Model save
Browse files- README.md +1 -1
- adapter_config.json +2 -2
- adapter_model.safetensors +1 -1
- all_results.json +4 -4
- runs/May09_01-29-54_sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute/events.out.tfevents.1746754269.sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute.103671.0 +3 -0
- train_results.json +4 -4
- trainer_state.json +1074 -1074
- 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/3to80ym3)
|
| 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_config.json
CHANGED
|
@@ -24,8 +24,8 @@
|
|
| 24 |
"rank_pattern": {},
|
| 25 |
"revision": null,
|
| 26 |
"target_modules": [
|
| 27 |
-
"
|
| 28 |
-
"
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM",
|
| 31 |
"trainable_token_indices": null,
|
|
|
|
| 24 |
"rank_pattern": {},
|
| 25 |
"revision": null,
|
| 26 |
"target_modules": [
|
| 27 |
+
"v_proj",
|
| 28 |
+
"q_proj"
|
| 29 |
],
|
| 30 |
"task_type": "CAUSAL_LM",
|
| 31 |
"trainable_token_indices": null,
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 71588104
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b40ac4fb9ca6db92eb1c8f97fccebeedd404d9b4ea366bea25cde654b788caff
|
| 3 |
size 71588104
|
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 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.0746982106854839,
|
| 5 |
+
"train_runtime": 1520.1443,
|
| 6 |
"train_samples": 12400,
|
| 7 |
+
"train_samples_per_second": 8.157,
|
| 8 |
+
"train_steps_per_second": 0.51
|
| 9 |
}
|
runs/May09_01-29-54_sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute/events.out.tfevents.1746754269.sky-3542-mykola-5e9d8a2f-head-4qgcm7zi-compute.103671.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f53a2af8f14c03425388f7ce02892f00fda987681a5e9f2a2b1b5da737699304
|
| 3 |
+
size 72147
|
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 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 1.0746982106854839,
|
| 5 |
+
"train_runtime": 1520.1443,
|
| 6 |
"train_samples": 12400,
|
| 7 |
+
"train_samples_per_second": 8.157,
|
| 8 |
+
"train_steps_per_second": 0.51
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -11,1398 +11,1398 @@
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.012903225806451613,
|
| 14 |
-
"grad_norm":
|
| 15 |
-
"learning_rate":
|
| 16 |
-
"log_odds_chosen": 1.
|
| 17 |
-
"log_odds_ratio": -0.
|
| 18 |
-
"logits/chosen": -1.
|
| 19 |
-
"logits/rejected": 15.
|
| 20 |
-
"logps/chosen": -1.
|
| 21 |
-
"logps/rejected": -2.
|
| 22 |
-
"loss": 2.
|
| 23 |
-
"nll_loss": 1.
|
| 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":
|
| 33 |
-
"learning_rate":
|
| 34 |
-
"log_odds_chosen": 1.
|
| 35 |
-
"log_odds_ratio": -0.
|
| 36 |
-
"logits/chosen": -
|
| 37 |
-
"logits/rejected": 16.
|
| 38 |
-
"logps/chosen": -1.
|
| 39 |
-
"logps/rejected": -2.
|
| 40 |
-
"loss":
|
| 41 |
-
"nll_loss": 1.
|
| 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":
|
| 51 |
-
"learning_rate":
|
| 52 |
-
"log_odds_chosen":
|
| 53 |
-
"log_odds_ratio": -0.
|
| 54 |
-
"logits/chosen": -
|
| 55 |
-
"logits/rejected": 16.
|
| 56 |
-
"logps/chosen": -1.
|
| 57 |
-
"logps/rejected": -2.
|
| 58 |
-
"loss":
|
| 59 |
-
"nll_loss": 1.
|
| 60 |
-
"rewards/accuracies": 0.
|
| 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": 0.
|
| 71 |
-
"log_odds_ratio": -0.
|
| 72 |
-
"logits/chosen":
|
| 73 |
-
"logits/rejected":
|
| 74 |
-
"logps/chosen": -1.
|
| 75 |
-
"logps/rejected": -
|
| 76 |
-
"loss": 1.
|
| 77 |
-
"nll_loss": 1.
|
| 78 |
-
"rewards/accuracies": 0.
|
| 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":
|
| 92 |
-
"logps/chosen": -1.
|
| 93 |
-
"logps/rejected": -
|
| 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": 0.
|
| 127 |
-
"logits/rejected":
|
| 128 |
-
"logps/chosen": -1.
|
| 129 |
-
"logps/rejected": -1.
|
| 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": 0.
|
| 145 |
-
"logits/rejected":
|
| 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":
|
| 161 |
-
"log_odds_ratio": -0.
|
| 162 |
-
"logits/chosen":
|
| 163 |
-
"logits/rejected":
|
| 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":
|
| 179 |
-
"log_odds_ratio": -0.
|
| 180 |
-
"logits/chosen":
|
| 181 |
-
"logits/rejected":
|
| 182 |
-
"logps/chosen": -1.
|
| 183 |
-
"logps/rejected": -
|
| 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":
|
| 195 |
-
"learning_rate":
|
| 196 |
-
"log_odds_chosen":
|
| 197 |
-
"log_odds_ratio": -0.
|
| 198 |
-
"logits/chosen":
|
| 199 |
-
"logits/rejected":
|
| 200 |
-
"logps/chosen": -1.
|
| 201 |
-
"logps/rejected": -
|
| 202 |
-
"loss": 1.
|
| 203 |
-
"nll_loss": 1.
|
| 204 |
-
"rewards/accuracies": 0.
|
| 205 |
-
"rewards/chosen": -0.
|
| 206 |
-
"rewards/margins": 0.
|
| 207 |
-
"rewards/rejected": -0.
|
| 208 |
"step": 110
|
| 209 |
},
|
| 210 |
{
|
| 211 |
"epoch": 0.15483870967741936,
|
| 212 |
-
"grad_norm":
|
| 213 |
-
"learning_rate":
|
| 214 |
-
"log_odds_chosen":
|
| 215 |
-
"log_odds_ratio": -0.
|
| 216 |
-
"logits/chosen":
|
| 217 |
-
"logits/rejected":
|
| 218 |
-
"logps/chosen": -1.
|
| 219 |
-
"logps/rejected": -
|
| 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":
|
| 231 |
-
"learning_rate":
|
| 232 |
-
"log_odds_chosen":
|
| 233 |
-
"log_odds_ratio": -0.
|
| 234 |
-
"logits/chosen": 0.
|
| 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":
|
| 249 |
-
"learning_rate":
|
| 250 |
-
"log_odds_chosen":
|
| 251 |
-
"log_odds_ratio": -0.
|
| 252 |
-
"logits/chosen": -
|
| 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":
|
| 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":
|
| 285 |
-
"learning_rate":
|
| 286 |
-
"log_odds_chosen":
|
| 287 |
-
"log_odds_ratio": -0.
|
| 288 |
-
"logits/chosen": -
|
| 289 |
-
"logits/rejected":
|
| 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": 4.
|
| 303 |
-
"learning_rate":
|
| 304 |
-
"log_odds_chosen":
|
| 305 |
-
"log_odds_ratio": -0.
|
| 306 |
-
"logits/chosen": -
|
| 307 |
-
"logits/rejected": 16.
|
| 308 |
-
"logps/chosen": -
|
| 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": 5.
|
| 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": -
|
| 343 |
-
"logits/rejected":
|
| 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": 4.
|
| 357 |
-
"learning_rate":
|
| 358 |
-
"log_odds_chosen":
|
| 359 |
-
"log_odds_ratio": -0.
|
| 360 |
-
"logits/chosen": -1.
|
| 361 |
-
"logits/rejected": 16.
|
| 362 |
-
"logps/chosen": -1.
|
| 363 |
-
"logps/rejected": -
|
| 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":
|
| 375 |
-
"learning_rate":
|
| 376 |
-
"log_odds_chosen":
|
| 377 |
-
"log_odds_ratio": -0.
|
| 378 |
-
"logits/chosen": -
|
| 379 |
-
"logits/rejected": 15.
|
| 380 |
-
"logps/chosen": -1.
|
| 381 |
-
"logps/rejected": -
|
| 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":
|
| 393 |
-
"learning_rate":
|
| 394 |
-
"log_odds_chosen":
|
| 395 |
-
"log_odds_ratio": -0.
|
| 396 |
-
"logits/chosen": -
|
| 397 |
-
"logits/rejected":
|
| 398 |
-
"logps/chosen": -1.
|
| 399 |
-
"logps/rejected": -
|
| 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":
|
| 411 |
-
"learning_rate":
|
| 412 |
-
"log_odds_chosen":
|
| 413 |
-
"log_odds_ratio": -0.
|
| 414 |
-
"logits/chosen": -
|
| 415 |
-
"logits/rejected": 15.
|
| 416 |
-
"logps/chosen": -1.
|
| 417 |
-
"logps/rejected": -2.
|
| 418 |
-
"loss": 1.
|
| 419 |
-
"nll_loss": 1.
|
| 420 |
-
"rewards/accuracies": 0.
|
| 421 |
-
"rewards/chosen": -0.
|
| 422 |
-
"rewards/margins": 0.
|
| 423 |
-
"rewards/rejected": -0.
|
| 424 |
"step": 230
|
| 425 |
},
|
| 426 |
{
|
| 427 |
"epoch": 0.3096774193548387,
|
| 428 |
-
"grad_norm": 4.
|
| 429 |
-
"learning_rate":
|
| 430 |
-
"log_odds_chosen":
|
| 431 |
-
"log_odds_ratio": -0.
|
| 432 |
-
"logits/chosen": -
|
| 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": 4.
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"log_odds_chosen":
|
| 449 |
-
"log_odds_ratio": -0.
|
| 450 |
-
"logits/chosen": -2.
|
| 451 |
-
"logits/rejected": 15.
|
| 452 |
-
"logps/chosen": -
|
| 453 |
-
"logps/rejected": -
|
| 454 |
-
"loss": 1.
|
| 455 |
-
"nll_loss": 1.
|
| 456 |
"rewards/accuracies": 0.981249988079071,
|
| 457 |
-
"rewards/chosen": -0.
|
| 458 |
-
"rewards/margins": 0.
|
| 459 |
-
"rewards/rejected": -0.
|
| 460 |
"step": 250
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 0.33548387096774196,
|
| 464 |
-
"grad_norm": 4.
|
| 465 |
-
"learning_rate":
|
| 466 |
-
"log_odds_chosen":
|
| 467 |
-
"log_odds_ratio": -0.
|
| 468 |
-
"logits/chosen": -2.
|
| 469 |
-
"logits/rejected": 15.
|
| 470 |
-
"logps/chosen": -1.
|
| 471 |
-
"logps/rejected": -
|
| 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":
|
| 483 |
-
"learning_rate":
|
| 484 |
-
"log_odds_chosen":
|
| 485 |
-
"log_odds_ratio": -0.
|
| 486 |
-
"logits/chosen": -
|
| 487 |
-
"logits/rejected": 15.
|
| 488 |
-
"logps/chosen": -
|
| 489 |
-
"logps/rejected": -
|
| 490 |
-
"loss": 1.
|
| 491 |
-
"nll_loss":
|
| 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": 5.
|
| 501 |
-
"learning_rate":
|
| 502 |
-
"log_odds_chosen":
|
| 503 |
-
"log_odds_ratio": -0.
|
| 504 |
-
"logits/chosen": -
|
| 505 |
-
"logits/rejected": 15.
|
| 506 |
-
"logps/chosen": -1.
|
| 507 |
-
"logps/rejected": -
|
| 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":
|
| 521 |
-
"log_odds_ratio": -0.
|
| 522 |
-
"logits/chosen": -
|
| 523 |
-
"logits/rejected": 15.
|
| 524 |
-
"logps/chosen": -
|
| 525 |
-
"logps/rejected": -
|
| 526 |
-
"loss": 1.
|
| 527 |
-
"nll_loss": 1.
|
| 528 |
"rewards/accuracies": 0.981249988079071,
|
| 529 |
-
"rewards/chosen": -0.
|
| 530 |
-
"rewards/margins": 0.
|
| 531 |
-
"rewards/rejected": -0.
|
| 532 |
"step": 290
|
| 533 |
},
|
| 534 |
{
|
| 535 |
"epoch": 0.3870967741935484,
|
| 536 |
-
"grad_norm":
|
| 537 |
-
"learning_rate":
|
| 538 |
-
"log_odds_chosen":
|
| 539 |
-
"log_odds_ratio": -0.
|
| 540 |
-
"logits/chosen": -
|
| 541 |
-
"logits/rejected": 15.
|
| 542 |
-
"logps/chosen": -1.
|
| 543 |
-
"logps/rejected": -
|
| 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": 5.
|
| 555 |
-
"learning_rate":
|
| 556 |
-
"log_odds_chosen": 2.
|
| 557 |
-
"log_odds_ratio": -0.
|
| 558 |
-
"logits/chosen": -
|
| 559 |
-
"logits/rejected":
|
| 560 |
-
"logps/chosen": -
|
| 561 |
-
"logps/rejected": -
|
| 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":
|
| 573 |
-
"learning_rate":
|
| 574 |
-
"log_odds_chosen":
|
| 575 |
-
"log_odds_ratio": -0.
|
| 576 |
-
"logits/chosen": -
|
| 577 |
-
"logits/rejected": 15.
|
| 578 |
-
"logps/chosen": -
|
| 579 |
-
"logps/rejected": -
|
| 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": 6.
|
| 591 |
-
"learning_rate":
|
| 592 |
-
"log_odds_chosen":
|
| 593 |
-
"log_odds_ratio": -0.
|
| 594 |
-
"logits/chosen": -
|
| 595 |
-
"logits/rejected": 15.
|
| 596 |
-
"logps/chosen": -1.
|
| 597 |
-
"logps/rejected": -
|
| 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":
|
| 611 |
-
"log_odds_ratio": -0.
|
| 612 |
-
"logits/chosen": -1.
|
| 613 |
-
"logits/rejected": 15.
|
| 614 |
-
"logps/chosen": -1.
|
| 615 |
-
"logps/rejected": -
|
| 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":
|
| 627 |
-
"learning_rate":
|
| 628 |
-
"log_odds_chosen":
|
| 629 |
-
"log_odds_ratio": -0.
|
| 630 |
-
"logits/chosen": -
|
| 631 |
-
"logits/rejected": 15.
|
| 632 |
-
"logps/chosen": -
|
| 633 |
-
"logps/rejected": -
|
| 634 |
-
"loss": 1.
|
| 635 |
-
"nll_loss": 1.
|
| 636 |
"rewards/accuracies": 0.9624999761581421,
|
| 637 |
-
"rewards/chosen": -0.
|
| 638 |
-
"rewards/margins": 0.
|
| 639 |
-
"rewards/rejected": -0.
|
| 640 |
"step": 350
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 0.4645161290322581,
|
| 644 |
-
"grad_norm": 5.
|
| 645 |
-
"learning_rate":
|
| 646 |
-
"log_odds_chosen":
|
| 647 |
-
"log_odds_ratio": -0.
|
| 648 |
-
"logits/chosen": -
|
| 649 |
-
"logits/rejected": 15.
|
| 650 |
-
"logps/chosen": -1.
|
| 651 |
-
"logps/rejected": -
|
| 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":
|
| 665 |
-
"log_odds_ratio": -0.
|
| 666 |
-
"logits/chosen": -
|
| 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": 4.
|
| 681 |
-
"learning_rate":
|
| 682 |
-
"log_odds_chosen": 2.
|
| 683 |
-
"log_odds_ratio": -0.
|
| 684 |
-
"logits/chosen": -
|
| 685 |
-
"logits/rejected": 15.
|
| 686 |
-
"logps/chosen": -
|
| 687 |
-
"logps/rejected": -2.
|
| 688 |
-
"loss":
|
| 689 |
-
"nll_loss":
|
| 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": 5.
|
| 699 |
-
"learning_rate":
|
| 700 |
-
"log_odds_chosen": 2.
|
| 701 |
-
"log_odds_ratio": -0.
|
| 702 |
-
"logits/chosen": -
|
| 703 |
-
"logits/rejected": 15.
|
| 704 |
-
"logps/chosen": -
|
| 705 |
-
"logps/rejected": -2.
|
| 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": 4.
|
| 717 |
-
"learning_rate":
|
| 718 |
-
"log_odds_chosen": 2.
|
| 719 |
-
"log_odds_ratio": -0.
|
| 720 |
-
"logits/chosen": -2.
|
| 721 |
-
"logits/rejected": 15.
|
| 722 |
-
"logps/chosen": -
|
| 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": 5.
|
| 735 |
-
"learning_rate":
|
| 736 |
-
"log_odds_chosen":
|
| 737 |
-
"log_odds_ratio": -0.
|
| 738 |
-
"logits/chosen": -2.
|
| 739 |
-
"logits/rejected": 15.
|
| 740 |
-
"logps/chosen": -
|
| 741 |
-
"logps/rejected": -
|
| 742 |
-
"loss":
|
| 743 |
-
"nll_loss":
|
| 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": 5.
|
| 753 |
-
"learning_rate":
|
| 754 |
-
"log_odds_chosen":
|
| 755 |
-
"log_odds_ratio": -0.
|
| 756 |
-
"logits/chosen": -
|
| 757 |
-
"logits/rejected": 15.
|
| 758 |
-
"logps/chosen": -1.
|
| 759 |
-
"logps/rejected": -
|
| 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":
|
| 773 |
-
"log_odds_ratio": -0.
|
| 774 |
-
"logits/chosen": -
|
| 775 |
-
"logits/rejected": 16.
|
| 776 |
-
"logps/chosen": -1.
|
| 777 |
-
"logps/rejected": -
|
| 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": 6.
|
| 789 |
-
"learning_rate":
|
| 790 |
-
"log_odds_chosen":
|
| 791 |
-
"log_odds_ratio": -0.
|
| 792 |
-
"logits/chosen": -
|
| 793 |
-
"logits/rejected": 15.
|
| 794 |
-
"logps/chosen": -1.
|
| 795 |
-
"logps/rejected": -3.
|
| 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": 5.
|
| 807 |
-
"learning_rate":
|
| 808 |
-
"log_odds_chosen":
|
| 809 |
-
"log_odds_ratio": -0.
|
| 810 |
-
"logits/chosen": -
|
| 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": 5.
|
| 825 |
-
"learning_rate":
|
| 826 |
-
"log_odds_chosen":
|
| 827 |
-
"log_odds_ratio": -0.
|
| 828 |
-
"logits/chosen": -
|
| 829 |
-
"logits/rejected": 15.
|
| 830 |
-
"logps/chosen": -
|
| 831 |
-
"logps/rejected": -
|
| 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": 6.
|
| 843 |
-
"learning_rate":
|
| 844 |
-
"log_odds_chosen":
|
| 845 |
-
"log_odds_ratio": -0.
|
| 846 |
-
"logits/chosen": -
|
| 847 |
-
"logits/rejected":
|
| 848 |
-
"logps/chosen": -1.
|
| 849 |
-
"logps/rejected": -
|
| 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": 5.
|
| 861 |
-
"learning_rate":
|
| 862 |
-
"log_odds_chosen":
|
| 863 |
-
"log_odds_ratio": -0.
|
| 864 |
-
"logits/chosen": -
|
| 865 |
-
"logits/rejected": 15.
|
| 866 |
-
"logps/chosen": -0.
|
| 867 |
-
"logps/rejected": -
|
| 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": 5.
|
| 879 |
-
"learning_rate":
|
| 880 |
-
"log_odds_chosen":
|
| 881 |
-
"log_odds_ratio": -0.
|
| 882 |
-
"logits/chosen": -
|
| 883 |
-
"logits/rejected":
|
| 884 |
-
"logps/chosen": -1.
|
| 885 |
-
"logps/rejected": -
|
| 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":
|
| 899 |
-
"log_odds_ratio": -0.
|
| 900 |
-
"logits/chosen": -
|
| 901 |
-
"logits/rejected": 15.
|
| 902 |
-
"logps/chosen": -1.
|
| 903 |
-
"logps/rejected": -
|
| 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": 5.
|
| 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": -
|
| 922 |
-
"loss":
|
| 923 |
-
"nll_loss":
|
| 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": -2.
|
| 937 |
-
"logits/rejected": 15.
|
| 938 |
-
"logps/chosen": -
|
| 939 |
-
"logps/rejected": -
|
| 940 |
-
"loss":
|
| 941 |
-
"nll_loss":
|
| 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": 5.
|
| 951 |
-
"learning_rate":
|
| 952 |
-
"log_odds_chosen":
|
| 953 |
-
"log_odds_ratio": -0.
|
| 954 |
-
"logits/chosen": -
|
| 955 |
-
"logits/rejected": 15.
|
| 956 |
-
"logps/chosen": -
|
| 957 |
-
"logps/rejected": -
|
| 958 |
-
"loss":
|
| 959 |
-
"nll_loss":
|
| 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":
|
| 969 |
-
"learning_rate":
|
| 970 |
-
"log_odds_chosen":
|
| 971 |
-
"log_odds_ratio": -0.
|
| 972 |
-
"logits/chosen": -
|
| 973 |
-
"logits/rejected":
|
| 974 |
-
"logps/chosen": -1.
|
| 975 |
-
"logps/rejected": -
|
| 976 |
-
"loss":
|
| 977 |
-
"nll_loss":
|
| 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": 5.
|
| 987 |
-
"learning_rate":
|
| 988 |
-
"log_odds_chosen":
|
| 989 |
-
"log_odds_ratio": -0.
|
| 990 |
-
"logits/chosen": -
|
| 991 |
-
"logits/rejected": 15.
|
| 992 |
-
"logps/chosen": -
|
| 993 |
-
"logps/rejected": -
|
| 994 |
-
"loss": 1.
|
| 995 |
-
"nll_loss": 1.
|
| 996 |
-
"rewards/accuracies": 0.
|
| 997 |
-
"rewards/chosen": -0.
|
| 998 |
-
"rewards/margins": 0.
|
| 999 |
-
"rewards/rejected": -0.
|
| 1000 |
"step": 550
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 0.7225806451612903,
|
| 1004 |
-
"grad_norm": 5.
|
| 1005 |
-
"learning_rate":
|
| 1006 |
-
"log_odds_chosen":
|
| 1007 |
-
"log_odds_ratio": -0.
|
| 1008 |
-
"logits/chosen": -
|
| 1009 |
-
"logits/rejected": 15.
|
| 1010 |
-
"logps/chosen": -
|
| 1011 |
-
"logps/rejected": -
|
| 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": -
|
| 1027 |
-
"logits/rejected": 15.
|
| 1028 |
-
"logps/chosen": -
|
| 1029 |
-
"logps/rejected": -
|
| 1030 |
-
"loss":
|
| 1031 |
-
"nll_loss":
|
| 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":
|
| 1041 |
-
"learning_rate":
|
| 1042 |
-
"log_odds_chosen":
|
| 1043 |
-
"log_odds_ratio": -0.
|
| 1044 |
-
"logits/chosen": -
|
| 1045 |
-
"logits/rejected":
|
| 1046 |
-
"logps/chosen": -0.
|
| 1047 |
-
"logps/rejected": -
|
| 1048 |
-
"loss":
|
| 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": 5.
|
| 1059 |
-
"learning_rate":
|
| 1060 |
-
"log_odds_chosen":
|
| 1061 |
-
"log_odds_ratio": -0.
|
| 1062 |
-
"logits/chosen": -2.
|
| 1063 |
-
"logits/rejected": 15.
|
| 1064 |
-
"logps/chosen": -0.
|
| 1065 |
-
"logps/rejected": -
|
| 1066 |
-
"loss":
|
| 1067 |
-
"nll_loss":
|
| 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": 5.
|
| 1077 |
-
"learning_rate":
|
| 1078 |
-
"log_odds_chosen":
|
| 1079 |
-
"log_odds_ratio": -0.
|
| 1080 |
-
"logits/chosen": -
|
| 1081 |
-
"logits/rejected": 15.
|
| 1082 |
-
"logps/chosen": -
|
| 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": -
|
| 1099 |
-
"logits/rejected": 15.
|
| 1100 |
-
"logps/chosen": -
|
| 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":
|
| 1113 |
-
"learning_rate":
|
| 1114 |
-
"log_odds_chosen":
|
| 1115 |
-
"log_odds_ratio": -0.
|
| 1116 |
-
"logits/chosen": -
|
| 1117 |
-
"logits/rejected": 15.
|
| 1118 |
-
"logps/chosen": -
|
| 1119 |
-
"logps/rejected": -
|
| 1120 |
-
"loss":
|
| 1121 |
-
"nll_loss":
|
| 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": -
|
| 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":
|
| 1149 |
-
"learning_rate":
|
| 1150 |
-
"log_odds_chosen":
|
| 1151 |
-
"log_odds_ratio": -0.
|
| 1152 |
-
"logits/chosen": -
|
| 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": 5.
|
| 1167 |
-
"learning_rate":
|
| 1168 |
-
"log_odds_chosen":
|
| 1169 |
-
"log_odds_ratio": -0.
|
| 1170 |
-
"logits/chosen": -
|
| 1171 |
-
"logits/rejected": 16.
|
| 1172 |
-
"logps/chosen": -1.
|
| 1173 |
-
"logps/rejected": -
|
| 1174 |
-
"loss": 1.
|
| 1175 |
-
"nll_loss": 1.
|
| 1176 |
-
"rewards/accuracies": 0.
|
| 1177 |
-
"rewards/chosen": -0.
|
| 1178 |
-
"rewards/margins": 0.
|
| 1179 |
-
"rewards/rejected": -0.
|
| 1180 |
"step": 650
|
| 1181 |
},
|
| 1182 |
{
|
| 1183 |
"epoch": 0.8516129032258064,
|
| 1184 |
-
"grad_norm": 5.
|
| 1185 |
-
"learning_rate":
|
| 1186 |
-
"log_odds_chosen":
|
| 1187 |
-
"log_odds_ratio": -0.
|
| 1188 |
-
"logits/chosen": -
|
| 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":
|
| 1203 |
-
"learning_rate":
|
| 1204 |
-
"log_odds_chosen":
|
| 1205 |
-
"log_odds_ratio": -0.
|
| 1206 |
-
"logits/chosen": -
|
| 1207 |
-
"logits/rejected": 15.
|
| 1208 |
-
"logps/chosen": -
|
| 1209 |
-
"logps/rejected": -
|
| 1210 |
-
"loss": 1.
|
| 1211 |
-
"nll_loss":
|
| 1212 |
"rewards/accuracies": 0.987500011920929,
|
| 1213 |
-
"rewards/chosen": -0.
|
| 1214 |
-
"rewards/margins": 0.
|
| 1215 |
-
"rewards/rejected": -0.
|
| 1216 |
"step": 670
|
| 1217 |
},
|
| 1218 |
{
|
| 1219 |
"epoch": 0.8774193548387097,
|
| 1220 |
-
"grad_norm":
|
| 1221 |
-
"learning_rate":
|
| 1222 |
-
"log_odds_chosen":
|
| 1223 |
-
"log_odds_ratio": -0.
|
| 1224 |
-
"logits/chosen": -
|
| 1225 |
-
"logits/rejected": 15.
|
| 1226 |
-
"logps/chosen": -1.
|
| 1227 |
-
"logps/rejected": -
|
| 1228 |
-
"loss": 1.
|
| 1229 |
-
"nll_loss": 1.
|
| 1230 |
"rewards/accuracies": 0.987500011920929,
|
| 1231 |
-
"rewards/chosen": -0.
|
| 1232 |
-
"rewards/margins": 0.
|
| 1233 |
-
"rewards/rejected": -0.
|
| 1234 |
"step": 680
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8903225806451613,
|
| 1238 |
-
"grad_norm": 5.
|
| 1239 |
-
"learning_rate":
|
| 1240 |
-
"log_odds_chosen":
|
| 1241 |
-
"log_odds_ratio": -0.
|
| 1242 |
-
"logits/chosen": -
|
| 1243 |
-
"logits/rejected": 15.
|
| 1244 |
-
"logps/chosen": -
|
| 1245 |
-
"logps/rejected": -
|
| 1246 |
-
"loss":
|
| 1247 |
-
"nll_loss":
|
| 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": 5.
|
| 1257 |
-
"learning_rate":
|
| 1258 |
-
"log_odds_chosen":
|
| 1259 |
-
"log_odds_ratio": -0.
|
| 1260 |
-
"logits/chosen": -
|
| 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": 5.
|
| 1275 |
-
"learning_rate":
|
| 1276 |
-
"log_odds_chosen":
|
| 1277 |
-
"log_odds_ratio": -0.
|
| 1278 |
-
"logits/chosen": -
|
| 1279 |
-
"logits/rejected": 15.
|
| 1280 |
-
"logps/chosen": -
|
| 1281 |
-
"logps/rejected": -
|
| 1282 |
-
"loss":
|
| 1283 |
-
"nll_loss":
|
| 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":
|
| 1293 |
-
"learning_rate":
|
| 1294 |
-
"log_odds_chosen":
|
| 1295 |
-
"log_odds_ratio": -0.
|
| 1296 |
-
"logits/chosen": -
|
| 1297 |
-
"logits/rejected": 15.
|
| 1298 |
-
"logps/chosen": -
|
| 1299 |
-
"logps/rejected": -
|
| 1300 |
-
"loss":
|
| 1301 |
-
"nll_loss":
|
| 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":
|
| 1311 |
-
"learning_rate":
|
| 1312 |
-
"log_odds_chosen":
|
| 1313 |
-
"log_odds_ratio": -0.
|
| 1314 |
-
"logits/chosen": -
|
| 1315 |
-
"logits/rejected": 15.
|
| 1316 |
-
"logps/chosen": -
|
| 1317 |
-
"logps/rejected": -
|
| 1318 |
-
"loss":
|
| 1319 |
-
"nll_loss":
|
| 1320 |
-
"rewards/accuracies": 0.
|
| 1321 |
-
"rewards/chosen": -0.
|
| 1322 |
-
"rewards/margins": 0.
|
| 1323 |
-
"rewards/rejected": -0.
|
| 1324 |
"step": 730
|
| 1325 |
},
|
| 1326 |
{
|
| 1327 |
"epoch": 0.9548387096774194,
|
| 1328 |
-
"grad_norm": 5.
|
| 1329 |
-
"learning_rate":
|
| 1330 |
-
"log_odds_chosen":
|
| 1331 |
-
"log_odds_ratio": -0.
|
| 1332 |
-
"logits/chosen": -
|
| 1333 |
-
"logits/rejected":
|
| 1334 |
-
"logps/chosen": -
|
| 1335 |
-
"logps/rejected": -
|
| 1336 |
-
"loss":
|
| 1337 |
-
"nll_loss":
|
| 1338 |
-
"rewards/accuracies": 0
|
| 1339 |
-
"rewards/chosen": -0.
|
| 1340 |
-
"rewards/margins": 0.
|
| 1341 |
-
"rewards/rejected": -0.
|
| 1342 |
"step": 740
|
| 1343 |
},
|
| 1344 |
{
|
| 1345 |
"epoch": 0.967741935483871,
|
| 1346 |
-
"grad_norm": 5.
|
| 1347 |
-
"learning_rate":
|
| 1348 |
-
"log_odds_chosen":
|
| 1349 |
-
"log_odds_ratio": -0.
|
| 1350 |
-
"logits/chosen": -
|
| 1351 |
-
"logits/rejected": 15.
|
| 1352 |
-
"logps/chosen": -
|
| 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": 5.
|
| 1365 |
-
"learning_rate":
|
| 1366 |
-
"log_odds_chosen":
|
| 1367 |
-
"log_odds_ratio": -0.
|
| 1368 |
-
"logits/chosen": -
|
| 1369 |
-
"logits/rejected": 15.
|
| 1370 |
-
"logps/chosen": -
|
| 1371 |
-
"logps/rejected": -
|
| 1372 |
-
"loss":
|
| 1373 |
-
"nll_loss":
|
| 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": 6.
|
| 1383 |
-
"learning_rate":
|
| 1384 |
-
"log_odds_chosen":
|
| 1385 |
-
"log_odds_ratio": -0.
|
| 1386 |
-
"logits/chosen": -
|
| 1387 |
-
"logits/rejected":
|
| 1388 |
-
"logps/chosen": -
|
| 1389 |
-
"logps/rejected": -
|
| 1390 |
-
"loss":
|
| 1391 |
-
"nll_loss":
|
| 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":
|
| 1405 |
-
"train_steps_per_second": 0.
|
| 1406 |
}
|
| 1407 |
],
|
| 1408 |
"logging_steps": 10,
|
|
|
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 0.012903225806451613,
|
| 14 |
+
"grad_norm": 24.415898031026043,
|
| 15 |
+
"learning_rate": 4.5000000000000003e-07,
|
| 16 |
+
"log_odds_chosen": 1.075781226158142,
|
| 17 |
+
"log_odds_ratio": -0.3363281190395355,
|
| 18 |
+
"logits/chosen": -1.048828125,
|
| 19 |
+
"logits/rejected": 15.6875,
|
| 20 |
+
"logps/chosen": -1.6554687023162842,
|
| 21 |
+
"logps/rejected": -2.598437547683716,
|
| 22 |
+
"loss": 2.0046,
|
| 23 |
+
"nll_loss": 1.986718773841858,
|
| 24 |
"rewards/accuracies": 0.949999988079071,
|
| 25 |
+
"rewards/chosen": -0.082763671875,
|
| 26 |
+
"rewards/margins": 0.04719238355755806,
|
| 27 |
+
"rewards/rejected": -0.12983398139476776,
|
| 28 |
"step": 10
|
| 29 |
},
|
| 30 |
{
|
| 31 |
"epoch": 0.025806451612903226,
|
| 32 |
+
"grad_norm": 13.677769862646576,
|
| 33 |
+
"learning_rate": 9.500000000000001e-07,
|
| 34 |
+
"log_odds_chosen": 1.037109375,
|
| 35 |
+
"log_odds_ratio": -0.34501951932907104,
|
| 36 |
+
"logits/chosen": -0.6742187738418579,
|
| 37 |
+
"logits/rejected": 16.96875,
|
| 38 |
+
"logps/chosen": -1.6023437976837158,
|
| 39 |
+
"logps/rejected": -2.5,
|
| 40 |
+
"loss": 1.8988,
|
| 41 |
+
"nll_loss": 1.8820312023162842,
|
| 42 |
"rewards/accuracies": 0.9750000238418579,
|
| 43 |
+
"rewards/chosen": -0.08017577975988388,
|
| 44 |
+
"rewards/margins": 0.044921875,
|
| 45 |
+
"rewards/rejected": -0.12504883110523224,
|
| 46 |
"step": 20
|
| 47 |
},
|
| 48 |
{
|
| 49 |
"epoch": 0.03870967741935484,
|
| 50 |
+
"grad_norm": 10.696630985039008,
|
| 51 |
+
"learning_rate": 1.45e-06,
|
| 52 |
+
"log_odds_chosen": 0.707226574420929,
|
| 53 |
+
"log_odds_ratio": -0.4332031309604645,
|
| 54 |
+
"logits/chosen": -0.3716796934604645,
|
| 55 |
+
"logits/rejected": 16.725000381469727,
|
| 56 |
+
"logps/chosen": -1.485937476158142,
|
| 57 |
+
"logps/rejected": -2.063281297683716,
|
| 58 |
+
"loss": 1.7568,
|
| 59 |
+
"nll_loss": 1.7351562976837158,
|
| 60 |
+
"rewards/accuracies": 0.90625,
|
| 61 |
+
"rewards/chosen": -0.07431640475988388,
|
| 62 |
+
"rewards/margins": 0.02889404259622097,
|
| 63 |
+
"rewards/rejected": -0.10317382961511612,
|
| 64 |
"step": 30
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"epoch": 0.05161290322580645,
|
| 68 |
+
"grad_norm": 6.801175564639496,
|
| 69 |
+
"learning_rate": 1.9500000000000004e-06,
|
| 70 |
+
"log_odds_chosen": 0.49785155057907104,
|
| 71 |
+
"log_odds_ratio": -0.498046875,
|
| 72 |
+
"logits/chosen": 0.8692382574081421,
|
| 73 |
+
"logits/rejected": 17.21875,
|
| 74 |
+
"logps/chosen": -1.31640625,
|
| 75 |
+
"logps/rejected": -1.7062499523162842,
|
| 76 |
+
"loss": 1.5594,
|
| 77 |
+
"nll_loss": 1.533593773841858,
|
| 78 |
+
"rewards/accuracies": 0.862500011920929,
|
| 79 |
+
"rewards/chosen": -0.06594238430261612,
|
| 80 |
+
"rewards/margins": 0.019378662109375,
|
| 81 |
+
"rewards/rejected": -0.08525390923023224,
|
| 82 |
"step": 40
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"epoch": 0.06451612903225806,
|
| 86 |
+
"grad_norm": 5.069050171950685,
|
| 87 |
+
"learning_rate": 2.4500000000000003e-06,
|
| 88 |
+
"log_odds_chosen": 0.35029298067092896,
|
| 89 |
+
"log_odds_ratio": -0.5658203363418579,
|
| 90 |
+
"logits/chosen": 1.32763671875,
|
| 91 |
+
"logits/rejected": 17.381250381469727,
|
| 92 |
+
"logps/chosen": -1.22265625,
|
| 93 |
+
"logps/rejected": -1.489843726158142,
|
| 94 |
+
"loss": 1.3957,
|
| 95 |
+
"nll_loss": 1.3671875,
|
| 96 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 97 |
+
"rewards/chosen": -0.06108398362994194,
|
| 98 |
+
"rewards/margins": 0.013356017880141735,
|
| 99 |
+
"rewards/rejected": -0.07441405951976776,
|
| 100 |
"step": 50
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 0.07741935483870968,
|
| 104 |
+
"grad_norm": 5.83545698860985,
|
| 105 |
+
"learning_rate": 2.95e-06,
|
| 106 |
+
"log_odds_chosen": 0.19368895888328552,
|
| 107 |
+
"log_odds_ratio": -0.637890636920929,
|
| 108 |
+
"logits/chosen": 0.9898437261581421,
|
| 109 |
+
"logits/rejected": 16.862499237060547,
|
| 110 |
+
"logps/chosen": -1.1828124523162842,
|
| 111 |
+
"logps/rejected": -1.33203125,
|
| 112 |
+
"loss": 1.3299,
|
| 113 |
+
"nll_loss": 1.296875,
|
| 114 |
+
"rewards/accuracies": 0.625,
|
| 115 |
+
"rewards/chosen": -0.05915527418255806,
|
| 116 |
+
"rewards/margins": 0.007449341006577015,
|
| 117 |
+
"rewards/rejected": -0.06660155951976776,
|
| 118 |
"step": 60
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 0.09032258064516129,
|
| 122 |
+
"grad_norm": 5.270396473856655,
|
| 123 |
+
"learning_rate": 3.45e-06,
|
| 124 |
+
"log_odds_chosen": 0.28056639432907104,
|
| 125 |
+
"log_odds_ratio": -0.590624988079071,
|
| 126 |
+
"logits/chosen": -0.01455078087747097,
|
| 127 |
+
"logits/rejected": 16.762500762939453,
|
| 128 |
+
"logps/chosen": -1.068750023841858,
|
| 129 |
+
"logps/rejected": -1.275781273841858,
|
| 130 |
+
"loss": 1.1896,
|
| 131 |
+
"nll_loss": 1.1593749523162842,
|
| 132 |
+
"rewards/accuracies": 0.699999988079071,
|
| 133 |
+
"rewards/chosen": -0.053466796875,
|
| 134 |
+
"rewards/margins": 0.010417175479233265,
|
| 135 |
+
"rewards/rejected": -0.06386718899011612,
|
| 136 |
"step": 70
|
| 137 |
},
|
| 138 |
{
|
| 139 |
"epoch": 0.1032258064516129,
|
| 140 |
+
"grad_norm": 5.648914075324417,
|
| 141 |
+
"learning_rate": 3.95e-06,
|
| 142 |
+
"log_odds_chosen": 0.39033204317092896,
|
| 143 |
+
"log_odds_ratio": -0.55078125,
|
| 144 |
+
"logits/chosen": -0.808667004108429,
|
| 145 |
+
"logits/rejected": 16.25,
|
| 146 |
+
"logps/chosen": -1.0945312976837158,
|
| 147 |
+
"logps/rejected": -1.372656226158142,
|
| 148 |
+
"loss": 1.1625,
|
| 149 |
+
"nll_loss": 1.13671875,
|
| 150 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 151 |
+
"rewards/chosen": -0.0546875,
|
| 152 |
+
"rewards/margins": 0.01395568810403347,
|
| 153 |
+
"rewards/rejected": -0.06857910007238388,
|
| 154 |
"step": 80
|
| 155 |
},
|
| 156 |
{
|
| 157 |
"epoch": 0.11612903225806452,
|
| 158 |
+
"grad_norm": 4.665506563337419,
|
| 159 |
+
"learning_rate": 4.450000000000001e-06,
|
| 160 |
+
"log_odds_chosen": 1.002343773841858,
|
| 161 |
+
"log_odds_ratio": -0.35888671875,
|
| 162 |
+
"logits/chosen": -1.157812476158142,
|
| 163 |
+
"logits/rejected": 15.862500190734863,
|
| 164 |
+
"logps/chosen": -1.06640625,
|
| 165 |
+
"logps/rejected": -1.828125,
|
| 166 |
+
"loss": 1.1173,
|
| 167 |
+
"nll_loss": 1.099609375,
|
| 168 |
+
"rewards/accuracies": 0.9437500238418579,
|
| 169 |
+
"rewards/chosen": -0.05329589918255806,
|
| 170 |
+
"rewards/margins": 0.03801269456744194,
|
| 171 |
+
"rewards/rejected": -0.09135742485523224,
|
| 172 |
"step": 90
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"epoch": 0.12903225806451613,
|
| 176 |
+
"grad_norm": 4.9194277923750285,
|
| 177 |
+
"learning_rate": 4.95e-06,
|
| 178 |
+
"log_odds_chosen": 1.3742187023162842,
|
| 179 |
+
"log_odds_ratio": -0.29130858182907104,
|
| 180 |
+
"logits/chosen": -1.532812476158142,
|
| 181 |
+
"logits/rejected": 15.25,
|
| 182 |
+
"logps/chosen": -1.092187523841858,
|
| 183 |
+
"logps/rejected": -2.2007813453674316,
|
| 184 |
+
"loss": 1.1352,
|
| 185 |
+
"nll_loss": 1.1218750476837158,
|
| 186 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 187 |
+
"rewards/chosen": -0.05461425706744194,
|
| 188 |
+
"rewards/margins": 0.0555419921875,
|
| 189 |
+
"rewards/rejected": -0.110107421875,
|
| 190 |
"step": 100
|
| 191 |
},
|
| 192 |
{
|
| 193 |
"epoch": 0.14193548387096774,
|
| 194 |
+
"grad_norm": 4.9620026922411204,
|
| 195 |
+
"learning_rate": 4.789131426105757e-06,
|
| 196 |
+
"log_odds_chosen": 1.75390625,
|
| 197 |
+
"log_odds_ratio": -0.212890625,
|
| 198 |
+
"logits/chosen": -0.905517578125,
|
| 199 |
+
"logits/rejected": 16.043750762939453,
|
| 200 |
+
"logps/chosen": -1.085546851158142,
|
| 201 |
+
"logps/rejected": -2.515625,
|
| 202 |
+
"loss": 1.1138,
|
| 203 |
+
"nll_loss": 1.103124976158142,
|
| 204 |
+
"rewards/accuracies": 0.981249988079071,
|
| 205 |
+
"rewards/chosen": -0.05424804612994194,
|
| 206 |
+
"rewards/margins": 0.07148437201976776,
|
| 207 |
+
"rewards/rejected": -0.125732421875,
|
| 208 |
"step": 110
|
| 209 |
},
|
| 210 |
{
|
| 211 |
"epoch": 0.15483870967741936,
|
| 212 |
+
"grad_norm": 4.526815051594843,
|
| 213 |
+
"learning_rate": 4.583492485141056e-06,
|
| 214 |
+
"log_odds_chosen": 1.5597655773162842,
|
| 215 |
+
"log_odds_ratio": -0.2994140684604645,
|
| 216 |
+
"logits/chosen": -1.553466796875,
|
| 217 |
+
"logits/rejected": 16.456249237060547,
|
| 218 |
+
"logps/chosen": -1.057031273841858,
|
| 219 |
+
"logps/rejected": -2.3335938453674316,
|
| 220 |
+
"loss": 1.0869,
|
| 221 |
+
"nll_loss": 1.071874976158142,
|
| 222 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 223 |
+
"rewards/chosen": -0.05280761793255806,
|
| 224 |
+
"rewards/margins": 0.06374511867761612,
|
| 225 |
+
"rewards/rejected": -0.11655273288488388,
|
| 226 |
"step": 120
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"epoch": 0.16774193548387098,
|
| 230 |
+
"grad_norm": 4.9259847932667595,
|
| 231 |
+
"learning_rate": 4.4022545316281195e-06,
|
| 232 |
+
"log_odds_chosen": 1.9718749523162842,
|
| 233 |
+
"log_odds_ratio": -0.19765624403953552,
|
| 234 |
+
"logits/chosen": -0.6851562261581421,
|
| 235 |
+
"logits/rejected": 16.143749237060547,
|
| 236 |
+
"logps/chosen": -1.0867187976837158,
|
| 237 |
+
"logps/rejected": -2.7265625,
|
| 238 |
+
"loss": 1.1121,
|
| 239 |
+
"nll_loss": 1.103124976158142,
|
| 240 |
+
"rewards/accuracies": 0.956250011920929,
|
| 241 |
+
"rewards/chosen": -0.0543212890625,
|
| 242 |
+
"rewards/margins": 0.08188476413488388,
|
| 243 |
+
"rewards/rejected": -0.13618163764476776,
|
| 244 |
"step": 130
|
| 245 |
},
|
| 246 |
{
|
| 247 |
"epoch": 0.18064516129032257,
|
| 248 |
+
"grad_norm": 5.039653421145649,
|
| 249 |
+
"learning_rate": 4.240944648399854e-06,
|
| 250 |
+
"log_odds_chosen": 1.8359375,
|
| 251 |
+
"log_odds_ratio": -0.20595702528953552,
|
| 252 |
+
"logits/chosen": -1.1907227039337158,
|
| 253 |
+
"logits/rejected": 15.668749809265137,
|
| 254 |
+
"logps/chosen": -1.072656273841858,
|
| 255 |
+
"logps/rejected": -2.571093797683716,
|
| 256 |
+
"loss": 1.0928,
|
| 257 |
+
"nll_loss": 1.08203125,
|
| 258 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 259 |
+
"rewards/chosen": -0.05368652194738388,
|
| 260 |
+
"rewards/margins": 0.07490234076976776,
|
| 261 |
+
"rewards/rejected": -0.128662109375,
|
| 262 |
"step": 140
|
| 263 |
},
|
| 264 |
{
|
| 265 |
"epoch": 0.1935483870967742,
|
| 266 |
+
"grad_norm": 4.8956379810361375,
|
| 267 |
+
"learning_rate": 4.096159602595203e-06,
|
| 268 |
+
"log_odds_chosen": 1.8679687976837158,
|
| 269 |
+
"log_odds_ratio": -0.22802734375,
|
| 270 |
+
"logits/chosen": -1.8488280773162842,
|
| 271 |
+
"logits/rejected": 15.550000190734863,
|
| 272 |
+
"logps/chosen": -1.05078125,
|
| 273 |
+
"logps/rejected": -2.5999999046325684,
|
| 274 |
+
"loss": 1.0779,
|
| 275 |
+
"nll_loss": 1.0656249523162842,
|
| 276 |
+
"rewards/accuracies": 0.96875,
|
| 277 |
+
"rewards/chosen": -0.0526123046875,
|
| 278 |
+
"rewards/margins": 0.07744140923023224,
|
| 279 |
+
"rewards/rejected": -0.12998047471046448,
|
| 280 |
"step": 150
|
| 281 |
},
|
| 282 |
{
|
| 283 |
"epoch": 0.2064516129032258,
|
| 284 |
+
"grad_norm": 5.648955197242222,
|
| 285 |
+
"learning_rate": 3.965257928590721e-06,
|
| 286 |
+
"log_odds_chosen": 2.026562452316284,
|
| 287 |
+
"log_odds_ratio": -0.21791991591453552,
|
| 288 |
+
"logits/chosen": -2.2445311546325684,
|
| 289 |
+
"logits/rejected": 14.962499618530273,
|
| 290 |
+
"logps/chosen": -1.0164062976837158,
|
| 291 |
+
"logps/rejected": -2.700000047683716,
|
| 292 |
+
"loss": 1.0501,
|
| 293 |
+
"nll_loss": 1.0402343273162842,
|
| 294 |
+
"rewards/accuracies": 0.96875,
|
| 295 |
+
"rewards/chosen": -0.05083007737994194,
|
| 296 |
+
"rewards/margins": 0.084228515625,
|
| 297 |
+
"rewards/rejected": -0.135009765625,
|
| 298 |
"step": 160
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 0.21935483870967742,
|
| 302 |
+
"grad_norm": 4.50921729284181,
|
| 303 |
+
"learning_rate": 3.846153846153846e-06,
|
| 304 |
+
"log_odds_chosen": 2.225781202316284,
|
| 305 |
+
"log_odds_ratio": -0.15439453721046448,
|
| 306 |
+
"logits/chosen": -2.3375000953674316,
|
| 307 |
+
"logits/rejected": 16.0625,
|
| 308 |
+
"logps/chosen": -0.977734386920929,
|
| 309 |
+
"logps/rejected": -2.809375047683716,
|
| 310 |
+
"loss": 1.0144,
|
| 311 |
+
"nll_loss": 1.006250023841858,
|
| 312 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 313 |
+
"rewards/chosen": -0.04885254055261612,
|
| 314 |
+
"rewards/margins": 0.09160156548023224,
|
| 315 |
+
"rewards/rejected": -0.14057616889476776,
|
| 316 |
"step": 170
|
| 317 |
},
|
| 318 |
{
|
| 319 |
"epoch": 0.23225806451612904,
|
| 320 |
+
"grad_norm": 5.153444368200689,
|
| 321 |
+
"learning_rate": 3.7371754637596797e-06,
|
| 322 |
+
"log_odds_chosen": 2.354687452316284,
|
| 323 |
+
"log_odds_ratio": -0.14643554389476776,
|
| 324 |
+
"logits/chosen": -1.9169921875,
|
| 325 |
+
"logits/rejected": 15.456250190734863,
|
| 326 |
+
"logps/chosen": -1.01953125,
|
| 327 |
+
"logps/rejected": -2.981250047683716,
|
| 328 |
+
"loss": 1.0307,
|
| 329 |
+
"nll_loss": 1.0234375,
|
| 330 |
+
"rewards/accuracies": 0.981249988079071,
|
| 331 |
+
"rewards/chosen": -0.05097656324505806,
|
| 332 |
+
"rewards/margins": 0.09814453125,
|
| 333 |
+
"rewards/rejected": -0.14912109076976776,
|
| 334 |
"step": 180
|
| 335 |
},
|
| 336 |
{
|
| 337 |
"epoch": 0.24516129032258063,
|
| 338 |
+
"grad_norm": 5.596223310698167,
|
| 339 |
+
"learning_rate": 3.63696483726654e-06,
|
| 340 |
+
"log_odds_chosen": 2.479687452316284,
|
| 341 |
+
"log_odds_ratio": -0.14628906548023224,
|
| 342 |
+
"logits/chosen": -1.4436156749725342,
|
| 343 |
+
"logits/rejected": 15.90625,
|
| 344 |
+
"logps/chosen": -1.0,
|
| 345 |
+
"logps/rejected": -3.0859375,
|
| 346 |
+
"loss": 1.0302,
|
| 347 |
+
"nll_loss": 1.023046851158142,
|
| 348 |
+
"rewards/accuracies": 0.96875,
|
| 349 |
+
"rewards/chosen": -0.04997558519244194,
|
| 350 |
+
"rewards/margins": 0.10419921576976776,
|
| 351 |
+
"rewards/rejected": -0.15419921278953552,
|
| 352 |
"step": 190
|
| 353 |
},
|
| 354 |
{
|
| 355 |
"epoch": 0.25806451612903225,
|
| 356 |
+
"grad_norm": 4.6505717562701765,
|
| 357 |
+
"learning_rate": 3.5444060250416796e-06,
|
| 358 |
+
"log_odds_chosen": 2.6781249046325684,
|
| 359 |
+
"log_odds_ratio": -0.11782226711511612,
|
| 360 |
+
"logits/chosen": -1.956640601158142,
|
| 361 |
+
"logits/rejected": 16.037500381469727,
|
| 362 |
+
"logps/chosen": -1.033203125,
|
| 363 |
+
"logps/rejected": -3.3046875,
|
| 364 |
+
"loss": 1.0531,
|
| 365 |
+
"nll_loss": 1.0480468273162842,
|
| 366 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 367 |
+
"rewards/chosen": -0.0516357421875,
|
| 368 |
+
"rewards/margins": 0.11372070014476776,
|
| 369 |
+
"rewards/rejected": -0.16542968153953552,
|
| 370 |
"step": 200
|
| 371 |
},
|
| 372 |
{
|
| 373 |
"epoch": 0.2709677419354839,
|
| 374 |
+
"grad_norm": 5.056481299672191,
|
| 375 |
+
"learning_rate": 3.458572319330373e-06,
|
| 376 |
+
"log_odds_chosen": 2.6109375953674316,
|
| 377 |
+
"log_odds_ratio": -0.16538086533546448,
|
| 378 |
+
"logits/chosen": -2.354687452316284,
|
| 379 |
+
"logits/rejected": 15.25,
|
| 380 |
+
"logps/chosen": -1.021875023841858,
|
| 381 |
+
"logps/rejected": -3.239062547683716,
|
| 382 |
+
"loss": 1.0347,
|
| 383 |
+
"nll_loss": 1.0265624523162842,
|
| 384 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 385 |
+
"rewards/chosen": -0.05104980617761612,
|
| 386 |
+
"rewards/margins": 0.11093749850988388,
|
| 387 |
+
"rewards/rejected": -0.162109375,
|
| 388 |
"step": 210
|
| 389 |
},
|
| 390 |
{
|
| 391 |
"epoch": 0.2838709677419355,
|
| 392 |
+
"grad_norm": 4.888314219389848,
|
| 393 |
+
"learning_rate": 3.3786868919974304e-06,
|
| 394 |
+
"log_odds_chosen": 2.6578125953674316,
|
| 395 |
+
"log_odds_ratio": -0.17612305283546448,
|
| 396 |
+
"logits/chosen": -1.467187523841858,
|
| 397 |
+
"logits/rejected": 15.787500381469727,
|
| 398 |
+
"logps/chosen": -1.0832030773162842,
|
| 399 |
+
"logps/rejected": -3.378124952316284,
|
| 400 |
+
"loss": 1.095,
|
| 401 |
+
"nll_loss": 1.0867187976837158,
|
| 402 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 403 |
+
"rewards/chosen": -0.05415039137005806,
|
| 404 |
+
"rewards/margins": 0.11494140326976776,
|
| 405 |
+
"rewards/rejected": -0.16904297471046448,
|
| 406 |
"step": 220
|
| 407 |
},
|
| 408 |
{
|
| 409 |
"epoch": 0.2967741935483871,
|
| 410 |
+
"grad_norm": 4.697252316410647,
|
| 411 |
+
"learning_rate": 3.3040930022754493e-06,
|
| 412 |
+
"log_odds_chosen": 2.167187452316284,
|
| 413 |
+
"log_odds_ratio": -0.19511719048023224,
|
| 414 |
+
"logits/chosen": -1.1691405773162842,
|
| 415 |
+
"logits/rejected": 15.512499809265137,
|
| 416 |
+
"logps/chosen": -1.071874976158142,
|
| 417 |
+
"logps/rejected": -2.8968749046325684,
|
| 418 |
+
"loss": 1.0854,
|
| 419 |
+
"nll_loss": 1.0773437023162842,
|
| 420 |
+
"rewards/accuracies": 0.96875,
|
| 421 |
+
"rewards/chosen": -0.05363769456744194,
|
| 422 |
+
"rewards/margins": 0.09116210788488388,
|
| 423 |
+
"rewards/rejected": -0.14492186903953552,
|
| 424 |
"step": 230
|
| 425 |
},
|
| 426 |
{
|
| 427 |
"epoch": 0.3096774193548387,
|
| 428 |
+
"grad_norm": 4.577631356182271,
|
| 429 |
+
"learning_rate": 3.234231136765755e-06,
|
| 430 |
+
"log_odds_chosen": 2.2632813453674316,
|
| 431 |
+
"log_odds_ratio": -0.19497069716453552,
|
| 432 |
+
"logits/chosen": -2.502148389816284,
|
| 433 |
+
"logits/rejected": 14.100000381469727,
|
| 434 |
+
"logps/chosen": -1.017968773841858,
|
| 435 |
+
"logps/rejected": -2.9296875,
|
| 436 |
+
"loss": 1.048,
|
| 437 |
+
"nll_loss": 1.0378906726837158,
|
| 438 |
+
"rewards/accuracies": 0.987500011920929,
|
| 439 |
+
"rewards/chosen": -0.05097656324505806,
|
| 440 |
+
"rewards/margins": 0.09550781548023224,
|
| 441 |
+
"rewards/rejected": -0.14628906548023224,
|
| 442 |
"step": 240
|
| 443 |
},
|
| 444 |
{
|
| 445 |
"epoch": 0.3225806451612903,
|
| 446 |
+
"grad_norm": 4.858199223746265,
|
| 447 |
+
"learning_rate": 3.1686212526223897e-06,
|
| 448 |
+
"log_odds_chosen": 2.8984375,
|
| 449 |
+
"log_odds_ratio": -0.1181640625,
|
| 450 |
+
"logits/chosen": -2.688671827316284,
|
| 451 |
+
"logits/rejected": 15.199999809265137,
|
| 452 |
+
"logps/chosen": -0.9878906011581421,
|
| 453 |
+
"logps/rejected": -3.464062452316284,
|
| 454 |
+
"loss": 1.0084,
|
| 455 |
+
"nll_loss": 1.0031249523162842,
|
| 456 |
"rewards/accuracies": 0.981249988079071,
|
| 457 |
+
"rewards/chosen": -0.0494384765625,
|
| 458 |
+
"rewards/margins": 0.1240234375,
|
| 459 |
+
"rewards/rejected": -0.17333984375,
|
| 460 |
"step": 250
|
| 461 |
},
|
| 462 |
{
|
| 463 |
"epoch": 0.33548387096774196,
|
| 464 |
+
"grad_norm": 4.519869758016946,
|
| 465 |
+
"learning_rate": 3.1068488300060003e-06,
|
| 466 |
+
"log_odds_chosen": 2.4437499046325684,
|
| 467 |
+
"log_odds_ratio": -0.14501953125,
|
| 468 |
+
"logits/chosen": -2.409374952316284,
|
| 469 |
+
"logits/rejected": 15.206250190734863,
|
| 470 |
+
"logps/chosen": -1.005859375,
|
| 471 |
+
"logps/rejected": -3.0453124046325684,
|
| 472 |
+
"loss": 1.03,
|
| 473 |
+
"nll_loss": 1.0222656726837158,
|
| 474 |
+
"rewards/accuracies": 0.987500011920929,
|
| 475 |
+
"rewards/chosen": -0.05021972581744194,
|
| 476 |
+
"rewards/margins": 0.10209961235523224,
|
| 477 |
+
"rewards/rejected": -0.15224608778953552,
|
| 478 |
"step": 260
|
| 479 |
},
|
| 480 |
{
|
| 481 |
"epoch": 0.34838709677419355,
|
| 482 |
+
"grad_norm": 5.239684216445077,
|
| 483 |
+
"learning_rate": 3.0485538042484623e-06,
|
| 484 |
+
"log_odds_chosen": 2.598437547683716,
|
| 485 |
+
"log_odds_ratio": -0.14907225966453552,
|
| 486 |
+
"logits/chosen": -2.1636719703674316,
|
| 487 |
+
"logits/rejected": 15.050000190734863,
|
| 488 |
+
"logps/chosen": -0.981640636920929,
|
| 489 |
+
"logps/rejected": -3.1781249046325684,
|
| 490 |
+
"loss": 1.0049,
|
| 491 |
+
"nll_loss": 0.996874988079071,
|
| 492 |
+
"rewards/accuracies": 0.987500011920929,
|
| 493 |
+
"rewards/chosen": -0.049072265625,
|
| 494 |
+
"rewards/margins": 0.10976562649011612,
|
| 495 |
+
"rewards/rejected": -0.15888671576976776,
|
| 496 |
"step": 270
|
| 497 |
},
|
| 498 |
{
|
| 499 |
"epoch": 0.36129032258064514,
|
| 500 |
+
"grad_norm": 5.276617193754623,
|
| 501 |
+
"learning_rate": 2.9934217004462487e-06,
|
| 502 |
+
"log_odds_chosen": 2.9078125953674316,
|
| 503 |
+
"log_odds_ratio": -0.12812499701976776,
|
| 504 |
+
"logits/chosen": -2.070117235183716,
|
| 505 |
+
"logits/rejected": 15.25,
|
| 506 |
+
"logps/chosen": -1.036718726158142,
|
| 507 |
+
"logps/rejected": -3.535937547683716,
|
| 508 |
+
"loss": 1.0451,
|
| 509 |
+
"nll_loss": 1.0390625,
|
| 510 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 511 |
+
"rewards/chosen": -0.05183105543255806,
|
| 512 |
+
"rewards/margins": 0.12495116889476776,
|
| 513 |
+
"rewards/rejected": -0.1767578125,
|
| 514 |
"step": 280
|
| 515 |
},
|
| 516 |
{
|
| 517 |
"epoch": 0.3741935483870968,
|
| 518 |
+
"grad_norm": 4.901202785480476,
|
| 519 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 520 |
+
"log_odds_chosen": 2.957812547683716,
|
| 521 |
+
"log_odds_ratio": -0.11459960788488388,
|
| 522 |
+
"logits/chosen": -2.172070264816284,
|
| 523 |
+
"logits/rejected": 15.556249618530273,
|
| 524 |
+
"logps/chosen": -0.998046875,
|
| 525 |
+
"logps/rejected": -3.5250000953674316,
|
| 526 |
+
"loss": 1.0108,
|
| 527 |
+
"nll_loss": 1.0050780773162842,
|
| 528 |
"rewards/accuracies": 0.981249988079071,
|
| 529 |
+
"rewards/chosen": -0.04987793043255806,
|
| 530 |
+
"rewards/margins": 0.12641601264476776,
|
| 531 |
+
"rewards/rejected": -0.17607422173023224,
|
| 532 |
"step": 290
|
| 533 |
},
|
| 534 |
{
|
| 535 |
"epoch": 0.3870967741935484,
|
| 536 |
+
"grad_norm": 6.3489235644192314,
|
| 537 |
+
"learning_rate": 2.891574659831201e-06,
|
| 538 |
+
"log_odds_chosen": 3.4140625,
|
| 539 |
+
"log_odds_ratio": -0.12001953274011612,
|
| 540 |
+
"logits/chosen": -2.2777342796325684,
|
| 541 |
+
"logits/rejected": 15.387499809265137,
|
| 542 |
+
"logps/chosen": -1.029687523841858,
|
| 543 |
+
"logps/rejected": -4.029687404632568,
|
| 544 |
+
"loss": 1.0416,
|
| 545 |
+
"nll_loss": 1.037109375,
|
| 546 |
+
"rewards/accuracies": 0.949999988079071,
|
| 547 |
+
"rewards/chosen": -0.05146484449505806,
|
| 548 |
+
"rewards/margins": 0.15004882216453552,
|
| 549 |
+
"rewards/rejected": -0.20146484673023224,
|
| 550 |
"step": 300
|
| 551 |
},
|
| 552 |
{
|
| 553 |
"epoch": 0.4,
|
| 554 |
+
"grad_norm": 5.3538840183503815,
|
| 555 |
+
"learning_rate": 2.844400619942872e-06,
|
| 556 |
+
"log_odds_chosen": 2.953125,
|
| 557 |
+
"log_odds_ratio": -0.14028319716453552,
|
| 558 |
+
"logits/chosen": -2.112109422683716,
|
| 559 |
+
"logits/rejected": 14.90625,
|
| 560 |
+
"logps/chosen": -0.994921863079071,
|
| 561 |
+
"logps/rejected": -3.535937547683716,
|
| 562 |
+
"loss": 1.0155,
|
| 563 |
+
"nll_loss": 1.008203148841858,
|
| 564 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 565 |
+
"rewards/chosen": -0.0496826171875,
|
| 566 |
+
"rewards/margins": 0.12700195610523224,
|
| 567 |
+
"rewards/rejected": -0.17685547471046448,
|
| 568 |
"step": 310
|
| 569 |
},
|
| 570 |
{
|
| 571 |
"epoch": 0.4129032258064516,
|
| 572 |
+
"grad_norm": 5.370323496962624,
|
| 573 |
+
"learning_rate": 2.799462554779272e-06,
|
| 574 |
+
"log_odds_chosen": 3.078125,
|
| 575 |
+
"log_odds_ratio": -0.1260986328125,
|
| 576 |
+
"logits/chosen": -2.033203125,
|
| 577 |
+
"logits/rejected": 15.4375,
|
| 578 |
+
"logps/chosen": -0.99609375,
|
| 579 |
+
"logps/rejected": -3.65625,
|
| 580 |
+
"loss": 1.0071,
|
| 581 |
+
"nll_loss": 1.0007812976837158,
|
| 582 |
+
"rewards/accuracies": 0.96875,
|
| 583 |
+
"rewards/chosen": -0.04982910305261612,
|
| 584 |
+
"rewards/margins": 0.13276366889476776,
|
| 585 |
+
"rewards/rejected": -0.18251952528953552,
|
| 586 |
"step": 320
|
| 587 |
},
|
| 588 |
{
|
| 589 |
"epoch": 0.4258064516129032,
|
| 590 |
+
"grad_norm": 6.23268929266231,
|
| 591 |
+
"learning_rate": 2.7565892320998566e-06,
|
| 592 |
+
"log_odds_chosen": 3.582812547683716,
|
| 593 |
+
"log_odds_ratio": -0.09384765475988388,
|
| 594 |
+
"logits/chosen": -2.2328124046325684,
|
| 595 |
+
"logits/rejected": 15.524999618530273,
|
| 596 |
+
"logps/chosen": -1.01953125,
|
| 597 |
+
"logps/rejected": -4.173437595367432,
|
| 598 |
+
"loss": 1.0513,
|
| 599 |
+
"nll_loss": 1.046484351158142,
|
| 600 |
+
"rewards/accuracies": 0.987500011920929,
|
| 601 |
+
"rewards/chosen": -0.051025390625,
|
| 602 |
+
"rewards/margins": 0.15756836533546448,
|
| 603 |
+
"rewards/rejected": -0.20859375596046448,
|
| 604 |
"step": 330
|
| 605 |
},
|
| 606 |
{
|
| 607 |
"epoch": 0.43870967741935485,
|
| 608 |
+
"grad_norm": 5.462484402984074,
|
| 609 |
+
"learning_rate": 2.7156272329678423e-06,
|
| 610 |
+
"log_odds_chosen": 3.59375,
|
| 611 |
+
"log_odds_ratio": -0.0960693359375,
|
| 612 |
+
"logits/chosen": -1.870507836341858,
|
| 613 |
+
"logits/rejected": 15.387499809265137,
|
| 614 |
+
"logps/chosen": -1.039453148841858,
|
| 615 |
+
"logps/rejected": -4.217187404632568,
|
| 616 |
+
"loss": 1.0323,
|
| 617 |
+
"nll_loss": 1.0281250476837158,
|
| 618 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 619 |
+
"rewards/chosen": -0.05192871019244194,
|
| 620 |
+
"rewards/margins": 0.15878906846046448,
|
| 621 |
+
"rewards/rejected": -0.21064452826976776,
|
| 622 |
"step": 340
|
| 623 |
},
|
| 624 |
{
|
| 625 |
"epoch": 0.45161290322580644,
|
| 626 |
+
"grad_norm": 5.006492348614002,
|
| 627 |
+
"learning_rate": 2.6764386378609465e-06,
|
| 628 |
+
"log_odds_chosen": 3.403125047683716,
|
| 629 |
+
"log_odds_ratio": -0.13632813096046448,
|
| 630 |
+
"logits/chosen": -2.469531297683716,
|
| 631 |
+
"logits/rejected": 15.212499618530273,
|
| 632 |
+
"logps/chosen": -0.9750000238418579,
|
| 633 |
+
"logps/rejected": -3.9609375,
|
| 634 |
+
"loss": 1.0064,
|
| 635 |
+
"nll_loss": 1.000390648841858,
|
| 636 |
"rewards/accuracies": 0.9624999761581421,
|
| 637 |
+
"rewards/chosen": -0.04877929762005806,
|
| 638 |
+
"rewards/margins": 0.149169921875,
|
| 639 |
+
"rewards/rejected": -0.19794921576976776,
|
| 640 |
"step": 350
|
| 641 |
},
|
| 642 |
{
|
| 643 |
"epoch": 0.4645161290322581,
|
| 644 |
+
"grad_norm": 5.708462747115479,
|
| 645 |
+
"learning_rate": 2.638899069846298e-06,
|
| 646 |
+
"log_odds_chosen": 3.543750047683716,
|
| 647 |
+
"log_odds_ratio": -0.11757812649011612,
|
| 648 |
+
"logits/chosen": -2.37890625,
|
| 649 |
+
"logits/rejected": 15.28125,
|
| 650 |
+
"logps/chosen": -1.0078125,
|
| 651 |
+
"logps/rejected": -4.129687309265137,
|
| 652 |
+
"loss": 1.0219,
|
| 653 |
+
"nll_loss": 1.0167968273162842,
|
| 654 |
+
"rewards/accuracies": 0.96875,
|
| 655 |
+
"rewards/chosen": -0.05039062350988388,
|
| 656 |
+
"rewards/margins": 0.15629883110523224,
|
| 657 |
+
"rewards/rejected": -0.20654296875,
|
| 658 |
"step": 360
|
| 659 |
},
|
| 660 |
{
|
| 661 |
"epoch": 0.4774193548387097,
|
| 662 |
+
"grad_norm": 15.819166923351714,
|
| 663 |
+
"learning_rate": 2.6028960314767675e-06,
|
| 664 |
+
"log_odds_chosen": 3.362499952316284,
|
| 665 |
+
"log_odds_ratio": -0.13581542670726776,
|
| 666 |
+
"logits/chosen": -2.4253907203674316,
|
| 667 |
+
"logits/rejected": 15.675000190734863,
|
| 668 |
+
"logps/chosen": -1.008203148841858,
|
| 669 |
+
"logps/rejected": -3.965625047683716,
|
| 670 |
+
"loss": 1.0221,
|
| 671 |
+
"nll_loss": 1.0148437023162842,
|
| 672 |
+
"rewards/accuracies": 0.956250011920929,
|
| 673 |
+
"rewards/chosen": -0.05039062350988388,
|
| 674 |
+
"rewards/margins": 0.14765624701976776,
|
| 675 |
+
"rewards/rejected": -0.1982421875,
|
| 676 |
"step": 370
|
| 677 |
},
|
| 678 |
{
|
| 679 |
"epoch": 0.49032258064516127,
|
| 680 |
+
"grad_norm": 4.605973549061064,
|
| 681 |
+
"learning_rate": 2.568327484687571e-06,
|
| 682 |
+
"log_odds_chosen": 2.30859375,
|
| 683 |
+
"log_odds_ratio": -0.17836913466453552,
|
| 684 |
+
"logits/chosen": -2.288281202316284,
|
| 685 |
+
"logits/rejected": 15.518750190734863,
|
| 686 |
+
"logps/chosen": -0.975390613079071,
|
| 687 |
+
"logps/rejected": -2.8828125,
|
| 688 |
+
"loss": 0.9912,
|
| 689 |
+
"nll_loss": 0.981640636920929,
|
| 690 |
+
"rewards/accuracies": 0.956250011920929,
|
| 691 |
+
"rewards/chosen": -0.0487060546875,
|
| 692 |
+
"rewards/margins": 0.09526367485523224,
|
| 693 |
+
"rewards/rejected": -0.14399413764476776,
|
| 694 |
"step": 380
|
| 695 |
},
|
| 696 |
{
|
| 697 |
"epoch": 0.5032258064516129,
|
| 698 |
+
"grad_norm": 5.353525659101831,
|
| 699 |
+
"learning_rate": 2.535100632816969e-06,
|
| 700 |
+
"log_odds_chosen": 2.34765625,
|
| 701 |
+
"log_odds_ratio": -0.158203125,
|
| 702 |
+
"logits/chosen": -2.428906202316284,
|
| 703 |
+
"logits/rejected": 15.625,
|
| 704 |
+
"logps/chosen": -0.9789062738418579,
|
| 705 |
+
"logps/rejected": -2.9312500953674316,
|
| 706 |
+
"loss": 1.0115,
|
| 707 |
+
"nll_loss": 1.003515601158142,
|
| 708 |
+
"rewards/accuracies": 0.96875,
|
| 709 |
+
"rewards/chosen": -0.04892577975988388,
|
| 710 |
+
"rewards/margins": 0.09760741889476776,
|
| 711 |
+
"rewards/rejected": -0.14653320610523224,
|
| 712 |
"step": 390
|
| 713 |
},
|
| 714 |
{
|
| 715 |
"epoch": 0.5161290322580645,
|
| 716 |
+
"grad_norm": 4.669342637045956,
|
| 717 |
+
"learning_rate": 2.5031308716087944e-06,
|
| 718 |
+
"log_odds_chosen": 2.7562499046325684,
|
| 719 |
+
"log_odds_ratio": -0.12431640923023224,
|
| 720 |
+
"logits/chosen": -2.8046875,
|
| 721 |
+
"logits/rejected": 15.206250190734863,
|
| 722 |
+
"logps/chosen": -0.9898437261581421,
|
| 723 |
+
"logps/rejected": -3.317187547683716,
|
| 724 |
+
"loss": 1.0071,
|
| 725 |
+
"nll_loss": 1.0011718273162842,
|
| 726 |
+
"rewards/accuracies": 0.981249988079071,
|
| 727 |
+
"rewards/chosen": -0.04946289211511612,
|
| 728 |
+
"rewards/margins": 0.11630859225988388,
|
| 729 |
+
"rewards/rejected": -0.16591796278953552,
|
| 730 |
"step": 400
|
| 731 |
},
|
| 732 |
{
|
| 733 |
"epoch": 0.5290322580645161,
|
| 734 |
+
"grad_norm": 5.425862615574437,
|
| 735 |
+
"learning_rate": 2.472340882170744e-06,
|
| 736 |
+
"log_odds_chosen": 3.4671874046325684,
|
| 737 |
+
"log_odds_ratio": -0.08171387016773224,
|
| 738 |
+
"logits/chosen": -2.4320311546325684,
|
| 739 |
+
"logits/rejected": 15.337499618530273,
|
| 740 |
+
"logps/chosen": -0.987500011920929,
|
| 741 |
+
"logps/rejected": -4.004687309265137,
|
| 742 |
+
"loss": 0.9981,
|
| 743 |
+
"nll_loss": 0.993359386920929,
|
| 744 |
+
"rewards/accuracies": 1.0,
|
| 745 |
+
"rewards/chosen": -0.04938964918255806,
|
| 746 |
+
"rewards/margins": 0.15092773735523224,
|
| 747 |
+
"rewards/rejected": -0.20039062201976776,
|
| 748 |
"step": 410
|
| 749 |
},
|
| 750 |
{
|
| 751 |
"epoch": 0.5419354838709678,
|
| 752 |
+
"grad_norm": 5.859653969362379,
|
| 753 |
+
"learning_rate": 2.4426598437301577e-06,
|
| 754 |
+
"log_odds_chosen": 3.825000047683716,
|
| 755 |
+
"log_odds_ratio": -0.10820312798023224,
|
| 756 |
+
"logits/chosen": -2.335156202316284,
|
| 757 |
+
"logits/rejected": 15.056249618530273,
|
| 758 |
+
"logps/chosen": -1.012109398841858,
|
| 759 |
+
"logps/rejected": -4.428124904632568,
|
| 760 |
+
"loss": 1.0235,
|
| 761 |
+
"nll_loss": 1.019140601158142,
|
| 762 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 763 |
+
"rewards/chosen": -0.05058593675494194,
|
| 764 |
+
"rewards/margins": 0.17080077528953552,
|
| 765 |
+
"rewards/rejected": -0.22119140625,
|
| 766 |
"step": 420
|
| 767 |
},
|
| 768 |
{
|
| 769 |
"epoch": 0.5548387096774193,
|
| 770 |
+
"grad_norm": 14.090172688528586,
|
| 771 |
+
"learning_rate": 2.4140227479263382e-06,
|
| 772 |
+
"log_odds_chosen": 3.9375,
|
| 773 |
+
"log_odds_ratio": -0.09453125298023224,
|
| 774 |
+
"logits/chosen": -1.5808594226837158,
|
| 775 |
+
"logits/rejected": 16.018749237060547,
|
| 776 |
+
"logps/chosen": -1.0070312023162842,
|
| 777 |
+
"logps/rejected": -4.506249904632568,
|
| 778 |
+
"loss": 1.0058,
|
| 779 |
+
"nll_loss": 1.002343773841858,
|
| 780 |
+
"rewards/accuracies": 0.987500011920929,
|
| 781 |
+
"rewards/chosen": -0.05031738430261612,
|
| 782 |
+
"rewards/margins": 0.1748046875,
|
| 783 |
+
"rewards/rejected": -0.22519531846046448,
|
| 784 |
"step": 430
|
| 785 |
},
|
| 786 |
{
|
| 787 |
"epoch": 0.567741935483871,
|
| 788 |
+
"grad_norm": 6.569549660660238,
|
| 789 |
+
"learning_rate": 2.3863697995167388e-06,
|
| 790 |
+
"log_odds_chosen": 3.371875047683716,
|
| 791 |
+
"log_odds_ratio": -0.11165771633386612,
|
| 792 |
+
"logits/chosen": -2.5687499046325684,
|
| 793 |
+
"logits/rejected": 15.324999809265137,
|
| 794 |
+
"logps/chosen": -1.009765625,
|
| 795 |
+
"logps/rejected": -3.965625047683716,
|
| 796 |
+
"loss": 1.0316,
|
| 797 |
+
"nll_loss": 1.0261719226837158,
|
| 798 |
+
"rewards/accuracies": 0.987500011920929,
|
| 799 |
+
"rewards/chosen": -0.05051269382238388,
|
| 800 |
+
"rewards/margins": 0.14790038764476776,
|
| 801 |
+
"rewards/rejected": -0.19843749701976776,
|
| 802 |
"step": 440
|
| 803 |
},
|
| 804 |
{
|
| 805 |
"epoch": 0.5806451612903226,
|
| 806 |
+
"grad_norm": 5.065013160531643,
|
| 807 |
+
"learning_rate": 2.359645890915044e-06,
|
| 808 |
+
"log_odds_chosen": 3.518749952316284,
|
| 809 |
+
"log_odds_ratio": -0.11590576171875,
|
| 810 |
+
"logits/chosen": -2.403515577316284,
|
| 811 |
+
"logits/rejected": 15.050000190734863,
|
| 812 |
+
"logps/chosen": -1.0281250476837158,
|
| 813 |
+
"logps/rejected": -4.128125190734863,
|
| 814 |
+
"loss": 1.0486,
|
| 815 |
+
"nll_loss": 1.043359398841858,
|
| 816 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 817 |
+
"rewards/chosen": -0.0513916015625,
|
| 818 |
+
"rewards/margins": 0.15498046576976776,
|
| 819 |
+
"rewards/rejected": -0.20644530653953552,
|
| 820 |
"step": 450
|
| 821 |
},
|
| 822 |
{
|
| 823 |
"epoch": 0.5935483870967742,
|
| 824 |
+
"grad_norm": 5.920025683401155,
|
| 825 |
+
"learning_rate": 2.333800140046683e-06,
|
| 826 |
+
"log_odds_chosen": 4.048437595367432,
|
| 827 |
+
"log_odds_ratio": -0.09355469048023224,
|
| 828 |
+
"logits/chosen": -2.242138624191284,
|
| 829 |
+
"logits/rejected": 15.487500190734863,
|
| 830 |
+
"logps/chosen": -0.994921863079071,
|
| 831 |
+
"logps/rejected": -4.592187404632568,
|
| 832 |
+
"loss": 1.0104,
|
| 833 |
+
"nll_loss": 1.0050780773162842,
|
| 834 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 835 |
+
"rewards/chosen": -0.04975586012005806,
|
| 836 |
+
"rewards/margins": 0.17998047173023224,
|
| 837 |
+
"rewards/rejected": -0.22968749701976776,
|
| 838 |
"step": 460
|
| 839 |
},
|
| 840 |
{
|
| 841 |
"epoch": 0.6064516129032258,
|
| 842 |
+
"grad_norm": 6.138559547910098,
|
| 843 |
+
"learning_rate": 2.3087854826980507e-06,
|
| 844 |
+
"log_odds_chosen": 3.6640625,
|
| 845 |
+
"log_odds_ratio": -0.1455078125,
|
| 846 |
+
"logits/chosen": -2.6703124046325684,
|
| 847 |
+
"logits/rejected": 14.762499809265137,
|
| 848 |
+
"logps/chosen": -1.03125,
|
| 849 |
+
"logps/rejected": -4.307812690734863,
|
| 850 |
+
"loss": 1.0394,
|
| 851 |
+
"nll_loss": 1.0320312976837158,
|
| 852 |
+
"rewards/accuracies": 0.925000011920929,
|
| 853 |
+
"rewards/chosen": -0.05156249925494194,
|
| 854 |
+
"rewards/margins": 0.16391602158546448,
|
| 855 |
+
"rewards/rejected": -0.21542969346046448,
|
| 856 |
"step": 470
|
| 857 |
},
|
| 858 |
{
|
| 859 |
"epoch": 0.6193548387096774,
|
| 860 |
+
"grad_norm": 5.567487993441089,
|
| 861 |
+
"learning_rate": 2.2845583119247696e-06,
|
| 862 |
+
"log_odds_chosen": 4.364062309265137,
|
| 863 |
+
"log_odds_ratio": -0.078857421875,
|
| 864 |
+
"logits/chosen": -2.549999952316284,
|
| 865 |
+
"logits/rejected": 15.662500381469727,
|
| 866 |
+
"logps/chosen": -0.9398437738418579,
|
| 867 |
+
"logps/rejected": -4.828125,
|
| 868 |
+
"loss": 0.9611,
|
| 869 |
+
"nll_loss": 0.9574218988418579,
|
| 870 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 871 |
+
"rewards/chosen": -0.04694824293255806,
|
| 872 |
+
"rewards/margins": 0.19462890923023224,
|
| 873 |
+
"rewards/rejected": -0.24169921875,
|
| 874 |
"step": 480
|
| 875 |
},
|
| 876 |
{
|
| 877 |
"epoch": 0.632258064516129,
|
| 878 |
+
"grad_norm": 5.862239199019293,
|
| 879 |
+
"learning_rate": 2.2610781582306733e-06,
|
| 880 |
+
"log_odds_chosen": 4.1484375,
|
| 881 |
+
"log_odds_ratio": -0.08124999701976776,
|
| 882 |
+
"logits/chosen": -2.7220215797424316,
|
| 883 |
+
"logits/rejected": 15.737500190734863,
|
| 884 |
+
"logps/chosen": -1.0242187976837158,
|
| 885 |
+
"logps/rejected": -4.734375,
|
| 886 |
+
"loss": 1.0391,
|
| 887 |
+
"nll_loss": 1.0343749523162842,
|
| 888 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 889 |
+
"rewards/chosen": -0.05124511569738388,
|
| 890 |
+
"rewards/margins": 0.18544921278953552,
|
| 891 |
+
"rewards/rejected": -0.23681640625,
|
| 892 |
"step": 490
|
| 893 |
},
|
| 894 |
{
|
| 895 |
"epoch": 0.6451612903225806,
|
| 896 |
+
"grad_norm": 5.118213222200493,
|
| 897 |
+
"learning_rate": 2.238307405179226e-06,
|
| 898 |
+
"log_odds_chosen": 3.965625047683716,
|
| 899 |
+
"log_odds_ratio": -0.10859374701976776,
|
| 900 |
+
"logits/chosen": -2.469921827316284,
|
| 901 |
+
"logits/rejected": 15.243749618530273,
|
| 902 |
+
"logps/chosen": -1.066796898841858,
|
| 903 |
+
"logps/rejected": -4.618750095367432,
|
| 904 |
+
"loss": 1.057,
|
| 905 |
+
"nll_loss": 1.051171898841858,
|
| 906 |
+
"rewards/accuracies": 0.96875,
|
| 907 |
+
"rewards/chosen": -0.05332031100988388,
|
| 908 |
+
"rewards/margins": 0.17753906548023224,
|
| 909 |
+
"rewards/rejected": -0.23095703125,
|
| 910 |
"step": 500
|
| 911 |
},
|
| 912 |
{
|
| 913 |
"epoch": 0.6580645161290323,
|
| 914 |
+
"grad_norm": 5.512232200800921,
|
| 915 |
+
"learning_rate": 2.2162110358896814e-06,
|
| 916 |
+
"log_odds_chosen": 3.957812547683716,
|
| 917 |
+
"log_odds_ratio": -0.08552245795726776,
|
| 918 |
+
"logits/chosen": -2.5843749046325684,
|
| 919 |
+
"logits/rejected": 15.137499809265137,
|
| 920 |
+
"logps/chosen": -0.9624999761581421,
|
| 921 |
+
"logps/rejected": -4.465624809265137,
|
| 922 |
+
"loss": 0.9945,
|
| 923 |
+
"nll_loss": 0.9906250238418579,
|
| 924 |
+
"rewards/accuracies": 0.981249988079071,
|
| 925 |
+
"rewards/chosen": -0.04819335788488388,
|
| 926 |
+
"rewards/margins": 0.175048828125,
|
| 927 |
+
"rewards/rejected": -0.22314453125,
|
| 928 |
"step": 510
|
| 929 |
},
|
| 930 |
{
|
| 931 |
"epoch": 0.6709677419354839,
|
| 932 |
+
"grad_norm": 5.353147766002135,
|
| 933 |
+
"learning_rate": 2.1947564065307354e-06,
|
| 934 |
+
"log_odds_chosen": 3.9203124046325684,
|
| 935 |
+
"log_odds_ratio": -0.08530273288488388,
|
| 936 |
+
"logits/chosen": -2.832812547683716,
|
| 937 |
+
"logits/rejected": 15.462499618530273,
|
| 938 |
+
"logps/chosen": -0.9613281488418579,
|
| 939 |
+
"logps/rejected": -4.428124904632568,
|
| 940 |
+
"loss": 0.9806,
|
| 941 |
+
"nll_loss": 0.9765625,
|
| 942 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 943 |
+
"rewards/chosen": -0.04807128757238388,
|
| 944 |
+
"rewards/margins": 0.17333984375,
|
| 945 |
+
"rewards/rejected": -0.22138671576976776,
|
| 946 |
"step": 520
|
| 947 |
},
|
| 948 |
{
|
| 949 |
"epoch": 0.6838709677419355,
|
| 950 |
+
"grad_norm": 5.53934453000343,
|
| 951 |
+
"learning_rate": 2.173913043478261e-06,
|
| 952 |
+
"log_odds_chosen": 3.6078124046325684,
|
| 953 |
+
"log_odds_ratio": -0.093505859375,
|
| 954 |
+
"logits/chosen": -1.6005859375,
|
| 955 |
+
"logits/rejected": 15.037500381469727,
|
| 956 |
+
"logps/chosen": -0.955859363079071,
|
| 957 |
+
"logps/rejected": -4.104687690734863,
|
| 958 |
+
"loss": 0.9783,
|
| 959 |
+
"nll_loss": 0.973828136920929,
|
| 960 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 961 |
+
"rewards/chosen": -0.04775390774011612,
|
| 962 |
+
"rewards/margins": 0.15732422471046448,
|
| 963 |
+
"rewards/rejected": -0.20488281548023224,
|
| 964 |
"step": 530
|
| 965 |
},
|
| 966 |
{
|
| 967 |
"epoch": 0.6967741935483871,
|
| 968 |
+
"grad_norm": 5.046849897934945,
|
| 969 |
+
"learning_rate": 2.1536524612697402e-06,
|
| 970 |
+
"log_odds_chosen": 3.75,
|
| 971 |
+
"log_odds_ratio": -0.094482421875,
|
| 972 |
+
"logits/chosen": -1.859375,
|
| 973 |
+
"logits/rejected": 15.962499618530273,
|
| 974 |
+
"logps/chosen": -1.030859351158142,
|
| 975 |
+
"logps/rejected": -4.309374809265137,
|
| 976 |
+
"loss": 0.9959,
|
| 977 |
+
"nll_loss": 0.9917968511581421,
|
| 978 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 979 |
+
"rewards/chosen": -0.05153808742761612,
|
| 980 |
+
"rewards/margins": 0.16396483778953552,
|
| 981 |
+
"rewards/rejected": -0.21562500298023224,
|
| 982 |
"step": 540
|
| 983 |
},
|
| 984 |
{
|
| 985 |
"epoch": 0.7096774193548387,
|
| 986 |
+
"grad_norm": 5.947499132612451,
|
| 987 |
+
"learning_rate": 2.1339479988816e-06,
|
| 988 |
+
"log_odds_chosen": 3.6875,
|
| 989 |
+
"log_odds_ratio": -0.06818847358226776,
|
| 990 |
+
"logits/chosen": -2.174999952316284,
|
| 991 |
+
"logits/rejected": 15.493749618530273,
|
| 992 |
+
"logps/chosen": -0.999218761920929,
|
| 993 |
+
"logps/rejected": -4.228125095367432,
|
| 994 |
+
"loss": 1.003,
|
| 995 |
+
"nll_loss": 1.0,
|
| 996 |
+
"rewards/accuracies": 0.987500011920929,
|
| 997 |
+
"rewards/chosen": -0.0499267578125,
|
| 998 |
+
"rewards/margins": 0.16171875596046448,
|
| 999 |
+
"rewards/rejected": -0.21142578125,
|
| 1000 |
"step": 550
|
| 1001 |
},
|
| 1002 |
{
|
| 1003 |
"epoch": 0.7225806451612903,
|
| 1004 |
+
"grad_norm": 5.351752547147555,
|
| 1005 |
+
"learning_rate": 2.114774672189068e-06,
|
| 1006 |
+
"log_odds_chosen": 3.862499952316284,
|
| 1007 |
+
"log_odds_ratio": -0.08652343600988388,
|
| 1008 |
+
"logits/chosen": -2.469531297683716,
|
| 1009 |
+
"logits/rejected": 15.331250190734863,
|
| 1010 |
+
"logps/chosen": -0.9957031011581421,
|
| 1011 |
+
"logps/rejected": -4.415625095367432,
|
| 1012 |
+
"loss": 1.0154,
|
| 1013 |
+
"nll_loss": 1.011328101158142,
|
| 1014 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1015 |
+
"rewards/chosen": -0.04978027194738388,
|
| 1016 |
+
"rewards/margins": 0.17119140923023224,
|
| 1017 |
+
"rewards/rejected": -0.22089843451976776,
|
| 1018 |
"step": 560
|
| 1019 |
},
|
| 1020 |
{
|
| 1021 |
"epoch": 0.7354838709677419,
|
| 1022 |
+
"grad_norm": 4.936722562656586,
|
| 1023 |
+
"learning_rate": 2.0961090407515926e-06,
|
| 1024 |
+
"log_odds_chosen": 4.0703125,
|
| 1025 |
+
"log_odds_ratio": -0.06992187350988388,
|
| 1026 |
+
"logits/chosen": -2.4644532203674316,
|
| 1027 |
+
"logits/rejected": 15.46875,
|
| 1028 |
+
"logps/chosen": -0.9624999761581421,
|
| 1029 |
+
"logps/rejected": -4.567187309265137,
|
| 1030 |
+
"loss": 0.9827,
|
| 1031 |
+
"nll_loss": 0.979687511920929,
|
| 1032 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1033 |
+
"rewards/chosen": -0.04814453050494194,
|
| 1034 |
+
"rewards/margins": 0.18017578125,
|
| 1035 |
+
"rewards/rejected": -0.22832031548023224,
|
| 1036 |
"step": 570
|
| 1037 |
},
|
| 1038 |
{
|
| 1039 |
"epoch": 0.7483870967741936,
|
| 1040 |
+
"grad_norm": 4.997513662861683,
|
| 1041 |
+
"learning_rate": 2.077929087308457e-06,
|
| 1042 |
+
"log_odds_chosen": 3.856250047683716,
|
| 1043 |
+
"log_odds_ratio": -0.07062987983226776,
|
| 1044 |
+
"logits/chosen": -2.723437547683716,
|
| 1045 |
+
"logits/rejected": 14.800000190734863,
|
| 1046 |
+
"logps/chosen": -0.94140625,
|
| 1047 |
+
"logps/rejected": -4.317187309265137,
|
| 1048 |
+
"loss": 0.9656,
|
| 1049 |
+
"nll_loss": 0.9624999761581421,
|
| 1050 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1051 |
+
"rewards/chosen": -0.047119140625,
|
| 1052 |
+
"rewards/margins": 0.16884765028953552,
|
| 1053 |
+
"rewards/rejected": -0.21562500298023224,
|
| 1054 |
"step": 580
|
| 1055 |
},
|
| 1056 |
{
|
| 1057 |
"epoch": 0.7612903225806451,
|
| 1058 |
+
"grad_norm": 5.381405087912754,
|
| 1059 |
+
"learning_rate": 2.0602141085758233e-06,
|
| 1060 |
+
"log_odds_chosen": 4.1015625,
|
| 1061 |
+
"log_odds_ratio": -0.06821288913488388,
|
| 1062 |
+
"logits/chosen": -2.957812547683716,
|
| 1063 |
+
"logits/rejected": 15.106249809265137,
|
| 1064 |
+
"logps/chosen": -0.9488281011581421,
|
| 1065 |
+
"logps/rejected": -4.574999809265137,
|
| 1066 |
+
"loss": 0.976,
|
| 1067 |
+
"nll_loss": 0.973437488079071,
|
| 1068 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1069 |
+
"rewards/chosen": -0.04743652418255806,
|
| 1070 |
+
"rewards/margins": 0.18144531548023224,
|
| 1071 |
+
"rewards/rejected": -0.22861328721046448,
|
| 1072 |
"step": 590
|
| 1073 |
},
|
| 1074 |
{
|
| 1075 |
"epoch": 0.7741935483870968,
|
| 1076 |
+
"grad_norm": 5.488636413342299,
|
| 1077 |
+
"learning_rate": 2.0429446161135922e-06,
|
| 1078 |
+
"log_odds_chosen": 4.181250095367432,
|
| 1079 |
+
"log_odds_ratio": -0.06918945163488388,
|
| 1080 |
+
"logits/chosen": -2.314160108566284,
|
| 1081 |
+
"logits/rejected": 15.449999809265137,
|
| 1082 |
+
"logps/chosen": -0.9921875,
|
| 1083 |
+
"logps/rejected": -4.721875190734863,
|
| 1084 |
+
"loss": 1.0089,
|
| 1085 |
+
"nll_loss": 1.0050780773162842,
|
| 1086 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1087 |
+
"rewards/chosen": -0.04958496242761612,
|
| 1088 |
+
"rewards/margins": 0.18642577528953552,
|
| 1089 |
+
"rewards/rejected": -0.23593750596046448,
|
| 1090 |
"step": 600
|
| 1091 |
},
|
| 1092 |
{
|
| 1093 |
"epoch": 0.7870967741935484,
|
| 1094 |
+
"grad_norm": 5.427887234676386,
|
| 1095 |
+
"learning_rate": 2.0261022461827697e-06,
|
| 1096 |
+
"log_odds_chosen": 4.599999904632568,
|
| 1097 |
+
"log_odds_ratio": -0.06470946967601776,
|
| 1098 |
+
"logits/chosen": -3.4375,
|
| 1099 |
+
"logits/rejected": 15.268750190734863,
|
| 1100 |
+
"logps/chosen": -0.998046875,
|
| 1101 |
+
"logps/rejected": -5.140625,
|
| 1102 |
+
"loss": 1.0151,
|
| 1103 |
+
"nll_loss": 1.0128905773162842,
|
| 1104 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1105 |
+
"rewards/chosen": -0.04990234225988388,
|
| 1106 |
+
"rewards/margins": 0.20703125,
|
| 1107 |
+
"rewards/rejected": -0.25712889432907104,
|
| 1108 |
"step": 610
|
| 1109 |
},
|
| 1110 |
{
|
| 1111 |
"epoch": 0.8,
|
| 1112 |
+
"grad_norm": 6.166954138248167,
|
| 1113 |
+
"learning_rate": 2.009669677645352e-06,
|
| 1114 |
+
"log_odds_chosen": 4.235937595367432,
|
| 1115 |
+
"log_odds_ratio": -0.08128662407398224,
|
| 1116 |
+
"logits/chosen": -2.9867186546325684,
|
| 1117 |
+
"logits/rejected": 15.600000381469727,
|
| 1118 |
+
"logps/chosen": -0.962890625,
|
| 1119 |
+
"logps/rejected": -4.724999904632568,
|
| 1120 |
+
"loss": 0.9886,
|
| 1121 |
+
"nll_loss": 0.985156238079071,
|
| 1122 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1123 |
+
"rewards/chosen": -0.04814453050494194,
|
| 1124 |
+
"rewards/margins": 0.18828125298023224,
|
| 1125 |
+
"rewards/rejected": -0.23642578721046448,
|
| 1126 |
"step": 620
|
| 1127 |
},
|
| 1128 |
{
|
| 1129 |
"epoch": 0.8129032258064516,
|
| 1130 |
+
"grad_norm": 5.497842991013724,
|
| 1131 |
+
"learning_rate": 1.9936305570722503e-06,
|
| 1132 |
+
"log_odds_chosen": 4.198437690734863,
|
| 1133 |
+
"log_odds_ratio": -0.05922241136431694,
|
| 1134 |
+
"logits/chosen": -2.833984375,
|
| 1135 |
+
"logits/rejected": 15.46875,
|
| 1136 |
+
"logps/chosen": -0.94921875,
|
| 1137 |
+
"logps/rejected": -4.667187690734863,
|
| 1138 |
+
"loss": 0.9591,
|
| 1139 |
+
"nll_loss": 0.956250011920929,
|
| 1140 |
+
"rewards/accuracies": 1.0,
|
| 1141 |
+
"rewards/chosen": -0.04746093600988388,
|
| 1142 |
+
"rewards/margins": 0.18569335341453552,
|
| 1143 |
+
"rewards/rejected": -0.23310546576976776,
|
| 1144 |
"step": 630
|
| 1145 |
},
|
| 1146 |
{
|
| 1147 |
"epoch": 0.8258064516129032,
|
| 1148 |
+
"grad_norm": 8.242142638367945,
|
| 1149 |
+
"learning_rate": 1.977969430323089e-06,
|
| 1150 |
+
"log_odds_chosen": 4.364062309265137,
|
| 1151 |
+
"log_odds_ratio": -0.06558837741613388,
|
| 1152 |
+
"logits/chosen": -2.439453125,
|
| 1153 |
+
"logits/rejected": 15.581250190734863,
|
| 1154 |
+
"logps/chosen": -1.0246093273162842,
|
| 1155 |
+
"logps/rejected": -4.918749809265137,
|
| 1156 |
+
"loss": 1.0151,
|
| 1157 |
+
"nll_loss": 1.011328101158142,
|
| 1158 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1159 |
+
"rewards/chosen": -0.05124511569738388,
|
| 1160 |
+
"rewards/margins": 0.19462890923023224,
|
| 1161 |
+
"rewards/rejected": -0.24580077826976776,
|
| 1162 |
"step": 640
|
| 1163 |
},
|
| 1164 |
{
|
| 1165 |
"epoch": 0.8387096774193549,
|
| 1166 |
+
"grad_norm": 5.407957699348334,
|
| 1167 |
+
"learning_rate": 1.962671679947149e-06,
|
| 1168 |
+
"log_odds_chosen": 4.728125095367432,
|
| 1169 |
+
"log_odds_ratio": -0.04775390774011612,
|
| 1170 |
+
"logits/chosen": -2.6484375,
|
| 1171 |
+
"logits/rejected": 16.125,
|
| 1172 |
+
"logps/chosen": -1.0050780773162842,
|
| 1173 |
+
"logps/rejected": -5.262499809265137,
|
| 1174 |
+
"loss": 1.0211,
|
| 1175 |
+
"nll_loss": 1.019140601158142,
|
| 1176 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1177 |
+
"rewards/chosen": -0.0501708984375,
|
| 1178 |
+
"rewards/margins": 0.21298828721046448,
|
| 1179 |
+
"rewards/rejected": -0.2630859315395355,
|
| 1180 |
"step": 650
|
| 1181 |
},
|
| 1182 |
{
|
| 1183 |
"epoch": 0.8516129032258064,
|
| 1184 |
+
"grad_norm": 5.429871193266642,
|
| 1185 |
+
"learning_rate": 1.9477234678290497e-06,
|
| 1186 |
+
"log_odds_chosen": 4.28125,
|
| 1187 |
+
"log_odds_ratio": -0.06833495944738388,
|
| 1188 |
+
"logits/chosen": -2.073046922683716,
|
| 1189 |
+
"logits/rejected": 15.493749618530273,
|
| 1190 |
+
"logps/chosen": -1.019921898841858,
|
| 1191 |
+
"logps/rejected": -4.854687690734863,
|
| 1192 |
+
"loss": 1.0178,
|
| 1193 |
+
"nll_loss": 1.0148437023162842,
|
| 1194 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1195 |
+
"rewards/chosen": -0.05095214769244194,
|
| 1196 |
+
"rewards/margins": 0.19204100966453552,
|
| 1197 |
+
"rewards/rejected": -0.24287109076976776,
|
| 1198 |
"step": 660
|
| 1199 |
},
|
| 1200 |
{
|
| 1201 |
"epoch": 0.864516129032258,
|
| 1202 |
+
"grad_norm": 5.38838761477697,
|
| 1203 |
+
"learning_rate": 1.9331116825676476e-06,
|
| 1204 |
+
"log_odds_chosen": 4.264062404632568,
|
| 1205 |
+
"log_odds_ratio": -0.07292480766773224,
|
| 1206 |
+
"logits/chosen": -2.7894530296325684,
|
| 1207 |
+
"logits/rejected": 15.3125,
|
| 1208 |
+
"logps/chosen": -0.9847656488418579,
|
| 1209 |
+
"logps/rejected": -4.784375190734863,
|
| 1210 |
+
"loss": 1.0014,
|
| 1211 |
+
"nll_loss": 0.998046875,
|
| 1212 |
"rewards/accuracies": 0.987500011920929,
|
| 1213 |
+
"rewards/chosen": -0.04924316331744194,
|
| 1214 |
+
"rewards/margins": 0.18974609673023224,
|
| 1215 |
+
"rewards/rejected": -0.2392578125,
|
| 1216 |
"step": 670
|
| 1217 |
},
|
| 1218 |
{
|
| 1219 |
"epoch": 0.8774193548387097,
|
| 1220 |
+
"grad_norm": 6.231144536945813,
|
| 1221 |
+
"learning_rate": 1.9188238911333405e-06,
|
| 1222 |
+
"log_odds_chosen": 4.495312690734863,
|
| 1223 |
+
"log_odds_ratio": -0.0684814453125,
|
| 1224 |
+
"logits/chosen": -2.408984422683716,
|
| 1225 |
+
"logits/rejected": 15.449999809265137,
|
| 1226 |
+
"logps/chosen": -1.029687523841858,
|
| 1227 |
+
"logps/rejected": -5.090624809265137,
|
| 1228 |
+
"loss": 1.0266,
|
| 1229 |
+
"nll_loss": 1.023828148841858,
|
| 1230 |
"rewards/accuracies": 0.987500011920929,
|
| 1231 |
+
"rewards/chosen": -0.05156249925494194,
|
| 1232 |
+
"rewards/margins": 0.20302733778953552,
|
| 1233 |
+
"rewards/rejected": -0.25439453125,
|
| 1234 |
"step": 680
|
| 1235 |
},
|
| 1236 |
{
|
| 1237 |
"epoch": 0.8903225806451613,
|
| 1238 |
+
"grad_norm": 5.900000733774859,
|
| 1239 |
+
"learning_rate": 1.9048482943986487e-06,
|
| 1240 |
+
"log_odds_chosen": 4.173437595367432,
|
| 1241 |
+
"log_odds_ratio": -0.10228271782398224,
|
| 1242 |
+
"logits/chosen": -2.411328077316284,
|
| 1243 |
+
"logits/rejected": 15.0625,
|
| 1244 |
+
"logps/chosen": -0.987109363079071,
|
| 1245 |
+
"logps/rejected": -4.704687595367432,
|
| 1246 |
+
"loss": 0.9854,
|
| 1247 |
+
"nll_loss": 0.981249988079071,
|
| 1248 |
"rewards/accuracies": 0.9750000238418579,
|
| 1249 |
+
"rewards/chosen": -0.04938964918255806,
|
| 1250 |
+
"rewards/margins": 0.18564453721046448,
|
| 1251 |
+
"rewards/rejected": -0.23476561903953552,
|
| 1252 |
"step": 690
|
| 1253 |
},
|
| 1254 |
{
|
| 1255 |
"epoch": 0.9032258064516129,
|
| 1256 |
+
"grad_norm": 5.5507604309289045,
|
| 1257 |
+
"learning_rate": 1.8911736861805847e-06,
|
| 1258 |
+
"log_odds_chosen": 3.984375,
|
| 1259 |
+
"log_odds_ratio": -0.08645019680261612,
|
| 1260 |
+
"logits/chosen": -2.534374952316284,
|
| 1261 |
+
"logits/rejected": 15.199999809265137,
|
| 1262 |
+
"logps/chosen": -0.946484386920929,
|
| 1263 |
+
"logps/rejected": -4.471875190734863,
|
| 1264 |
+
"loss": 0.9628,
|
| 1265 |
+
"nll_loss": 0.958984375,
|
| 1266 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1267 |
+
"rewards/chosen": -0.04731445387005806,
|
| 1268 |
+
"rewards/margins": 0.17587891221046448,
|
| 1269 |
+
"rewards/rejected": -0.22343750298023224,
|
| 1270 |
"step": 700
|
| 1271 |
},
|
| 1272 |
{
|
| 1273 |
"epoch": 0.9161290322580645,
|
| 1274 |
+
"grad_norm": 5.72722978193488,
|
| 1275 |
+
"learning_rate": 1.8777894154716994e-06,
|
| 1276 |
+
"log_odds_chosen": 4.260937690734863,
|
| 1277 |
+
"log_odds_ratio": -0.08770751953125,
|
| 1278 |
+
"logits/chosen": -2.785937547683716,
|
| 1279 |
+
"logits/rejected": 15.212499618530273,
|
| 1280 |
+
"logps/chosen": -0.975781261920929,
|
| 1281 |
+
"logps/rejected": -4.790625095367432,
|
| 1282 |
+
"loss": 0.9996,
|
| 1283 |
+
"nll_loss": 0.9957031011581421,
|
| 1284 |
+
"rewards/accuracies": 0.96875,
|
| 1285 |
+
"rewards/chosen": -0.048828125,
|
| 1286 |
+
"rewards/margins": 0.19072265923023224,
|
| 1287 |
+
"rewards/rejected": -0.23945312201976776,
|
| 1288 |
"step": 710
|
| 1289 |
},
|
| 1290 |
{
|
| 1291 |
"epoch": 0.9290322580645162,
|
| 1292 |
+
"grad_norm": 5.182886168475235,
|
| 1293 |
+
"learning_rate": 1.864685351570504e-06,
|
| 1294 |
+
"log_odds_chosen": 4.417187690734863,
|
| 1295 |
+
"log_odds_ratio": -0.05722656100988388,
|
| 1296 |
+
"logits/chosen": -2.477343797683716,
|
| 1297 |
+
"logits/rejected": 15.387499809265137,
|
| 1298 |
+
"logps/chosen": -0.953906238079071,
|
| 1299 |
+
"logps/rejected": -4.900000095367432,
|
| 1300 |
+
"loss": 0.9764,
|
| 1301 |
+
"nll_loss": 0.973437488079071,
|
| 1302 |
+
"rewards/accuracies": 1.0,
|
| 1303 |
+
"rewards/chosen": -0.04770507663488388,
|
| 1304 |
+
"rewards/margins": 0.19716796278953552,
|
| 1305 |
+
"rewards/rejected": -0.24492187798023224,
|
| 1306 |
"step": 720
|
| 1307 |
},
|
| 1308 |
{
|
| 1309 |
"epoch": 0.9419354838709677,
|
| 1310 |
+
"grad_norm": 6.373900642995302,
|
| 1311 |
+
"learning_rate": 1.8518518518518519e-06,
|
| 1312 |
+
"log_odds_chosen": 3.9078125953674316,
|
| 1313 |
+
"log_odds_ratio": -0.07728271186351776,
|
| 1314 |
+
"logits/chosen": -2.785937547683716,
|
| 1315 |
+
"logits/rejected": 15.181249618530273,
|
| 1316 |
+
"logps/chosen": -0.979687511920929,
|
| 1317 |
+
"logps/rejected": -4.4296875,
|
| 1318 |
+
"loss": 0.9999,
|
| 1319 |
+
"nll_loss": 0.9957031011581421,
|
| 1320 |
+
"rewards/accuracies": 0.981249988079071,
|
| 1321 |
+
"rewards/chosen": -0.049072265625,
|
| 1322 |
+
"rewards/margins": 0.17246094346046448,
|
| 1323 |
+
"rewards/rejected": -0.22158202528953552,
|
| 1324 |
"step": 730
|
| 1325 |
},
|
| 1326 |
{
|
| 1327 |
"epoch": 0.9548387096774194,
|
| 1328 |
+
"grad_norm": 5.301970324773416,
|
| 1329 |
+
"learning_rate": 1.8392797319442742e-06,
|
| 1330 |
+
"log_odds_chosen": 3.7484374046325684,
|
| 1331 |
+
"log_odds_ratio": -0.07939453423023224,
|
| 1332 |
+
"logits/chosen": -3.1468749046325684,
|
| 1333 |
+
"logits/rejected": 14.743749618530273,
|
| 1334 |
+
"logps/chosen": -0.958203136920929,
|
| 1335 |
+
"logps/rejected": -4.229687690734863,
|
| 1336 |
+
"loss": 0.9549,
|
| 1337 |
+
"nll_loss": 0.9527343511581421,
|
| 1338 |
+
"rewards/accuracies": 1.0,
|
| 1339 |
+
"rewards/chosen": -0.04790038987994194,
|
| 1340 |
+
"rewards/margins": 0.16396483778953552,
|
| 1341 |
+
"rewards/rejected": -0.21162109076976776,
|
| 1342 |
"step": 740
|
| 1343 |
},
|
| 1344 |
{
|
| 1345 |
"epoch": 0.967741935483871,
|
| 1346 |
+
"grad_norm": 5.607236715482459,
|
| 1347 |
+
"learning_rate": 1.8269602381046742e-06,
|
| 1348 |
+
"log_odds_chosen": 4.028124809265137,
|
| 1349 |
+
"log_odds_ratio": -0.06224365159869194,
|
| 1350 |
+
"logits/chosen": -2.905468702316284,
|
| 1351 |
+
"logits/rejected": 15.387499809265137,
|
| 1352 |
+
"logps/chosen": -0.992968738079071,
|
| 1353 |
+
"logps/rejected": -4.550000190734863,
|
| 1354 |
+
"loss": 1.0119,
|
| 1355 |
+
"nll_loss": 1.010156273841858,
|
| 1356 |
+
"rewards/accuracies": 0.9937499761581421,
|
| 1357 |
+
"rewards/chosen": -0.04965820163488388,
|
| 1358 |
+
"rewards/margins": 0.17753906548023224,
|
| 1359 |
+
"rewards/rejected": -0.2275390625,
|
| 1360 |
"step": 750
|
| 1361 |
},
|
| 1362 |
{
|
| 1363 |
"epoch": 0.9806451612903225,
|
| 1364 |
+
"grad_norm": 5.46826560416884,
|
| 1365 |
+
"learning_rate": 1.8148850216015693e-06,
|
| 1366 |
+
"log_odds_chosen": 3.9203124046325684,
|
| 1367 |
+
"log_odds_ratio": -0.08592529594898224,
|
| 1368 |
+
"logits/chosen": -2.5835938453674316,
|
| 1369 |
+
"logits/rejected": 15.125,
|
| 1370 |
+
"logps/chosen": -0.974609375,
|
| 1371 |
+
"logps/rejected": -4.432812690734863,
|
| 1372 |
+
"loss": 0.9957,
|
| 1373 |
+
"nll_loss": 0.9925781488418579,
|
| 1374 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1375 |
+
"rewards/chosen": -0.04873047024011612,
|
| 1376 |
+
"rewards/margins": 0.17304687201976776,
|
| 1377 |
+
"rewards/rejected": -0.22158202528953552,
|
| 1378 |
"step": 760
|
| 1379 |
},
|
| 1380 |
{
|
| 1381 |
"epoch": 0.9935483870967742,
|
| 1382 |
+
"grad_norm": 6.701663744929856,
|
| 1383 |
+
"learning_rate": 1.803046114936547e-06,
|
| 1384 |
+
"log_odds_chosen": 4.103125095367432,
|
| 1385 |
+
"log_odds_ratio": -0.0823974609375,
|
| 1386 |
+
"logits/chosen": -3.176562547683716,
|
| 1387 |
+
"logits/rejected": 14.993749618530273,
|
| 1388 |
+
"logps/chosen": -0.97265625,
|
| 1389 |
+
"logps/rejected": -4.615624904632568,
|
| 1390 |
+
"loss": 0.9939,
|
| 1391 |
+
"nll_loss": 0.9906250238418579,
|
| 1392 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1393 |
+
"rewards/chosen": -0.048583984375,
|
| 1394 |
+
"rewards/margins": 0.18212890625,
|
| 1395 |
+
"rewards/rejected": -0.23056641221046448,
|
| 1396 |
"step": 770
|
| 1397 |
},
|
| 1398 |
{
|
| 1399 |
"epoch": 1.0,
|
| 1400 |
"step": 775,
|
| 1401 |
"total_flos": 0.0,
|
| 1402 |
+
"train_loss": 1.0746982106854839,
|
| 1403 |
+
"train_runtime": 1520.1443,
|
| 1404 |
+
"train_samples_per_second": 8.157,
|
| 1405 |
+
"train_steps_per_second": 0.51
|
| 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 sha256:1c9f48d2605a16695bdd6a65f12d736b4e7c6a0fb9d67dcff3355744a75a6822
|
| 3 |
size 7032
|