.
|
| 1369 |
-
"logits/rejected": 15.
|
| 1370 |
-
"logps/chosen": -1.
|
| 1371 |
-
"logps/rejected": -
|
| 1372 |
-
"loss": 1.
|
| 1373 |
-
"nll_loss": 1.
|
| 1374 |
-
"rewards/accuracies": 0.
|
| 1375 |
-
"rewards/chosen": -0.
|
| 1376 |
-
"rewards/margins": 0.
|
| 1377 |
-
"rewards/rejected": -0.
|
| 1378 |
"step": 760
|
| 1379 |
},
|
| 1380 |
{
|
| 1381 |
"epoch": 0.9935483870967742,
|
| 1382 |
-
"grad_norm":
|
| 1383 |
-
"learning_rate":
|
| 1384 |
-
"log_odds_chosen":
|
| 1385 |
-
"log_odds_ratio": -0.
|
| 1386 |
-
"logits/chosen": -
|
| 1387 |
-
"logits/rejected": 15.
|
| 1388 |
-
"logps/chosen": -1.
|
| 1389 |
-
"logps/rejected": -
|
| 1390 |
-
"loss": 1.
|
| 1391 |
-
"nll_loss": 1.
|
| 1392 |
-
"rewards/accuracies": 0.
|
| 1393 |
-
"rewards/chosen": -0.
|
| 1394 |
-
"rewards/margins": 0.
|
| 1395 |
-
"rewards/rejected": -0.
|
| 1396 |
"step": 770
|
| 1397 |
},
|
| 1398 |
{
|
| 1399 |
"epoch": 1.0,
|
| 1400 |
"step": 775,
|
| 1401 |
"total_flos": 0.0,
|
| 1402 |
-
"train_loss": 1.
|
| 1403 |
-
"train_runtime":
|
| 1404 |
-
"train_samples_per_second": 8.
|
| 1405 |
-
"train_steps_per_second": 0.
|
| 1406 |
}
|
| 1407 |
],
|
| 1408 |
"logging_steps": 10,
|
|
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.012903225806451613,
|
| 14 |
+
"grad_norm": 9.303802312326823,
|
| 15 |
+
"learning_rate": 9.000000000000001e-07,
|
| 16 |
+
"log_odds_chosen": 1.083593726158142,
|
| 17 |
+
"log_odds_ratio": -0.33476561307907104,
|
| 18 |
+
"logits/chosen": -1.0812499523162842,
|
| 19 |
"logits/rejected": 15.675000190734863,
|
| 20 |
+
"logps/chosen": -1.661718726158142,
|
| 21 |
+
"logps/rejected": -2.6156249046325684,
|
| 22 |
+
"loss": 2.0143,
|
| 23 |
+
"nll_loss": 1.99609375,
|
| 24 |
"rewards/accuracies": 0.949999988079071,
|
| 25 |
+
"rewards/chosen": -0.083251953125,
|
| 26 |
+
"rewards/margins": 0.04765624925494194,
|
| 27 |
+
"rewards/rejected": -0.13066406548023224,
|
| 28 |
"step": 10
|
| 29 |
},
|
| 30 |
{
|
| 31 |
"epoch": 0.025806451612903226,
|
| 32 |
+
"grad_norm": 7.100174397164698,
|
| 33 |
+
"learning_rate": 1.9000000000000002e-06,
|
| 34 |
+
"log_odds_chosen": 1.1789062023162842,
|
| 35 |
+
"log_odds_ratio": -0.31318360567092896,
|
| 36 |
+
"logits/chosen": -1.015625,
|
| 37 |
+
"logits/rejected": 16.8125,
|
| 38 |
+
"logps/chosen": -1.6867187023162842,
|
| 39 |
+
"logps/rejected": -2.729687452316284,
|
| 40 |
+
"loss": 2.0014,
|
| 41 |
+
"nll_loss": 1.9835937023162842,
|
| 42 |
"rewards/accuracies": 0.9750000238418579,
|
| 43 |
+
"rewards/chosen": -0.08432617038488388,
|
| 44 |
+
"rewards/margins": 0.05207519605755806,
|
| 45 |
+
"rewards/rejected": -0.13637694716453552,
|
| 46 |
"step": 20
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.03870967741935484,
|
| 50 |
+
"grad_norm": 7.008035670995982,
|
| 51 |
+
"learning_rate": 2.9e-06,
|
| 52 |
+
"log_odds_chosen": 0.9878906011581421,
|
| 53 |
+
"log_odds_ratio": -0.35566407442092896,
|
| 54 |
+
"logits/chosen": -1.28765869140625,
|
| 55 |
+
"logits/rejected": 16.299999237060547,
|
| 56 |
+
"logps/chosen": -1.661718726158142,
|
| 57 |
+
"logps/rejected": -2.512500047683716,
|
| 58 |
+
"loss": 1.9846,
|
| 59 |
+
"nll_loss": 1.966406226158142,
|
| 60 |
"rewards/accuracies": 0.9375,
|
| 61 |
+
"rewards/chosen": -0.08310546725988388,
|
| 62 |
+
"rewards/margins": 0.04267578199505806,
|
| 63 |
+
"rewards/rejected": -0.12568359076976776,
|
| 64 |
"step": 30
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"epoch": 0.05161290322580645,
|
| 68 |
+
"grad_norm": 5.261804897002812,
|
| 69 |
+
"learning_rate": 3.900000000000001e-06,
|
| 70 |
+
"log_odds_chosen": 0.8687499761581421,
|
| 71 |
+
"log_odds_ratio": -0.3785156309604645,
|
| 72 |
+
"logits/chosen": -0.40996092557907104,
|
| 73 |
+
"logits/rejected": 16.706249237060547,
|
| 74 |
+
"logps/chosen": -1.548437476158142,
|
| 75 |
+
"logps/rejected": -2.284374952316284,
|
| 76 |
+
"loss": 1.8731,
|
| 77 |
+
"nll_loss": 1.8546874523162842,
|
| 78 |
"rewards/accuracies": 0.9624999761581421,
|
| 79 |
+
"rewards/chosen": -0.07744140923023224,
|
| 80 |
+
"rewards/margins": 0.036865234375,
|
| 81 |
+
"rewards/rejected": -0.11435546725988388,
|
| 82 |
"step": 40
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"epoch": 0.06451612903225806,
|
| 86 |
+
"grad_norm": 3.9860274644043203,
|
| 87 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 88 |
+
"log_odds_chosen": 0.748828113079071,
|
| 89 |
+
"log_odds_ratio": -0.4212890565395355,
|
| 90 |
+
"logits/chosen": 0.379638671875,
|
| 91 |
+
"logits/rejected": 16.956249237060547,
|
| 92 |
+
"logps/chosen": -1.467187523841858,
|
| 93 |
+
"logps/rejected": -2.0882811546325684,
|
| 94 |
+
"loss": 1.7385,
|
| 95 |
+
"nll_loss": 1.7179687023162842,
|
| 96 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 97 |
+
"rewards/chosen": -0.07333984225988388,
|
| 98 |
+
"rewards/margins": 0.0311279296875,
|
| 99 |
+
"rewards/rejected": -0.10444335639476776,
|
| 100 |
"step": 50
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.07741935483870968,
|
| 104 |
+
"grad_norm": 3.8803923662969386,
|
| 105 |
+
"learning_rate": 5.9e-06,
|
| 106 |
+
"log_odds_chosen": 0.49042969942092896,
|
| 107 |
+
"log_odds_ratio": -0.5087890625,
|
| 108 |
+
"logits/chosen": 0.905078113079071,
|
| 109 |
+
"logits/rejected": 16.831249237060547,
|
| 110 |
+
"logps/chosen": -1.4015624523162842,
|
| 111 |
+
"logps/rejected": -1.791406273841858,
|
| 112 |
+
"loss": 1.6458,
|
| 113 |
+
"nll_loss": 1.62109375,
|
| 114 |
+
"rewards/accuracies": 0.824999988079071,
|
| 115 |
+
"rewards/chosen": -0.070068359375,
|
| 116 |
+
"rewards/margins": 0.019622802734375,
|
| 117 |
+
"rewards/rejected": -0.089599609375,
|
| 118 |
"step": 60
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 0.09032258064516129,
|
| 122 |
+
"grad_norm": 2.9291160475725464,
|
| 123 |
+
"learning_rate": 6.9e-06,
|
| 124 |
+
"log_odds_chosen": 0.42890626192092896,
|
| 125 |
+
"log_odds_ratio": -0.525585949420929,
|
| 126 |
+
"logits/chosen": 1.0153319835662842,
|
| 127 |
+
"logits/rejected": 17.274999618530273,
|
| 128 |
+
"logps/chosen": -1.251562476158142,
|
| 129 |
+
"logps/rejected": -1.5867187976837158,
|
| 130 |
+
"loss": 1.4603,
|
| 131 |
+
"nll_loss": 1.435156226158142,
|
| 132 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 133 |
+
"rewards/chosen": -0.0626220703125,
|
| 134 |
+
"rewards/margins": 0.01672973670065403,
|
| 135 |
+
"rewards/rejected": -0.07929687201976776,
|
| 136 |
"step": 70
|
| 137 |
},
|
| 138 |
{
|
| 139 |
"epoch": 0.1032258064516129,
|
| 140 |
+
"grad_norm": 2.3895368722437746,
|
| 141 |
+
"learning_rate": 7.9e-06,
|
| 142 |
+
"log_odds_chosen": 0.20942382514476776,
|
| 143 |
+
"log_odds_ratio": -0.6214843988418579,
|
| 144 |
+
"logits/chosen": 1.1715819835662842,
|
| 145 |
+
"logits/rejected": 17.331249237060547,
|
| 146 |
+
"logps/chosen": -1.2156250476837158,
|
| 147 |
+
"logps/rejected": -1.375,
|
| 148 |
+
"loss": 1.3688,
|
| 149 |
+
"nll_loss": 1.3390624523162842,
|
| 150 |
+
"rewards/accuracies": 0.668749988079071,
|
| 151 |
+
"rewards/chosen": -0.06086425855755806,
|
| 152 |
+
"rewards/margins": 0.0077911377884447575,
|
| 153 |
+
"rewards/rejected": -0.06862793117761612,
|
| 154 |
"step": 80
|
| 155 |
},
|
| 156 |
{
|
| 157 |
"epoch": 0.11612903225806452,
|
| 158 |
+
"grad_norm": 2.1265814928397626,
|
| 159 |
+
"learning_rate": 8.900000000000001e-06,
|
| 160 |
+
"log_odds_chosen": 0.1634521484375,
|
| 161 |
+
"log_odds_ratio": -0.649609386920929,
|
| 162 |
+
"logits/chosen": 0.767163097858429,
|
| 163 |
+
"logits/rejected": 17.131250381469727,
|
| 164 |
+
"logps/chosen": -1.14453125,
|
| 165 |
+
"logps/rejected": -1.26953125,
|
| 166 |
+
"loss": 1.292,
|
| 167 |
+
"nll_loss": 1.260156273841858,
|
| 168 |
+
"rewards/accuracies": 0.625,
|
| 169 |
+
"rewards/chosen": -0.05727539211511612,
|
| 170 |
+
"rewards/margins": 0.006190490908920765,
|
| 171 |
+
"rewards/rejected": -0.06345214694738388,
|
| 172 |
"step": 90
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"epoch": 0.12903225806451613,
|
| 176 |
+
"grad_norm": 2.459971487020149,
|
| 177 |
+
"learning_rate": 9.9e-06,
|
| 178 |
+
"log_odds_chosen": 0.14106445014476776,
|
| 179 |
+
"log_odds_ratio": -0.6585937738418579,
|
| 180 |
+
"logits/chosen": 0.06308593600988388,
|
| 181 |
+
"logits/rejected": 16.306249618530273,
|
| 182 |
+
"logps/chosen": -1.1515624523162842,
|
| 183 |
+
"logps/rejected": -1.2664062976837158,
|
| 184 |
+
"loss": 1.2808,
|
| 185 |
+
"nll_loss": 1.247656226158142,
|
| 186 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 187 |
+
"rewards/chosen": -0.05751953274011612,
|
| 188 |
+
"rewards/margins": 0.005828857421875,
|
| 189 |
+
"rewards/rejected": -0.06342773139476776,
|
| 190 |
"step": 100
|
| 191 |
},
|
| 192 |
{
|
| 193 |
"epoch": 0.14193548387096774,
|
| 194 |
+
"grad_norm": 2.4830433577540436,
|
| 195 |
+
"learning_rate": 9.578262852211514e-06,
|
| 196 |
+
"log_odds_chosen": 0.37322998046875,
|
| 197 |
+
"log_odds_ratio": -0.560351550579071,
|
| 198 |
+
"logits/chosen": 0.01640624925494194,
|
| 199 |
+
"logits/rejected": 16.587499618530273,
|
| 200 |
+
"logps/chosen": -1.126953125,
|
| 201 |
+
"logps/rejected": -1.4070312976837158,
|
| 202 |
+
"loss": 1.2312,
|
| 203 |
+
"nll_loss": 1.2023437023162842,
|
| 204 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 205 |
+
"rewards/chosen": -0.05644531175494194,
|
| 206 |
+
"rewards/margins": 0.013935851864516735,
|
| 207 |
"rewards/rejected": -0.07038573920726776,
|
| 208 |
"step": 110
|
| 209 |
},
|
| 210 |
{
|
| 211 |
"epoch": 0.15483870967741936,
|
| 212 |
+
"grad_norm": 8.467524109258711,
|
| 213 |
+
"learning_rate": 9.166984970282113e-06,
|
| 214 |
+
"log_odds_chosen": 0.886914074420929,
|
| 215 |
+
"log_odds_ratio": -0.39765626192092896,
|
| 216 |
+
"logits/chosen": -0.9283202886581421,
|
| 217 |
+
"logits/rejected": 16.756250381469727,
|
| 218 |
+
"logps/chosen": -1.103124976158142,
|
| 219 |
+
"logps/rejected": -1.77734375,
|
| 220 |
+
"loss": 1.1735,
|
| 221 |
+
"nll_loss": 1.1531250476837158,
|
| 222 |
+
"rewards/accuracies": 0.90625,
|
| 223 |
+
"rewards/chosen": -0.05517578125,
|
| 224 |
+
"rewards/margins": 0.03373413160443306,
|
| 225 |
+
"rewards/rejected": -0.08891601860523224,
|
| 226 |
"step": 120
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.16774193548387098,
|
| 230 |
+
"grad_norm": 2.492915426271658,
|
| 231 |
+
"learning_rate": 8.804509063256239e-06,
|
| 232 |
+
"log_odds_chosen": 1.274999976158142,
|
| 233 |
+
"log_odds_ratio": -0.29912108182907104,
|
| 234 |
+
"logits/chosen": -0.2676025331020355,
|
| 235 |
+
"logits/rejected": 16.34375,
|
| 236 |
+
"logps/chosen": -1.1281249523162842,
|
| 237 |
+
"logps/rejected": -2.14453125,
|
| 238 |
+
"loss": 1.1897,
|
| 239 |
+
"nll_loss": 1.174218773841858,
|
| 240 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 241 |
+
"rewards/chosen": -0.05637206882238388,
|
| 242 |
+
"rewards/margins": 0.05068359524011612,
|
| 243 |
+
"rewards/rejected": -0.10708007961511612,
|
| 244 |
"step": 130
|
| 245 |
},
|
| 246 |
{
|
| 247 |
"epoch": 0.18064516129032257,
|
| 248 |
+
"grad_norm": 2.259429852293854,
|
| 249 |
+
"learning_rate": 8.481889296799709e-06,
|
| 250 |
+
"log_odds_chosen": 1.3566405773162842,
|
| 251 |
+
"log_odds_ratio": -0.2734375,
|
| 252 |
+
"logits/chosen": -0.8609374761581421,
|
| 253 |
+
"logits/rejected": 15.743749618530273,
|
| 254 |
+
"logps/chosen": -1.103906273841858,
|
| 255 |
+
"logps/rejected": -2.171875,
|
| 256 |
+
"loss": 1.1491,
|
| 257 |
+
"nll_loss": 1.1355469226837158,
|
| 258 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 259 |
+
"rewards/chosen": -0.05512695387005806,
|
| 260 |
+
"rewards/margins": 0.053466796875,
|
| 261 |
+
"rewards/rejected": -0.10859374701976776,
|
| 262 |
"step": 140
|
| 263 |
},
|
| 264 |
{
|
| 265 |
"epoch": 0.1935483870967742,
|
| 266 |
+
"grad_norm": 2.3418100298894013,
|
| 267 |
+
"learning_rate": 8.192319205190406e-06,
|
| 268 |
+
"log_odds_chosen": 1.53125,
|
| 269 |
+
"log_odds_ratio": -0.24775390326976776,
|
| 270 |
+
"logits/chosen": -1.3630859851837158,
|
| 271 |
+
"logits/rejected": 15.75,
|
| 272 |
+
"logps/chosen": -1.092187523841858,
|
| 273 |
+
"logps/rejected": -2.328125,
|
| 274 |
+
"loss": 1.137,
|
| 275 |
+
"nll_loss": 1.1242187023162842,
|
| 276 |
+
"rewards/accuracies": 0.981249988079071,
|
| 277 |
+
"rewards/chosen": -0.05458984524011612,
|
| 278 |
+
"rewards/margins": 0.06184082105755806,
|
| 279 |
+
"rewards/rejected": -0.11635742336511612,
|
| 280 |
"step": 150
|
| 281 |
},
|
| 282 |
{
|
| 283 |
"epoch": 0.2064516129032258,
|
| 284 |
+
"grad_norm": 3.440439256067863,
|
| 285 |
+
"learning_rate": 7.930515857181441e-06,
|
| 286 |
+
"log_odds_chosen": 1.54296875,
|
| 287 |
+
"log_odds_ratio": -0.2630859315395355,
|
| 288 |
+
"logits/chosen": -1.760156273841858,
|
| 289 |
+
"logits/rejected": 15.199999809265137,
|
| 290 |
+
"logps/chosen": -1.055078148841858,
|
| 291 |
+
"logps/rejected": -2.293750047683716,
|
| 292 |
+
"loss": 1.1039,
|
| 293 |
+
"nll_loss": 1.0906250476837158,
|
| 294 |
+
"rewards/accuracies": 0.96875,
|
| 295 |
+
"rewards/chosen": -0.05275879055261612,
|
| 296 |
+
"rewards/margins": 0.06201171875,
|
| 297 |
+
"rewards/rejected": -0.11484374850988388,
|
| 298 |
"step": 160
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 0.21935483870967742,
|
| 302 |
+
"grad_norm": 2.237987495849613,
|
| 303 |
+
"learning_rate": 7.692307692307692e-06,
|
| 304 |
+
"log_odds_chosen": 1.7804687023162842,
|
| 305 |
+
"log_odds_ratio": -0.20322266221046448,
|
| 306 |
+
"logits/chosen": -1.688085913658142,
|
| 307 |
+
"logits/rejected": 16.331249237060547,
|
| 308 |
+
"logps/chosen": -1.01171875,
|
| 309 |
+
"logps/rejected": -2.440624952316284,
|
| 310 |
+
"loss": 1.0603,
|
| 311 |
+
"nll_loss": 1.050390601158142,
|
| 312 |
+
"rewards/accuracies": 0.981249988079071,
|
| 313 |
+
"rewards/chosen": -0.05051269382238388,
|
| 314 |
+
"rewards/margins": 0.07155761867761612,
|
| 315 |
+
"rewards/rejected": -0.12216796725988388,
|
| 316 |
"step": 170
|
| 317 |
},
|
| 318 |
{
|
| 319 |
"epoch": 0.23225806451612904,
|
| 320 |
+
"grad_norm": 3.0199229105723564,
|
| 321 |
+
"learning_rate": 7.4743509275193595e-06,
|
| 322 |
+
"log_odds_chosen": 1.7257812023162842,
|
| 323 |
+
"log_odds_ratio": -0.20849609375,
|
| 324 |
+
"logits/chosen": -1.4043457508087158,
|
| 325 |
+
"logits/rejected": 15.712499618530273,
|
| 326 |
+
"logps/chosen": -1.048828125,
|
| 327 |
+
"logps/rejected": -2.4312500953674316,
|
| 328 |
+
"loss": 1.075,
|
| 329 |
+
"nll_loss": 1.0652344226837158,
|
| 330 |
+
"rewards/accuracies": 0.981249988079071,
|
| 331 |
+
"rewards/chosen": -0.05239257961511612,
|
| 332 |
+
"rewards/margins": 0.06918945163488388,
|
| 333 |
+
"rewards/rejected": -0.12163086235523224,
|
| 334 |
"step": 180
|
| 335 |
},
|
| 336 |
{
|
| 337 |
"epoch": 0.24516129032258063,
|
| 338 |
+
"grad_norm": 6.554283287023061,
|
| 339 |
+
"learning_rate": 7.27392967453308e-06,
|
| 340 |
+
"log_odds_chosen": 1.7296874523162842,
|
| 341 |
+
"log_odds_ratio": -0.21250000596046448,
|
| 342 |
+
"logits/chosen": -0.91064453125,
|
| 343 |
+
"logits/rejected": 16.212499618530273,
|
| 344 |
+
"logps/chosen": -1.030859351158142,
|
| 345 |
+
"logps/rejected": -2.418750047683716,
|
| 346 |
+
"loss": 1.0727,
|
| 347 |
+
"nll_loss": 1.0617187023162842,
|
| 348 |
+
"rewards/accuracies": 0.96875,
|
| 349 |
+
"rewards/chosen": -0.05158691480755806,
|
| 350 |
+
"rewards/margins": 0.06936035305261612,
|
| 351 |
+
"rewards/rejected": -0.12114258110523224,
|
| 352 |
"step": 190
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.25806451612903225,
|
| 356 |
+
"grad_norm": 2.3355013011756913,
|
| 357 |
+
"learning_rate": 7.088812050083359e-06,
|
| 358 |
+
"log_odds_chosen": 1.876562476158142,
|
| 359 |
+
"log_odds_ratio": -0.18281249701976776,
|
| 360 |
+
"logits/chosen": -1.1605956554412842,
|
| 361 |
+
"logits/rejected": 16.456249237060547,
|
| 362 |
+
"logps/chosen": -1.064453125,
|
| 363 |
+
"logps/rejected": -2.590625047683716,
|
| 364 |
+
"loss": 1.0956,
|
| 365 |
+
"nll_loss": 1.0867187976837158,
|
| 366 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 367 |
+
"rewards/chosen": -0.05319824069738388,
|
| 368 |
+
"rewards/margins": 0.076416015625,
|
| 369 |
+
"rewards/rejected": -0.12958984076976776,
|
| 370 |
"step": 200
|
| 371 |
},
|
| 372 |
{
|
| 373 |
"epoch": 0.2709677419354839,
|
| 374 |
+
"grad_norm": 2.663340541202873,
|
| 375 |
+
"learning_rate": 6.917144638660746e-06,
|
| 376 |
+
"log_odds_chosen": 1.8046875,
|
| 377 |
+
"log_odds_ratio": -0.21416015923023224,
|
| 378 |
+
"logits/chosen": -1.562109351158142,
|
| 379 |
+
"logits/rejected": 15.65625,
|
| 380 |
+
"logps/chosen": -1.050390601158142,
|
| 381 |
+
"logps/rejected": -2.5093750953674316,
|
| 382 |
+
"loss": 1.0719,
|
| 383 |
+
"nll_loss": 1.060937523841858,
|
| 384 |
+
"rewards/accuracies": 0.956250011920929,
|
| 385 |
+
"rewards/chosen": -0.05251464992761612,
|
| 386 |
+
"rewards/margins": 0.07290039211511612,
|
| 387 |
+
"rewards/rejected": -0.12548828125,
|
| 388 |
"step": 210
|
| 389 |
},
|
| 390 |
{
|
| 391 |
"epoch": 0.2838709677419355,
|
| 392 |
+
"grad_norm": 2.4359455832988774,
|
| 393 |
+
"learning_rate": 6.757373783994861e-06,
|
| 394 |
+
"log_odds_chosen": 1.8781249523162842,
|
| 395 |
+
"log_odds_ratio": -0.21025390923023224,
|
| 396 |
+
"logits/chosen": -0.8188232183456421,
|
| 397 |
+
"logits/rejected": 16.125,
|
| 398 |
+
"logps/chosen": -1.103906273841858,
|
| 399 |
+
"logps/rejected": -2.6578125953674316,
|
| 400 |
+
"loss": 1.13,
|
| 401 |
+
"nll_loss": 1.119531273841858,
|
| 402 |
+
"rewards/accuracies": 0.949999988079071,
|
| 403 |
+
"rewards/chosen": -0.05522460862994194,
|
| 404 |
+
"rewards/margins": 0.07768554985523224,
|
| 405 |
+
"rewards/rejected": -0.13300780951976776,
|
| 406 |
"step": 220
|
| 407 |
},
|
| 408 |
{
|
| 409 |
"epoch": 0.2967741935483871,
|
| 410 |
+
"grad_norm": 2.5381406849446044,
|
| 411 |
+
"learning_rate": 6.6081860045508985e-06,
|
| 412 |
+
"log_odds_chosen": 1.8273437023162842,
|
| 413 |
+
"log_odds_ratio": -0.19570311903953552,
|
| 414 |
+
"logits/chosen": -0.516894519329071,
|
| 415 |
+
"logits/rejected": 15.912500381469727,
|
| 416 |
+
"logps/chosen": -1.114843726158142,
|
| 417 |
+
"logps/rejected": -2.606250047683716,
|
| 418 |
+
"loss": 1.1259,
|
| 419 |
+
"nll_loss": 1.1160156726837158,
|
| 420 |
"rewards/accuracies": 0.9750000238418579,
|
| 421 |
+
"rewards/chosen": -0.05568847805261612,
|
| 422 |
+
"rewards/margins": 0.07460937649011612,
|
| 423 |
+
"rewards/rejected": -0.13032226264476776,
|
| 424 |
"step": 230
|
| 425 |
},
|
| 426 |
{
|
| 427 |
"epoch": 0.3096774193548387,
|
| 428 |
+
"grad_norm": 2.249756481488481,
|
| 429 |
+
"learning_rate": 6.46846227353151e-06,
|
| 430 |
+
"log_odds_chosen": 1.7234375476837158,
|
| 431 |
+
"log_odds_ratio": -0.22285155951976776,
|
| 432 |
+
"logits/chosen": -1.727636694908142,
|
| 433 |
+
"logits/rejected": 14.53125,
|
| 434 |
+
"logps/chosen": -1.047265648841858,
|
| 435 |
+
"logps/rejected": -2.4453125,
|
| 436 |
+
"loss": 1.0836,
|
| 437 |
+
"nll_loss": 1.0730469226837158,
|
| 438 |
+
"rewards/accuracies": 0.987500011920929,
|
| 439 |
+
"rewards/chosen": -0.05241699144244194,
|
| 440 |
+
"rewards/margins": 0.06984863430261612,
|
| 441 |
+
"rewards/rejected": -0.12221679836511612,
|
| 442 |
"step": 240
|
| 443 |
},
|
| 444 |
{
|
| 445 |
"epoch": 0.3225806451612903,
|
| 446 |
+
"grad_norm": 2.4182796191238567,
|
| 447 |
+
"learning_rate": 6.337242505244779e-06,
|
| 448 |
+
"log_odds_chosen": 2.01171875,
|
| 449 |
+
"log_odds_ratio": -0.17333984375,
|
| 450 |
+
"logits/chosen": -2.0010743141174316,
|
| 451 |
+
"logits/rejected": 15.53125,
|
| 452 |
+
"logps/chosen": -1.0105469226837158,
|
| 453 |
+
"logps/rejected": -2.6484375,
|
| 454 |
+
"loss": 1.0448,
|
| 455 |
+
"nll_loss": 1.03515625,
|
| 456 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 457 |
+
"rewards/chosen": -0.05056152492761612,
|
| 458 |
+
"rewards/margins": 0.08193359524011612,
|
| 459 |
+
"rewards/rejected": -0.13247069716453552,
|
| 460 |
"step": 250
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 0.33548387096774196,
|
| 464 |
+
"grad_norm": 2.4058296328212965,
|
| 465 |
+
"learning_rate": 6.213697660012001e-06,
|
| 466 |
+
"log_odds_chosen": 1.8125,
|
| 467 |
+
"log_odds_ratio": -0.2138671875,
|
| 468 |
+
"logits/chosen": -1.9609375,
|
| 469 |
+
"logits/rejected": 15.487500190734863,
|
| 470 |
+
"logps/chosen": -1.032812476158142,
|
| 471 |
+
"logps/rejected": -2.5,
|
| 472 |
+
"loss": 1.0654,
|
| 473 |
+
"nll_loss": 1.054296851158142,
|
| 474 |
+
"rewards/accuracies": 0.981249988079071,
|
| 475 |
+
"rewards/chosen": -0.05168456956744194,
|
| 476 |
+
"rewards/margins": 0.0733642578125,
|
| 477 |
+
"rewards/rejected": -0.12495116889476776,
|
| 478 |
"step": 260
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.34838709677419355,
|
| 482 |
+
"grad_norm": 2.448833917691866,
|
| 483 |
+
"learning_rate": 6.097107608496925e-06,
|
| 484 |
+
"log_odds_chosen": 2.004687547683716,
|
| 485 |
+
"log_odds_ratio": -0.18144531548023224,
|
| 486 |
+
"logits/chosen": -1.5515625476837158,
|
| 487 |
+
"logits/rejected": 15.40625,
|
| 488 |
+
"logps/chosen": -1.0011718273162842,
|
| 489 |
+
"logps/rejected": -2.6343750953674316,
|
| 490 |
+
"loss": 1.0323,
|
| 491 |
+
"nll_loss": 1.0226562023162842,
|
| 492 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 493 |
+
"rewards/chosen": -0.050048828125,
|
| 494 |
+
"rewards/margins": 0.08164062350988388,
|
| 495 |
+
"rewards/rejected": -0.1318359375,
|
| 496 |
"step": 270
|
| 497 |
},
|
| 498 |
{
|
| 499 |
"epoch": 0.36129032258064514,
|
| 500 |
+
"grad_norm": 2.7061200750489762,
|
| 501 |
+
"learning_rate": 5.986843400892497e-06,
|
| 502 |
+
"log_odds_chosen": 2.0953125953674316,
|
| 503 |
+
"log_odds_ratio": -0.16904297471046448,
|
| 504 |
+
"logits/chosen": -1.35595703125,
|
| 505 |
+
"logits/rejected": 15.625,
|
| 506 |
+
"logps/chosen": -1.0671875476837158,
|
| 507 |
+
"logps/rejected": -2.792187452316284,
|
| 508 |
+
"loss": 1.0819,
|
| 509 |
+
"nll_loss": 1.073828101158142,
|
| 510 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 511 |
+
"rewards/chosen": -0.05336914211511612,
|
| 512 |
+
"rewards/margins": 0.08642578125,
|
| 513 |
+
"rewards/rejected": -0.13984374701976776,
|
| 514 |
"step": 280
|
| 515 |
},
|
| 516 |
{
|
| 517 |
"epoch": 0.3741935483870968,
|
| 518 |
+
"grad_norm": 2.4367758380161573,
|
| 519 |
+
"learning_rate": 5.882352941176471e-06,
|
| 520 |
+
"log_odds_chosen": 2.0132813453674316,
|
| 521 |
+
"log_odds_ratio": -0.1748046875,
|
| 522 |
+
"logits/chosen": -1.527734398841858,
|
| 523 |
+
"logits/rejected": 15.881250381469727,
|
| 524 |
+
"logps/chosen": -1.030859351158142,
|
| 525 |
+
"logps/rejected": -2.6656250953674316,
|
| 526 |
+
"loss": 1.0479,
|
| 527 |
+
"nll_loss": 1.0390625,
|
| 528 |
+
"rewards/accuracies": 0.987500011920929,
|
| 529 |
+
"rewards/chosen": -0.051513671875,
|
| 530 |
+
"rewards/margins": 0.08173827826976776,
|
| 531 |
+
"rewards/rejected": -0.13325195014476776,
|
| 532 |
"step": 290
|
| 533 |
},
|
| 534 |
{
|
| 535 |
"epoch": 0.3870967741935484,
|
| 536 |
+
"grad_norm": 3.3727479977644363,
|
| 537 |
+
"learning_rate": 5.783149319662402e-06,
|
| 538 |
+
"log_odds_chosen": 2.188281297683716,
|
| 539 |
+
"log_odds_ratio": -0.182373046875,
|
| 540 |
+
"logits/chosen": -1.51806640625,
|
| 541 |
+
"logits/rejected": 15.743749618530273,
|
| 542 |
+
"logps/chosen": -1.0539062023162842,
|
| 543 |
+
"logps/rejected": -2.879687547683716,
|
| 544 |
+
"loss": 1.0747,
|
| 545 |
+
"nll_loss": 1.0656249523162842,
|
| 546 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 547 |
+
"rewards/chosen": -0.05266113206744194,
|
| 548 |
+
"rewards/margins": 0.09116210788488388,
|
| 549 |
+
"rewards/rejected": -0.14374999701976776,
|
| 550 |
"step": 300
|
| 551 |
},
|
| 552 |
{
|
| 553 |
"epoch": 0.4,
|
| 554 |
+
"grad_norm": 2.812934888942385,
|
| 555 |
+
"learning_rate": 5.688801239885744e-06,
|
| 556 |
+
"log_odds_chosen": 2.254687547683716,
|
| 557 |
+
"log_odds_ratio": -0.17075195908546448,
|
| 558 |
+
"logits/chosen": -1.6027343273162842,
|
| 559 |
+
"logits/rejected": 15.181249618530273,
|
| 560 |
+
"logps/chosen": -1.0187499523162842,
|
| 561 |
+
"logps/rejected": -2.8968749046325684,
|
| 562 |
+
"loss": 1.0446,
|
| 563 |
+
"nll_loss": 1.036718726158142,
|
| 564 |
+
"rewards/accuracies": 0.96875,
|
| 565 |
+
"rewards/chosen": -0.05092773586511612,
|
| 566 |
+
"rewards/margins": 0.09372558444738388,
|
| 567 |
+
"rewards/rejected": -0.14467772841453552,
|
| 568 |
"step": 310
|
| 569 |
},
|
| 570 |
{
|
| 571 |
"epoch": 0.4129032258064516,
|
| 572 |
+
"grad_norm": 2.751505791676006,
|
| 573 |
+
"learning_rate": 5.598925109558544e-06,
|
| 574 |
+
"log_odds_chosen": 2.2125000953674316,
|
| 575 |
+
"log_odds_ratio": -0.169677734375,
|
| 576 |
+
"logits/chosen": -1.304540991783142,
|
| 577 |
+
"logits/rejected": 15.78125,
|
| 578 |
+
"logps/chosen": -1.0320312976837158,
|
| 579 |
+
"logps/rejected": -2.8687500953674316,
|
| 580 |
+
"loss": 1.0443,
|
| 581 |
+
"nll_loss": 1.036718726158142,
|
| 582 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 583 |
+
"rewards/chosen": -0.05166015774011612,
|
| 584 |
+
"rewards/margins": 0.09177245944738388,
|
| 585 |
+
"rewards/rejected": -0.14340820908546448,
|
| 586 |
"step": 320
|
| 587 |
},
|
| 588 |
{
|
| 589 |
"epoch": 0.4258064516129032,
|
| 590 |
+
"grad_norm": 3.3773353240681008,
|
| 591 |
+
"learning_rate": 5.513178464199713e-06,
|
| 592 |
+
"log_odds_chosen": 2.3843750953674316,
|
| 593 |
+
"log_odds_ratio": -0.14497070014476776,
|
| 594 |
+
"logits/chosen": -1.611181616783142,
|
| 595 |
+
"logits/rejected": 15.831250190734863,
|
| 596 |
+
"logps/chosen": -1.0519530773162842,
|
| 597 |
+
"logps/rejected": -3.057812452316284,
|
| 598 |
+
"loss": 1.0883,
|
| 599 |
+
"nll_loss": 1.08203125,
|
| 600 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 601 |
+
"rewards/chosen": -0.0526123046875,
|
| 602 |
+
"rewards/margins": 0.10014648735523224,
|
| 603 |
+
"rewards/rejected": -0.15273436903953552,
|
| 604 |
"step": 330
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.43870967741935485,
|
| 608 |
+
"grad_norm": 2.9870987656093315,
|
| 609 |
+
"learning_rate": 5.4312544659356846e-06,
|
| 610 |
+
"log_odds_chosen": 2.4984374046325684,
|
| 611 |
+
"log_odds_ratio": -0.13447265326976776,
|
| 612 |
+
"logits/chosen": -1.1388671398162842,
|
| 613 |
+
"logits/rejected": 15.675000190734863,
|
| 614 |
+
"logps/chosen": -1.064062476158142,
|
| 615 |
+
"logps/rejected": -3.1781249046325684,
|
| 616 |
+
"loss": 1.0626,
|
| 617 |
+
"nll_loss": 1.0574219226837158,
|
| 618 |
+
"rewards/accuracies": 0.987500011920929,
|
| 619 |
+
"rewards/chosen": -0.05322265625,
|
| 620 |
+
"rewards/margins": 0.10561523586511612,
|
| 621 |
+
"rewards/rejected": -0.15878906846046448,
|
| 622 |
"step": 340
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 0.45161290322580644,
|
| 626 |
+
"grad_norm": 2.83867528933509,
|
| 627 |
+
"learning_rate": 5.352877275721893e-06,
|
| 628 |
+
"log_odds_chosen": 2.5406250953674316,
|
| 629 |
+
"log_odds_ratio": -0.14262695610523224,
|
| 630 |
+
"logits/chosen": -1.796777367591858,
|
| 631 |
+
"logits/rejected": 15.524999618530273,
|
| 632 |
+
"logps/chosen": -1.0109374523162842,
|
| 633 |
+
"logps/rejected": -3.1484375,
|
| 634 |
+
"loss": 1.0423,
|
| 635 |
+
"nll_loss": 1.03515625,
|
| 636 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 637 |
+
"rewards/chosen": -0.05056152492761612,
|
| 638 |
+
"rewards/margins": 0.10683593899011612,
|
| 639 |
+
"rewards/rejected": -0.15742187201976776,
|
| 640 |
"step": 350
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 0.4645161290322581,
|
| 644 |
+
"grad_norm": 2.9913815230527505,
|
| 645 |
+
"learning_rate": 5.277798139692596e-06,
|
| 646 |
+
"log_odds_chosen": 2.6484375,
|
| 647 |
+
"log_odds_ratio": -0.13837890326976776,
|
| 648 |
+
"logits/chosen": -1.791015625,
|
| 649 |
+
"logits/rejected": 15.574999809265137,
|
| 650 |
+
"logps/chosen": -1.0304687023162842,
|
| 651 |
+
"logps/rejected": -3.284374952316284,
|
| 652 |
+
"loss": 1.05,
|
| 653 |
+
"nll_loss": 1.0437500476837158,
|
| 654 |
+
"rewards/accuracies": 0.96875,
|
| 655 |
+
"rewards/chosen": -0.051513671875,
|
| 656 |
+
"rewards/margins": 0.11279296875,
|
| 657 |
+
"rewards/rejected": -0.16445311903953552,
|
| 658 |
"step": 360
|
| 659 |
},
|
| 660 |
{
|
| 661 |
"epoch": 0.4774193548387097,
|
| 662 |
+
"grad_norm": 6.430960743025478,
|
| 663 |
+
"learning_rate": 5.205792062953535e-06,
|
| 664 |
+
"log_odds_chosen": 2.629687547683716,
|
| 665 |
+
"log_odds_ratio": -0.14753417670726776,
|
| 666 |
+
"logits/chosen": -1.828125,
|
| 667 |
+
"logits/rejected": 15.9375,
|
| 668 |
+
"logps/chosen": -1.038671851158142,
|
| 669 |
+
"logps/rejected": -3.278125047683716,
|
| 670 |
+
"loss": 1.0531,
|
| 671 |
+
"nll_loss": 1.0457031726837158,
|
| 672 |
+
"rewards/accuracies": 0.96875,
|
| 673 |
+
"rewards/chosen": -0.05190429836511612,
|
| 674 |
+
"rewards/margins": 0.112060546875,
|
| 675 |
+
"rewards/rejected": -0.16396483778953552,
|
| 676 |
"step": 370
|
| 677 |
},
|
| 678 |
{
|
| 679 |
"epoch": 0.49032258064516127,
|
| 680 |
+
"grad_norm": 2.3749325025222587,
|
| 681 |
+
"learning_rate": 5.136654969375142e-06,
|
| 682 |
+
"log_odds_chosen": 2.176562547683716,
|
| 683 |
+
"log_odds_ratio": -0.16562500596046448,
|
| 684 |
+
"logits/chosen": -1.7742187976837158,
|
| 685 |
+
"logits/rejected": 15.612500190734863,
|
| 686 |
+
"logps/chosen": -1.0148437023162842,
|
| 687 |
+
"logps/rejected": -2.7984375953674316,
|
| 688 |
+
"loss": 1.0246,
|
| 689 |
+
"nll_loss": 1.015234351158142,
|
| 690 |
+
"rewards/accuracies": 0.987500011920929,
|
| 691 |
+
"rewards/chosen": -0.05070800706744194,
|
| 692 |
+
"rewards/margins": 0.08916015923023224,
|
| 693 |
+
"rewards/rejected": -0.139892578125,
|
| 694 |
"step": 380
|
| 695 |
},
|
| 696 |
{
|
| 697 |
"epoch": 0.5032258064516129,
|
| 698 |
+
"grad_norm": 2.9704201026587405,
|
| 699 |
+
"learning_rate": 5.070201265633938e-06,
|
| 700 |
+
"log_odds_chosen": 2.112499952316284,
|
| 701 |
+
"log_odds_ratio": -0.1689453125,
|
| 702 |
+
"logits/chosen": -1.837499976158142,
|
| 703 |
+
"logits/rejected": 15.824999809265137,
|
| 704 |
+
"logps/chosen": -1.0246093273162842,
|
| 705 |
+
"logps/rejected": -2.7671875953674316,
|
| 706 |
+
"loss": 1.0501,
|
| 707 |
+
"nll_loss": 1.0421874523162842,
|
| 708 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 709 |
+
"rewards/chosen": -0.05129394680261612,
|
| 710 |
+
"rewards/margins": 0.08686523139476776,
|
| 711 |
+
"rewards/rejected": -0.13833007216453552,
|
| 712 |
"step": 390
|
| 713 |
},
|
| 714 |
{
|
| 715 |
"epoch": 0.5161290322580645,
|
| 716 |
+
"grad_norm": 2.7572182399385303,
|
| 717 |
+
"learning_rate": 5.006261743217589e-06,
|
| 718 |
+
"log_odds_chosen": 2.305468797683716,
|
| 719 |
+
"log_odds_ratio": -0.14760741591453552,
|
| 720 |
+
"logits/chosen": -2.180468797683716,
|
| 721 |
+
"logits/rejected": 15.449999809265137,
|
| 722 |
+
"logps/chosen": -1.017968773841858,
|
| 723 |
+
"logps/rejected": -2.9203124046325684,
|
| 724 |
+
"loss": 1.0379,
|
| 725 |
+
"nll_loss": 1.030859351158142,
|
| 726 |
+
"rewards/accuracies": 0.987500011920929,
|
| 727 |
+
"rewards/chosen": -0.05087890475988388,
|
| 728 |
+
"rewards/margins": 0.09526367485523224,
|
| 729 |
+
"rewards/rejected": -0.14633789658546448,
|
| 730 |
"step": 400
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.5290322580645161,
|
| 734 |
+
"grad_norm": 2.988129143495499,
|
| 735 |
+
"learning_rate": 4.944681764341488e-06,
|
| 736 |
+
"log_odds_chosen": 2.643749952316284,
|
| 737 |
+
"log_odds_ratio": -0.12158203125,
|
| 738 |
+
"logits/chosen": -1.9125487804412842,
|
| 739 |
+
"logits/rejected": 15.543749809265137,
|
| 740 |
+
"logps/chosen": -1.0125000476837158,
|
| 741 |
+
"logps/rejected": -3.2484374046325684,
|
| 742 |
+
"loss": 1.0289,
|
| 743 |
+
"nll_loss": 1.023046851158142,
|
| 744 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 745 |
+
"rewards/chosen": -0.0506591796875,
|
| 746 |
+
"rewards/margins": 0.11171875149011612,
|
| 747 |
+
"rewards/rejected": -0.16245117783546448,
|
| 748 |
"step": 410
|
| 749 |
},
|
| 750 |
{
|
| 751 |
"epoch": 0.5419354838709678,
|
| 752 |
+
"grad_norm": 3.3731155020181154,
|
| 753 |
+
"learning_rate": 4.885319687460315e-06,
|
| 754 |
+
"log_odds_chosen": 2.682812452316284,
|
| 755 |
+
"log_odds_ratio": -0.14687499403953552,
|
| 756 |
+
"logits/chosen": -1.4832031726837158,
|
| 757 |
+
"logits/rejected": 15.387499809265137,
|
| 758 |
+
"logps/chosen": -1.058203101158142,
|
| 759 |
+
"logps/rejected": -3.375,
|
| 760 |
+
"loss": 1.0678,
|
| 761 |
+
"nll_loss": 1.060156226158142,
|
| 762 |
+
"rewards/accuracies": 0.956250011920929,
|
| 763 |
+
"rewards/chosen": -0.05280761793255806,
|
| 764 |
+
"rewards/margins": 0.115966796875,
|
| 765 |
+
"rewards/rejected": -0.16875000298023224,
|
| 766 |
"step": 420
|
| 767 |
},
|
| 768 |
{
|
| 769 |
"epoch": 0.5548387096774193,
|
| 770 |
+
"grad_norm": 11.540099842491083,
|
| 771 |
+
"learning_rate": 4.8280454958526765e-06,
|
| 772 |
+
"log_odds_chosen": 2.6937499046325684,
|
| 773 |
+
"log_odds_ratio": -0.13825683295726776,
|
| 774 |
+
"logits/chosen": -0.983593761920929,
|
| 775 |
+
"logits/rejected": 16.268749237060547,
|
| 776 |
+
"logps/chosen": -1.044531226158142,
|
| 777 |
+
"logps/rejected": -3.346874952316284,
|
| 778 |
+
"loss": 1.0426,
|
| 779 |
+
"nll_loss": 1.0363280773162842,
|
| 780 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 781 |
+
"rewards/chosen": -0.05219726637005806,
|
| 782 |
+
"rewards/margins": 0.11508788913488388,
|
| 783 |
+
"rewards/rejected": -0.16728515923023224,
|
| 784 |
"step": 430
|
| 785 |
},
|
| 786 |
{
|
| 787 |
"epoch": 0.567741935483871,
|
| 788 |
+
"grad_norm": 3.6791814532749805,
|
| 789 |
+
"learning_rate": 4.7727395990334775e-06,
|
| 790 |
+
"log_odds_chosen": 2.5210938453674316,
|
| 791 |
+
"log_odds_ratio": -0.14609375596046448,
|
| 792 |
+
"logits/chosen": -1.860937476158142,
|
| 793 |
+
"logits/rejected": 15.625,
|
| 794 |
+
"logps/chosen": -1.04296875,
|
| 795 |
+
"logps/rejected": -3.1859374046325684,
|
| 796 |
+
"loss": 1.0679,
|
| 797 |
+
"nll_loss": 1.060546875,
|
| 798 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 799 |
+
"rewards/chosen": -0.05214843899011612,
|
| 800 |
+
"rewards/margins": 0.10712890326976776,
|
| 801 |
+
"rewards/rejected": -0.15937499701976776,
|
| 802 |
"step": 440
|
| 803 |
},
|
| 804 |
{
|
| 805 |
"epoch": 0.5806451612903226,
|
| 806 |
+
"grad_norm": 2.7953578580150538,
|
| 807 |
+
"learning_rate": 4.719291781830088e-06,
|
| 808 |
+
"log_odds_chosen": 2.6546874046325684,
|
| 809 |
+
"log_odds_ratio": -0.14821776747703552,
|
| 810 |
+
"logits/chosen": -1.664709448814392,
|
| 811 |
+
"logits/rejected": 15.356249809265137,
|
| 812 |
+
"logps/chosen": -1.074609398841858,
|
| 813 |
+
"logps/rejected": -3.3515625,
|
| 814 |
+
"loss": 1.0897,
|
| 815 |
+
"nll_loss": 1.08203125,
|
| 816 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 817 |
+
"rewards/chosen": -0.05378418043255806,
|
| 818 |
+
"rewards/margins": 0.11376953125,
|
| 819 |
+
"rewards/rejected": -0.16757813096046448,
|
| 820 |
"step": 450
|
| 821 |
},
|
| 822 |
{
|
| 823 |
"epoch": 0.5935483870967742,
|
| 824 |
+
"grad_norm": 3.005212830785336,
|
| 825 |
+
"learning_rate": 4.667600280093366e-06,
|
| 826 |
+
"log_odds_chosen": 3.120312452316284,
|
| 827 |
+
"log_odds_ratio": -0.12114258110523224,
|
| 828 |
+
"logits/chosen": -1.77734375,
|
| 829 |
+
"logits/rejected": 15.712499618530273,
|
| 830 |
+
"logps/chosen": -1.0207030773162842,
|
| 831 |
+
"logps/rejected": -3.7265625,
|
| 832 |
+
"loss": 1.0381,
|
| 833 |
+
"nll_loss": 1.0324218273162842,
|
| 834 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 835 |
+
"rewards/chosen": -0.05104980617761612,
|
| 836 |
+
"rewards/margins": 0.13530273735523224,
|
| 837 |
+
"rewards/rejected": -0.18632812798023224,
|
| 838 |
"step": 460
|
| 839 |
},
|
| 840 |
{
|
| 841 |
"epoch": 0.6064516129032258,
|
| 842 |
+
"grad_norm": 3.459726587570506,
|
| 843 |
+
"learning_rate": 4.6175709653961015e-06,
|
| 844 |
+
"log_odds_chosen": 2.854687452316284,
|
| 845 |
+
"log_odds_ratio": -0.16249999403953552,
|
| 846 |
+
"logits/chosen": -1.89453125,
|
| 847 |
+
"logits/rejected": 15.081250190734863,
|
| 848 |
+
"logps/chosen": -1.0656249523162842,
|
| 849 |
+
"logps/rejected": -3.5640625953674316,
|
| 850 |
+
"loss": 1.0723,
|
| 851 |
+
"nll_loss": 1.0652344226837158,
|
| 852 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 853 |
+
"rewards/chosen": -0.05327148362994194,
|
| 854 |
+
"rewards/margins": 0.124755859375,
|
| 855 |
+
"rewards/rejected": -0.17812499403953552,
|
| 856 |
"step": 470
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.6193548387096774,
|
| 860 |
+
"grad_norm": 3.0057455144663665,
|
| 861 |
+
"learning_rate": 4.569116623849539e-06,
|
| 862 |
+
"log_odds_chosen": 3.1343750953674316,
|
| 863 |
+
"log_odds_ratio": -0.11713866889476776,
|
| 864 |
+
"logits/chosen": -1.96875,
|
| 865 |
+
"logits/rejected": 15.893750190734863,
|
| 866 |
+
"logps/chosen": -0.9652343988418579,
|
| 867 |
+
"logps/rejected": -3.667187452316284,
|
| 868 |
+
"loss": 0.9906,
|
| 869 |
+
"nll_loss": 0.984375,
|
| 870 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 871 |
+
"rewards/chosen": -0.04826660081744194,
|
| 872 |
+
"rewards/margins": 0.13496093451976776,
|
| 873 |
+
"rewards/rejected": -0.18330077826976776,
|
| 874 |
"step": 480
|
| 875 |
},
|
| 876 |
{
|
| 877 |
"epoch": 0.632258064516129,
|
| 878 |
+
"grad_norm": 3.351920982571353,
|
| 879 |
+
"learning_rate": 4.522156316461347e-06,
|
| 880 |
+
"log_odds_chosen": 2.721874952316284,
|
| 881 |
+
"log_odds_ratio": -0.12812499701976776,
|
| 882 |
+
"logits/chosen": -1.908203125,
|
| 883 |
+
"logits/rejected": 16.024999618530273,
|
| 884 |
+
"logps/chosen": -1.060937523841858,
|
| 885 |
+
"logps/rejected": -3.395312547683716,
|
| 886 |
+
"loss": 1.0772,
|
| 887 |
+
"nll_loss": 1.070703148841858,
|
| 888 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 889 |
+
"rewards/chosen": -0.05305175855755806,
|
| 890 |
+
"rewards/margins": 0.11660156399011612,
|
| 891 |
+
"rewards/rejected": -0.16972656548023224,
|
| 892 |
"step": 490
|
| 893 |
},
|
| 894 |
{
|
| 895 |
"epoch": 0.6451612903225806,
|
| 896 |
+
"grad_norm": 23.089712113718853,
|
| 897 |
+
"learning_rate": 4.476614810358452e-06,
|
| 898 |
+
"log_odds_chosen": 2.9468750953674316,
|
| 899 |
+
"log_odds_ratio": -0.14716796576976776,
|
| 900 |
+
"logits/chosen": -1.800390601158142,
|
| 901 |
+
"logits/rejected": 15.518750190734863,
|
| 902 |
+
"logps/chosen": -1.0910155773162842,
|
| 903 |
+
"logps/rejected": -3.6624999046325684,
|
| 904 |
+
"loss": 1.0891,
|
| 905 |
+
"nll_loss": 1.082421898841858,
|
| 906 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 907 |
+
"rewards/chosen": -0.05458984524011612,
|
| 908 |
+
"rewards/margins": 0.12851563096046448,
|
| 909 |
+
"rewards/rejected": -0.18310546875,
|
| 910 |
"step": 500
|
| 911 |
},
|
| 912 |
{
|
| 913 |
"epoch": 0.6580645161290323,
|
| 914 |
+
"grad_norm": 2.9945598420779853,
|
| 915 |
+
"learning_rate": 4.432422071779363e-06,
|
| 916 |
+
"log_odds_chosen": 3.317187547683716,
|
| 917 |
+
"log_odds_ratio": -0.10064697265625,
|
| 918 |
+
"logits/chosen": -2.010937452316284,
|
| 919 |
+
"logits/rejected": 15.393750190734863,
|
| 920 |
+
"logps/chosen": -0.9996093511581421,
|
| 921 |
+
"logps/rejected": -3.8921875953674316,
|
| 922 |
+
"loss": 1.0307,
|
| 923 |
+
"nll_loss": 1.02734375,
|
| 924 |
+
"rewards/accuracies": 0.987500011920929,
|
| 925 |
+
"rewards/chosen": -0.04997558519244194,
|
| 926 |
+
"rewards/margins": 0.14467772841453552,
|
| 927 |
+
"rewards/rejected": -0.19462890923023224,
|
| 928 |
"step": 510
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 0.6709677419354839,
|
| 932 |
+
"grad_norm": 7.310800315863408,
|
| 933 |
+
"learning_rate": 4.389512813061471e-06,
|
| 934 |
+
"log_odds_chosen": 3.265625,
|
| 935 |
+
"log_odds_ratio": -0.10830078274011612,
|
| 936 |
+
"logits/chosen": -2.186718702316284,
|
| 937 |
+
"logits/rejected": 15.712499618530273,
|
| 938 |
+
"logps/chosen": -1.0011718273162842,
|
| 939 |
+
"logps/rejected": -3.840625047683716,
|
| 940 |
+
"loss": 1.0184,
|
| 941 |
+
"nll_loss": 1.014062523841858,
|
| 942 |
+
"rewards/accuracies": 0.987500011920929,
|
| 943 |
+
"rewards/chosen": -0.05007324367761612,
|
| 944 |
+
"rewards/margins": 0.14218750596046448,
|
| 945 |
+
"rewards/rejected": -0.19218750298023224,
|
| 946 |
"step": 520
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 0.6838709677419355,
|
| 950 |
+
"grad_norm": 2.9539080812054275,
|
| 951 |
+
"learning_rate": 4.347826086956522e-06,
|
| 952 |
+
"log_odds_chosen": 3.0718750953674316,
|
| 953 |
+
"log_odds_ratio": -0.12153320014476776,
|
| 954 |
+
"logits/chosen": -0.94921875,
|
| 955 |
+
"logits/rejected": 15.306249618530273,
|
| 956 |
+
"logps/chosen": -0.98828125,
|
| 957 |
+
"logps/rejected": -3.637500047683716,
|
| 958 |
+
"loss": 1.0124,
|
| 959 |
+
"nll_loss": 1.005859375,
|
| 960 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 961 |
+
"rewards/chosen": -0.04941406100988388,
|
| 962 |
+
"rewards/margins": 0.13232421875,
|
| 963 |
+
"rewards/rejected": -0.18173828721046448,
|
| 964 |
"step": 530
|
| 965 |
},
|
| 966 |
{
|
| 967 |
"epoch": 0.6967741935483871,
|
| 968 |
+
"grad_norm": 2.8332863620261124,
|
| 969 |
+
"learning_rate": 4.3073049225394805e-06,
|
| 970 |
+
"log_odds_chosen": 3.135937452316284,
|
| 971 |
+
"log_odds_ratio": -0.1195068359375,
|
| 972 |
+
"logits/chosen": -1.21875,
|
| 973 |
+
"logits/rejected": 16.206249237060547,
|
| 974 |
+
"logps/chosen": -1.0390625,
|
| 975 |
+
"logps/rejected": -3.745312452316284,
|
| 976 |
+
"loss": 1.0254,
|
| 977 |
+
"nll_loss": 1.019140601158142,
|
| 978 |
+
"rewards/accuracies": 0.981249988079071,
|
| 979 |
+
"rewards/chosen": -0.05195312574505806,
|
| 980 |
+
"rewards/margins": 0.13520507514476776,
|
| 981 |
+
"rewards/rejected": -0.18710938096046448,
|
| 982 |
"step": 540
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.7096774193548387,
|
| 986 |
+
"grad_norm": 3.316803771710403,
|
| 987 |
+
"learning_rate": 4.2678959977632e-06,
|
| 988 |
+
"log_odds_chosen": 3.0453124046325684,
|
| 989 |
+
"log_odds_ratio": -0.105224609375,
|
| 990 |
+
"logits/chosen": -1.7107422351837158,
|
| 991 |
+
"logits/rejected": 15.649999618530273,
|
| 992 |
+
"logps/chosen": -1.0378906726837158,
|
| 993 |
+
"logps/rejected": -3.674999952316284,
|
| 994 |
+
"loss": 1.0426,
|
| 995 |
+
"nll_loss": 1.037500023841858,
|
| 996 |
"rewards/accuracies": 0.987500011920929,
|
| 997 |
+
"rewards/chosen": -0.05195312574505806,
|
| 998 |
+
"rewards/margins": 0.13188476860523224,
|
| 999 |
+
"rewards/rejected": -0.18369141221046448,
|
| 1000 |
"step": 550
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 0.7225806451612903,
|
| 1004 |
+
"grad_norm": 2.730881069069837,
|
| 1005 |
+
"learning_rate": 4.229549344378136e-06,
|
| 1006 |
+
"log_odds_chosen": 3.1968750953674316,
|
| 1007 |
+
"log_odds_ratio": -0.110595703125,
|
| 1008 |
+
"logits/chosen": -1.8035156726837158,
|
| 1009 |
+
"logits/rejected": 15.556249618530273,
|
| 1010 |
+
"logps/chosen": -1.029687523841858,
|
| 1011 |
+
"logps/rejected": -3.8203125,
|
| 1012 |
+
"loss": 1.0507,
|
| 1013 |
+
"nll_loss": 1.044921875,
|
| 1014 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1015 |
+
"rewards/chosen": -0.05146484449505806,
|
| 1016 |
+
"rewards/margins": 0.13959960639476776,
|
| 1017 |
+
"rewards/rejected": -0.19091796875,
|
| 1018 |
"step": 560
|
| 1019 |
},
|
| 1020 |
{
|
| 1021 |
"epoch": 0.7354838709677419,
|
| 1022 |
+
"grad_norm": 5.293665189129163,
|
| 1023 |
+
"learning_rate": 4.192218081503185e-06,
|
| 1024 |
+
"log_odds_chosen": 3.5,
|
| 1025 |
+
"log_odds_ratio": -0.08615722507238388,
|
| 1026 |
+
"logits/chosen": -1.8792235851287842,
|
| 1027 |
+
"logits/rejected": 15.631250381469727,
|
| 1028 |
+
"logps/chosen": -0.999218761920929,
|
| 1029 |
+
"logps/rejected": -4.064062595367432,
|
| 1030 |
+
"loss": 1.017,
|
| 1031 |
+
"nll_loss": 1.0125000476837158,
|
| 1032 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1033 |
+
"rewards/chosen": -0.04990234225988388,
|
| 1034 |
+
"rewards/margins": 0.15317383408546448,
|
| 1035 |
+
"rewards/rejected": -0.203125,
|
| 1036 |
"step": 570
|
| 1037 |
},
|
| 1038 |
{
|
| 1039 |
"epoch": 0.7483870967741936,
|
| 1040 |
+
"grad_norm": 2.7468345646696406,
|
| 1041 |
+
"learning_rate": 4.155858174616914e-06,
|
| 1042 |
+
"log_odds_chosen": 3.2906250953674316,
|
| 1043 |
+
"log_odds_ratio": -0.0904541015625,
|
| 1044 |
+
"logits/chosen": -2.1302733421325684,
|
| 1045 |
+
"logits/rejected": 15.012499809265137,
|
| 1046 |
+
"logps/chosen": -0.977343738079071,
|
| 1047 |
+
"logps/rejected": -3.8218750953674316,
|
| 1048 |
+
"loss": 1.0009,
|
| 1049 |
+
"nll_loss": 0.997265636920929,
|
| 1050 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1051 |
+
"rewards/chosen": -0.048828125,
|
| 1052 |
+
"rewards/margins": 0.14199218153953552,
|
| 1053 |
+
"rewards/rejected": -0.19082030653953552,
|
| 1054 |
"step": 580
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
"epoch": 0.7612903225806451,
|
| 1058 |
+
"grad_norm": 3.0853307261867635,
|
| 1059 |
+
"learning_rate": 4.120428217151647e-06,
|
| 1060 |
+
"log_odds_chosen": 3.4124999046325684,
|
| 1061 |
+
"log_odds_ratio": -0.09726562350988388,
|
| 1062 |
+
"logits/chosen": -2.3573241233825684,
|
| 1063 |
+
"logits/rejected": 15.337499618530273,
|
| 1064 |
+
"logps/chosen": -0.975390613079071,
|
| 1065 |
+
"logps/rejected": -3.9468750953674316,
|
| 1066 |
+
"loss": 1.0064,
|
| 1067 |
+
"nll_loss": 1.0011718273162842,
|
| 1068 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1069 |
+
"rewards/chosen": -0.0487060546875,
|
| 1070 |
+
"rewards/margins": 0.14853516221046448,
|
| 1071 |
+
"rewards/rejected": -0.197265625,
|
| 1072 |
"step": 590
|
| 1073 |
},
|
| 1074 |
{
|
| 1075 |
"epoch": 0.7741935483870968,
|
| 1076 |
+
"grad_norm": 2.7361001609622426,
|
| 1077 |
+
"learning_rate": 4.0858892322271844e-06,
|
| 1078 |
+
"log_odds_chosen": 3.567187547683716,
|
| 1079 |
+
"log_odds_ratio": -0.08948974311351776,
|
| 1080 |
+
"logits/chosen": -1.8064453601837158,
|
| 1081 |
+
"logits/rejected": 15.625,
|
| 1082 |
+
"logps/chosen": -1.0207030773162842,
|
| 1083 |
+
"logps/rejected": -4.165625095367432,
|
| 1084 |
+
"loss": 1.0415,
|
| 1085 |
+
"nll_loss": 1.035546898841858,
|
| 1086 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1087 |
+
"rewards/chosen": -0.05104980617761612,
|
| 1088 |
+
"rewards/margins": 0.15751953423023224,
|
| 1089 |
+
"rewards/rejected": -0.20839843153953552,
|
| 1090 |
"step": 600
|
| 1091 |
},
|
| 1092 |
{
|
| 1093 |
"epoch": 0.7870967741935484,
|
| 1094 |
+
"grad_norm": 3.4721246535677714,
|
| 1095 |
+
"learning_rate": 4.052204492365539e-06,
|
| 1096 |
+
"log_odds_chosen": 3.9296875,
|
| 1097 |
+
"log_odds_ratio": -0.07633056491613388,
|
| 1098 |
+
"logits/chosen": -2.706249952316284,
|
| 1099 |
+
"logits/rejected": 15.518750190734863,
|
| 1100 |
+
"logps/chosen": -1.033593773841858,
|
| 1101 |
+
"logps/rejected": -4.540625095367432,
|
| 1102 |
+
"loss": 1.0518,
|
| 1103 |
+
"nll_loss": 1.0476562976837158,
|
| 1104 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1105 |
+
"rewards/chosen": -0.0516357421875,
|
| 1106 |
+
"rewards/margins": 0.17529296875,
|
| 1107 |
+
"rewards/rejected": -0.22685547173023224,
|
| 1108 |
"step": 610
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.8,
|
| 1112 |
+
"grad_norm": 3.6600611060580515,
|
| 1113 |
+
"learning_rate": 4.019339355290704e-06,
|
| 1114 |
+
"log_odds_chosen": 3.6156249046325684,
|
| 1115 |
+
"log_odds_ratio": -0.1014404296875,
|
| 1116 |
+
"logits/chosen": -2.306640625,
|
| 1117 |
+
"logits/rejected": 15.862500190734863,
|
| 1118 |
+
"logps/chosen": -1.014062523841858,
|
| 1119 |
+
"logps/rejected": -4.1953125,
|
| 1120 |
+
"loss": 1.0323,
|
| 1121 |
+
"nll_loss": 1.027734398841858,
|
| 1122 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1123 |
+
"rewards/chosen": -0.05063476413488388,
|
| 1124 |
+
"rewards/margins": 0.15908202528953552,
|
| 1125 |
+
"rewards/rejected": -0.20966796576976776,
|
| 1126 |
"step": 620
|
| 1127 |
},
|
| 1128 |
{
|
| 1129 |
"epoch": 0.8129032258064516,
|
| 1130 |
+
"grad_norm": 3.044997462446035,
|
| 1131 |
+
"learning_rate": 3.9872611141445005e-06,
|
| 1132 |
+
"log_odds_chosen": 3.4546875953674316,
|
| 1133 |
+
"log_odds_ratio": -0.09210205078125,
|
| 1134 |
+
"logits/chosen": -2.1363282203674316,
|
| 1135 |
+
"logits/rejected": 15.725000381469727,
|
| 1136 |
+
"logps/chosen": -0.983203113079071,
|
| 1137 |
+
"logps/rejected": -3.9921875,
|
| 1138 |
+
"loss": 0.9936,
|
| 1139 |
+
"nll_loss": 0.9898437261581421,
|
| 1140 |
+
"rewards/accuracies": 1.0,
|
| 1141 |
+
"rewards/chosen": -0.04914550855755806,
|
| 1142 |
+
"rewards/margins": 0.15019531548023224,
|
| 1143 |
+
"rewards/rejected": -0.19951172173023224,
|
| 1144 |
"step": 630
|
| 1145 |
},
|
| 1146 |
{
|
| 1147 |
"epoch": 0.8258064516129032,
|
| 1148 |
+
"grad_norm": 3.972247288109564,
|
| 1149 |
+
"learning_rate": 3.955938860646178e-06,
|
| 1150 |
+
"log_odds_chosen": 3.6546874046325684,
|
| 1151 |
+
"log_odds_ratio": -0.09271240234375,
|
| 1152 |
+
"logits/chosen": -1.830078125,
|
| 1153 |
+
"logits/rejected": 15.8125,
|
| 1154 |
+
"logps/chosen": -1.047265648841858,
|
| 1155 |
+
"logps/rejected": -4.262499809265137,
|
| 1156 |
+
"loss": 1.0426,
|
| 1157 |
+
"nll_loss": 1.0382812023162842,
|
| 1158 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1159 |
+
"rewards/chosen": -0.0523681640625,
|
| 1160 |
+
"rewards/margins": 0.16064453125,
|
| 1161 |
+
"rewards/rejected": -0.21318359673023224,
|
| 1162 |
"step": 640
|
| 1163 |
},
|
| 1164 |
{
|
| 1165 |
"epoch": 0.8387096774193549,
|
| 1166 |
+
"grad_norm": 3.2312016718085466,
|
| 1167 |
+
"learning_rate": 3.925343359894298e-06,
|
| 1168 |
+
"log_odds_chosen": 4.076562404632568,
|
| 1169 |
+
"log_odds_ratio": -0.06367187201976776,
|
| 1170 |
+
"logits/chosen": -1.952734351158142,
|
| 1171 |
+
"logits/rejected": 16.381250381469727,
|
| 1172 |
+
"logps/chosen": -1.0363280773162842,
|
| 1173 |
+
"logps/rejected": -4.678124904632568,
|
| 1174 |
+
"loss": 1.0545,
|
| 1175 |
+
"nll_loss": 1.051171898841858,
|
| 1176 |
"rewards/accuracies": 0.987500011920929,
|
| 1177 |
+
"rewards/chosen": -0.05180663987994194,
|
| 1178 |
+
"rewards/margins": 0.18183593451976776,
|
| 1179 |
+
"rewards/rejected": -0.23378905653953552,
|
| 1180 |
"step": 650
|
| 1181 |
},
|
| 1182 |
{
|
| 1183 |
"epoch": 0.8516129032258064,
|
| 1184 |
+
"grad_norm": 3.1542224686668425,
|
| 1185 |
+
"learning_rate": 3.895446935658099e-06,
|
| 1186 |
+
"log_odds_chosen": 3.5687499046325684,
|
| 1187 |
+
"log_odds_ratio": -0.10651855170726776,
|
| 1188 |
+
"logits/chosen": -1.4596679210662842,
|
| 1189 |
+
"logits/rejected": 15.725000381469727,
|
| 1190 |
+
"logps/chosen": -1.042578101158142,
|
| 1191 |
+
"logps/rejected": -4.209374904632568,
|
| 1192 |
+
"loss": 1.047,
|
| 1193 |
+
"nll_loss": 1.041015625,
|
| 1194 |
+
"rewards/accuracies": 0.96875,
|
| 1195 |
+
"rewards/chosen": -0.05217285081744194,
|
| 1196 |
+
"rewards/margins": 0.15805664658546448,
|
| 1197 |
+
"rewards/rejected": -0.21044921875,
|
| 1198 |
"step": 660
|
| 1199 |
},
|
| 1200 |
{
|
| 1201 |
"epoch": 0.864516129032258,
|
| 1202 |
+
"grad_norm": 2.76746682960287,
|
| 1203 |
+
"learning_rate": 3.866223365135295e-06,
|
| 1204 |
+
"log_odds_chosen": 3.481250047683716,
|
| 1205 |
+
"log_odds_ratio": -0.09975586086511612,
|
| 1206 |
+
"logits/chosen": -2.1744141578674316,
|
| 1207 |
+
"logits/rejected": 15.537500381469727,
|
| 1208 |
+
"logps/chosen": -1.017968773841858,
|
| 1209 |
+
"logps/rejected": -4.071875095367432,
|
| 1210 |
+
"loss": 1.0355,
|
| 1211 |
+
"nll_loss": 1.03125,
|
| 1212 |
"rewards/accuracies": 0.981249988079071,
|
| 1213 |
+
"rewards/chosen": -0.0509033203125,
|
| 1214 |
+
"rewards/margins": 0.15253905951976776,
|
| 1215 |
+
"rewards/rejected": -0.20341797173023224,
|
| 1216 |
"step": 670
|
| 1217 |
},
|
| 1218 |
{
|
| 1219 |
"epoch": 0.8774193548387097,
|
| 1220 |
+
"grad_norm": 3.339669169989708,
|
| 1221 |
+
"learning_rate": 3.837647782266681e-06,
|
| 1222 |
+
"log_odds_chosen": 3.8062500953674316,
|
| 1223 |
+
"log_odds_ratio": -0.08485107123851776,
|
| 1224 |
+
"logits/chosen": -1.7482421398162842,
|
| 1225 |
+
"logits/rejected": 15.668749809265137,
|
| 1226 |
+
"logps/chosen": -1.081640601158142,
|
| 1227 |
+
"logps/rejected": -4.474999904632568,
|
| 1228 |
+
"loss": 1.0689,
|
| 1229 |
+
"nll_loss": 1.064062476158142,
|
| 1230 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1231 |
+
"rewards/chosen": -0.05405273288488388,
|
| 1232 |
+
"rewards/margins": 0.16982421278953552,
|
| 1233 |
+
"rewards/rejected": -0.22402343153953552,
|
| 1234 |
"step": 680
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8903225806451613,
|
| 1238 |
+
"grad_norm": 3.216293567417128,
|
| 1239 |
+
"learning_rate": 3.8096965887972973e-06,
|
| 1240 |
+
"log_odds_chosen": 3.6656250953674316,
|
| 1241 |
+
"log_odds_ratio": -0.1185302734375,
|
| 1242 |
+
"logits/chosen": -1.7449219226837158,
|
| 1243 |
+
"logits/rejected": 15.34375,
|
| 1244 |
+
"logps/chosen": -1.060937523841858,
|
| 1245 |
+
"logps/rejected": -4.292187690734863,
|
| 1246 |
+
"loss": 1.0248,
|
| 1247 |
+
"nll_loss": 1.019921898841858,
|
| 1248 |
"rewards/accuracies": 0.9750000238418579,
|
| 1249 |
+
"rewards/chosen": -0.05307617038488388,
|
| 1250 |
+
"rewards/margins": 0.16152343153953552,
|
| 1251 |
+
"rewards/rejected": -0.21445313096046448,
|
| 1252 |
"step": 690
|
| 1253 |
},
|
| 1254 |
{
|
| 1255 |
"epoch": 0.9032258064516129,
|
| 1256 |
+
"grad_norm": 3.165994900484303,
|
| 1257 |
+
"learning_rate": 3.7823473723611694e-06,
|
| 1258 |
+
"log_odds_chosen": 3.528125047683716,
|
| 1259 |
+
"log_odds_ratio": -0.09687499701976776,
|
| 1260 |
+
"logits/chosen": -1.868749976158142,
|
| 1261 |
+
"logits/rejected": 15.487500190734863,
|
| 1262 |
+
"logps/chosen": -0.983593761920929,
|
| 1263 |
+
"logps/rejected": -4.068749904632568,
|
| 1264 |
+
"loss": 0.9975,
|
| 1265 |
+
"nll_loss": 0.993359386920929,
|
| 1266 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1267 |
+
"rewards/chosen": -0.0491943359375,
|
| 1268 |
+
"rewards/margins": 0.154296875,
|
| 1269 |
+
"rewards/rejected": -0.20351561903953552,
|
| 1270 |
"step": 700
|
| 1271 |
},
|
| 1272 |
{
|
| 1273 |
"epoch": 0.9161290322580645,
|
| 1274 |
+
"grad_norm": 3.1804898751694877,
|
| 1275 |
+
"learning_rate": 3.7555788309433988e-06,
|
| 1276 |
+
"log_odds_chosen": 3.59375,
|
| 1277 |
+
"log_odds_ratio": -0.10922851413488388,
|
| 1278 |
+
"logits/chosen": -2.244140625,
|
| 1279 |
+
"logits/rejected": 15.425000190734863,
|
| 1280 |
+
"logps/chosen": -1.0078125,
|
| 1281 |
+
"logps/rejected": -4.1875,
|
| 1282 |
+
"loss": 1.0329,
|
| 1283 |
+
"nll_loss": 1.0265624523162842,
|
| 1284 |
+
"rewards/accuracies": 0.96875,
|
| 1285 |
+
"rewards/chosen": -0.0504150390625,
|
| 1286 |
+
"rewards/margins": 0.15878906846046448,
|
| 1287 |
+
"rewards/rejected": -0.20937499403953552,
|
| 1288 |
"step": 710
|
| 1289 |
},
|
| 1290 |
{
|
| 1291 |
"epoch": 0.9290322580645162,
|
| 1292 |
+
"grad_norm": 2.8833066776928478,
|
| 1293 |
+
"learning_rate": 3.729370703141008e-06,
|
| 1294 |
+
"log_odds_chosen": 3.8125,
|
| 1295 |
+
"log_odds_ratio": -0.07578124850988388,
|
| 1296 |
+
"logits/chosen": -1.795996069908142,
|
| 1297 |
+
"logits/rejected": 15.637499809265137,
|
| 1298 |
+
"logps/chosen": -0.9906250238418579,
|
| 1299 |
+
"logps/rejected": -4.357812404632568,
|
| 1300 |
+
"loss": 1.0145,
|
| 1301 |
+
"nll_loss": 1.0105469226837158,
|
| 1302 |
+
"rewards/accuracies": 1.0,
|
| 1303 |
+
"rewards/chosen": -0.04951171949505806,
|
| 1304 |
+
"rewards/margins": 0.16845703125,
|
| 1305 |
+
"rewards/rejected": -0.21787109971046448,
|
| 1306 |
"step": 720
|
| 1307 |
},
|
| 1308 |
{
|
| 1309 |
"epoch": 0.9419354838709677,
|
| 1310 |
+
"grad_norm": 3.5368171320315573,
|
| 1311 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 1312 |
+
"log_odds_chosen": 3.534374952316284,
|
| 1313 |
+
"log_odds_ratio": -0.09101562201976776,
|
| 1314 |
+
"logits/chosen": -2.108203172683716,
|
| 1315 |
+
"logits/rejected": 15.462499618530273,
|
| 1316 |
+
"logps/chosen": -1.011328101158142,
|
| 1317 |
+
"logps/rejected": -4.114062309265137,
|
| 1318 |
+
"loss": 1.0332,
|
| 1319 |
+
"nll_loss": 1.029296875,
|
| 1320 |
"rewards/accuracies": 0.987500011920929,
|
| 1321 |
+
"rewards/chosen": -0.05058593675494194,
|
| 1322 |
+
"rewards/margins": 0.15507812798023224,
|
| 1323 |
+
"rewards/rejected": -0.20576171576976776,
|
| 1324 |
"step": 730
|
| 1325 |
},
|
| 1326 |
{
|
| 1327 |
"epoch": 0.9548387096774194,
|
| 1328 |
+
"grad_norm": 2.88455714823336,
|
| 1329 |
+
"learning_rate": 3.6785594638885484e-06,
|
| 1330 |
+
"log_odds_chosen": 3.4359374046325684,
|
| 1331 |
+
"log_odds_ratio": -0.09562988579273224,
|
| 1332 |
+
"logits/chosen": -2.369335889816284,
|
| 1333 |
+
"logits/rejected": 15.0625,
|
| 1334 |
+
"logps/chosen": -0.989062488079071,
|
| 1335 |
+
"logps/rejected": -3.987499952316284,
|
| 1336 |
+
"loss": 0.9878,
|
| 1337 |
+
"nll_loss": 0.9839843511581421,
|
| 1338 |
"rewards/accuracies": 0.987500011920929,
|
| 1339 |
+
"rewards/chosen": -0.0494384765625,
|
| 1340 |
+
"rewards/margins": 0.14960937201976776,
|
| 1341 |
+
"rewards/rejected": -0.19912108778953552,
|
| 1342 |
"step": 740
|
| 1343 |
},
|
| 1344 |
{
|
| 1345 |
"epoch": 0.967741935483871,
|
| 1346 |
+
"grad_norm": 3.3000700003310293,
|
| 1347 |
+
"learning_rate": 3.6539204762093485e-06,
|
| 1348 |
+
"log_odds_chosen": 3.807812452316284,
|
| 1349 |
+
"log_odds_ratio": -0.07308349758386612,
|
| 1350 |
+
"logits/chosen": -2.13671875,
|
| 1351 |
+
"logits/rejected": 15.693750381469727,
|
| 1352 |
+
"logps/chosen": -1.0265624523162842,
|
| 1353 |
+
"logps/rejected": -4.396874904632568,
|
| 1354 |
+
"loss": 1.0483,
|
| 1355 |
+
"nll_loss": 1.044921875,
|
| 1356 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1357 |
+
"rewards/chosen": -0.05136718600988388,
|
| 1358 |
+
"rewards/margins": 0.16835936903953552,
|
| 1359 |
+
"rewards/rejected": -0.21962890028953552,
|
| 1360 |
"step": 750
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.9806451612903225,
|
| 1364 |
+
"grad_norm": 3.1320983844464814,
|
| 1365 |
+
"learning_rate": 3.6297700432031386e-06,
|
| 1366 |
+
"log_odds_chosen": 3.5453124046325684,
|
| 1367 |
+
"log_odds_ratio": -0.101806640625,
|
| 1368 |
+
"logits/chosen": -1.9267578125,
|
| 1369 |
+
"logits/rejected": 15.40625,
|
| 1370 |
+
"logps/chosen": -1.008203148841858,
|
| 1371 |
+
"logps/rejected": -4.128125190734863,
|
| 1372 |
+
"loss": 1.0326,
|
| 1373 |
+
"nll_loss": 1.0285155773162842,
|
| 1374 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1375 |
+
"rewards/chosen": -0.05043945461511612,
|
| 1376 |
+
"rewards/margins": 0.15605469048023224,
|
| 1377 |
+
"rewards/rejected": -0.20624999701976776,
|
| 1378 |
"step": 760
|
| 1379 |
},
|
| 1380 |
{
|
| 1381 |
"epoch": 0.9935483870967742,
|
| 1382 |
+
"grad_norm": 5.872666548921313,
|
| 1383 |
+
"learning_rate": 3.606092229873094e-06,
|
| 1384 |
+
"log_odds_chosen": 3.6624999046325684,
|
| 1385 |
+
"log_odds_ratio": -0.09074707329273224,
|
| 1386 |
+
"logits/chosen": -2.439453125,
|
| 1387 |
+
"logits/rejected": 15.293749809265137,
|
| 1388 |
+
"logps/chosen": -1.0261719226837158,
|
| 1389 |
+
"logps/rejected": -4.246874809265137,
|
| 1390 |
+
"loss": 1.0374,
|
| 1391 |
+
"nll_loss": 1.0324218273162842,
|
| 1392 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1393 |
+
"rewards/chosen": -0.05126953125,
|
| 1394 |
+
"rewards/margins": 0.16127929091453552,
|
| 1395 |
+
"rewards/rejected": -0.21230468153953552,
|
| 1396 |
"step": 770
|
| 1397 |
},
|
| 1398 |
{
|
| 1399 |
"epoch": 1.0,
|
| 1400 |
"step": 775,
|
| 1401 |
"total_flos": 0.0,
|
| 1402 |
+
"train_loss": 1.1359047379032259,
|
| 1403 |
+
"train_runtime": 1458.2503,
|
| 1404 |
+
"train_samples_per_second": 8.503,
|
| 1405 |
+
"train_steps_per_second": 0.531
|
| 1406 |
}
|
| 1407 |
],
|
| 1408 |
"logging_steps": 10,
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 7032
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a359549e54a9062c7fb67a97107aebc3c832827e266a3b5b10f4639a22841b97
|
| 3 |
size 7032
|