Model save
Browse files- README.md +12 -12
- adapter_model.safetensors +1 -1
- all_results.json +17 -17
- eval_results.json +14 -14
- train_results.json +3 -3
- trainer_state.json +999 -999
README.md
CHANGED
|
@@ -15,17 +15,17 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
-
- Loss: 0.
|
| 19 |
-
- Rewards/chosen: -
|
| 20 |
-
- Rewards/rejected: -
|
| 21 |
-
- Rewards/accuracies: 0.
|
| 22 |
-
- Rewards/margins:
|
| 23 |
-
- Logps/rejected: -
|
| 24 |
-
- Logps/chosen: -
|
| 25 |
-
- Logits/rejected: -2.
|
| 26 |
-
- Logits/chosen: -2.
|
| 27 |
-
- Use Label:
|
| 28 |
-
- Pred Label:
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
@@ -62,7 +62,7 @@ The following hyperparameters were used during training:
|
|
| 62 |
|
| 63 |
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen | Use Label | Pred Label |
|
| 64 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:---------:|:----------:|
|
| 65 |
-
| 0.
|
| 66 |
|
| 67 |
|
| 68 |
### Framework versions
|
|
|
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.3439
|
| 19 |
+
- Rewards/chosen: -1.1633
|
| 20 |
+
- Rewards/rejected: -3.5290
|
| 21 |
+
- Rewards/accuracies: 0.7420
|
| 22 |
+
- Rewards/margins: 2.3657
|
| 23 |
+
- Logps/rejected: -294.5901
|
| 24 |
+
- Logps/chosen: -295.8908
|
| 25 |
+
- Logits/rejected: -2.7390
|
| 26 |
+
- Logits/chosen: -2.7421
|
| 27 |
+
- Use Label: 9180.7998
|
| 28 |
+
- Pred Label: 6851.2002
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
|
|
| 62 |
|
| 63 |
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen | Use Label | Pred Label |
|
| 64 |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|:---------:|:----------:|
|
| 65 |
+
| 0.333 | 1.0 | 955 | 0.3439 | -1.1633 | -3.5290 | 0.7420 | 2.3657 | -294.5901 | -295.8908 | -2.7390 | -2.7421 | 8950.7998 | 6581.2002 |
|
| 66 |
|
| 67 |
|
| 68 |
### Framework versions
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 218138576
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:940dab0d56f0e97f38a28e754a805f85470394ea5666743d488b36e659bc665a
|
| 3 |
size 218138576
|
all_results.json
CHANGED
|
@@ -1,23 +1,23 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_logits/chosen": -2.
|
| 4 |
-
"eval_logits/rejected": -2.
|
| 5 |
-
"eval_logps/chosen": -
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 0.
|
| 8 |
-
"eval_pred_label":
|
| 9 |
-
"eval_rewards/accuracies": 0.
|
| 10 |
-
"eval_rewards/chosen": -
|
| 11 |
-
"eval_rewards/margins":
|
| 12 |
-
"eval_rewards/rejected": -
|
| 13 |
-
"eval_runtime":
|
| 14 |
"eval_samples": 2000,
|
| 15 |
-
"eval_samples_per_second": 4.
|
| 16 |
-
"eval_steps_per_second": 0.
|
| 17 |
-
"eval_use_label":
|
| 18 |
-
"train_loss": 0.
|
| 19 |
-
"train_runtime":
|
| 20 |
"train_samples": 61135,
|
| 21 |
-
"train_samples_per_second": 2.
|
| 22 |
"train_steps_per_second": 0.038
|
| 23 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -2.742107391357422,
|
| 4 |
+
"eval_logits/rejected": -2.739009141921997,
|
| 5 |
+
"eval_logps/chosen": -295.89080810546875,
|
| 6 |
+
"eval_logps/rejected": -294.5900573730469,
|
| 7 |
+
"eval_loss": 0.3439472019672394,
|
| 8 |
+
"eval_pred_label": 6851.2001953125,
|
| 9 |
+
"eval_rewards/accuracies": 0.7419999837875366,
|
| 10 |
+
"eval_rewards/chosen": -1.163341999053955,
|
| 11 |
+
"eval_rewards/margins": 2.3656928539276123,
|
| 12 |
+
"eval_rewards/rejected": -3.5290346145629883,
|
| 13 |
+
"eval_runtime": 460.2253,
|
| 14 |
"eval_samples": 2000,
|
| 15 |
+
"eval_samples_per_second": 4.346,
|
| 16 |
+
"eval_steps_per_second": 0.272,
|
| 17 |
+
"eval_use_label": 9180.7998046875,
|
| 18 |
+
"train_loss": 0.39464539333163756,
|
| 19 |
+
"train_runtime": 25354.6696,
|
| 20 |
"train_samples": 61135,
|
| 21 |
+
"train_samples_per_second": 2.411,
|
| 22 |
"train_steps_per_second": 0.038
|
| 23 |
}
|
eval_results.json
CHANGED
|
@@ -1,18 +1,18 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_logits/chosen": -2.
|
| 4 |
-
"eval_logits/rejected": -2.
|
| 5 |
-
"eval_logps/chosen": -
|
| 6 |
-
"eval_logps/rejected": -
|
| 7 |
-
"eval_loss": 0.
|
| 8 |
-
"eval_pred_label":
|
| 9 |
-
"eval_rewards/accuracies": 0.
|
| 10 |
-
"eval_rewards/chosen": -
|
| 11 |
-
"eval_rewards/margins":
|
| 12 |
-
"eval_rewards/rejected": -
|
| 13 |
-
"eval_runtime":
|
| 14 |
"eval_samples": 2000,
|
| 15 |
-
"eval_samples_per_second": 4.
|
| 16 |
-
"eval_steps_per_second": 0.
|
| 17 |
-
"eval_use_label":
|
| 18 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_logits/chosen": -2.742107391357422,
|
| 4 |
+
"eval_logits/rejected": -2.739009141921997,
|
| 5 |
+
"eval_logps/chosen": -295.89080810546875,
|
| 6 |
+
"eval_logps/rejected": -294.5900573730469,
|
| 7 |
+
"eval_loss": 0.3439472019672394,
|
| 8 |
+
"eval_pred_label": 6851.2001953125,
|
| 9 |
+
"eval_rewards/accuracies": 0.7419999837875366,
|
| 10 |
+
"eval_rewards/chosen": -1.163341999053955,
|
| 11 |
+
"eval_rewards/margins": 2.3656928539276123,
|
| 12 |
+
"eval_rewards/rejected": -3.5290346145629883,
|
| 13 |
+
"eval_runtime": 460.2253,
|
| 14 |
"eval_samples": 2000,
|
| 15 |
+
"eval_samples_per_second": 4.346,
|
| 16 |
+
"eval_steps_per_second": 0.272,
|
| 17 |
+
"eval_use_label": 9180.7998046875
|
| 18 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 61135,
|
| 6 |
-
"train_samples_per_second": 2.
|
| 7 |
"train_steps_per_second": 0.038
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.39464539333163756,
|
| 4 |
+
"train_runtime": 25354.6696,
|
| 5 |
"train_samples": 61135,
|
| 6 |
+
"train_samples_per_second": 2.411,
|
| 7 |
"train_steps_per_second": 0.038
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -80,13 +80,13 @@
|
|
| 80 |
"logps/chosen": -281.32928466796875,
|
| 81 |
"logps/rejected": -277.8607482910156,
|
| 82 |
"loss": 0.6339,
|
| 83 |
-
"pred_label": 0.
|
| 84 |
"rewards/accuracies": 0.6875,
|
| 85 |
"rewards/chosen": 0.02641097828745842,
|
| 86 |
"rewards/margins": 0.2079576551914215,
|
| 87 |
"rewards/rejected": -0.1815466731786728,
|
| 88 |
"step": 40,
|
| 89 |
-
"use_label":
|
| 90 |
},
|
| 91 |
{
|
| 92 |
"epoch": 0.05,
|
|
@@ -96,1479 +96,1479 @@
|
|
| 96 |
"logps/chosen": -266.80517578125,
|
| 97 |
"logps/rejected": -261.9176025390625,
|
| 98 |
"loss": 0.5914,
|
| 99 |
-
"pred_label":
|
| 100 |
"rewards/accuracies": 0.643750011920929,
|
| 101 |
"rewards/chosen": -0.06968289613723755,
|
| 102 |
"rewards/margins": 0.33043327927589417,
|
| 103 |
"rewards/rejected": -0.4001162648200989,
|
| 104 |
"step": 50,
|
| 105 |
-
"use_label":
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.06,
|
| 109 |
"learning_rate": 3.125e-05,
|
| 110 |
-
"logits/chosen": -2.
|
| 111 |
-
"logits/rejected": -2.
|
| 112 |
-
"logps/chosen": -300.
|
| 113 |
-
"logps/rejected": -290.
|
| 114 |
-
"loss": 0.
|
| 115 |
-
"pred_label":
|
| 116 |
"rewards/accuracies": 0.699999988079071,
|
| 117 |
-
"rewards/chosen": -0.
|
| 118 |
-
"rewards/margins": 0.
|
| 119 |
-
"rewards/rejected": -0.
|
| 120 |
"step": 60,
|
| 121 |
-
"use_label":
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.07,
|
| 125 |
"learning_rate": 3.6458333333333336e-05,
|
| 126 |
-
"logits/chosen": -2.
|
| 127 |
-
"logits/rejected": -2.
|
| 128 |
-
"logps/chosen": -294.
|
| 129 |
-
"logps/rejected": -278.
|
| 130 |
-
"loss": 0.
|
| 131 |
-
"pred_label":
|
| 132 |
-
"rewards/accuracies": 0.
|
| 133 |
-
"rewards/chosen": -0.
|
| 134 |
-
"rewards/margins": 0.
|
| 135 |
-
"rewards/rejected": -0.
|
| 136 |
"step": 70,
|
| 137 |
-
"use_label":
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 0.08,
|
| 141 |
"learning_rate": 4.166666666666667e-05,
|
| 142 |
-
"logits/chosen": -2.
|
| 143 |
-
"logits/rejected": -2.
|
| 144 |
-
"logps/chosen": -306.
|
| 145 |
-
"logps/rejected": -288.
|
| 146 |
-
"loss": 0.
|
| 147 |
-
"pred_label":
|
| 148 |
-
"rewards/accuracies": 0.
|
| 149 |
-
"rewards/chosen": -0.
|
| 150 |
-
"rewards/margins": 0.
|
| 151 |
-
"rewards/rejected": -0.
|
| 152 |
"step": 80,
|
| 153 |
-
"use_label":
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.09,
|
| 157 |
"learning_rate": 4.6875e-05,
|
| 158 |
-
"logits/chosen": -2.
|
| 159 |
-
"logits/rejected": -2.
|
| 160 |
-
"logps/chosen": -295.
|
| 161 |
-
"logps/rejected": -
|
| 162 |
-
"loss": 0.
|
| 163 |
-
"pred_label":
|
| 164 |
-
"rewards/accuracies": 0.
|
| 165 |
-
"rewards/chosen": 0.
|
| 166 |
-
"rewards/margins": 0.
|
| 167 |
-
"rewards/rejected": -0.
|
| 168 |
"step": 90,
|
| 169 |
-
"use_label":
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 0.1,
|
| 173 |
"learning_rate": 4.976717112922003e-05,
|
| 174 |
-
"logits/chosen": -2.
|
| 175 |
-
"logits/rejected": -2.
|
| 176 |
-
"logps/chosen": -276.
|
| 177 |
-
"logps/rejected": -
|
| 178 |
-
"loss": 0.
|
| 179 |
-
"pred_label":
|
| 180 |
-
"rewards/accuracies": 0.
|
| 181 |
-
"rewards/chosen": -0.
|
| 182 |
-
"rewards/margins":
|
| 183 |
-
"rewards/rejected": -1.
|
| 184 |
"step": 100,
|
| 185 |
-
"use_label":
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 0.12,
|
| 189 |
"learning_rate": 4.918509895227008e-05,
|
| 190 |
-
"logits/chosen": -2.
|
| 191 |
-
"logits/rejected": -2.
|
| 192 |
-
"logps/chosen": -
|
| 193 |
-
"logps/rejected": -
|
| 194 |
-
"loss": 0.
|
| 195 |
-
"pred_label":
|
| 196 |
-
"rewards/accuracies": 0.
|
| 197 |
-
"rewards/chosen": -0.
|
| 198 |
-
"rewards/margins": 0.
|
| 199 |
-
"rewards/rejected": -
|
| 200 |
"step": 110,
|
| 201 |
-
"use_label":
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.13,
|
| 205 |
"learning_rate": 4.860302677532014e-05,
|
| 206 |
-
"logits/chosen": -2.
|
| 207 |
-
"logits/rejected": -2.
|
| 208 |
-
"logps/chosen": -
|
| 209 |
-
"logps/rejected": -
|
| 210 |
-
"loss": 0.
|
| 211 |
-
"pred_label":
|
| 212 |
-
"rewards/accuracies": 0.
|
| 213 |
-
"rewards/chosen": -0.
|
| 214 |
-
"rewards/margins": 0.
|
| 215 |
-
"rewards/rejected": -
|
| 216 |
"step": 120,
|
| 217 |
-
"use_label":
|
| 218 |
},
|
| 219 |
{
|
| 220 |
"epoch": 0.14,
|
| 221 |
"learning_rate": 4.80209545983702e-05,
|
| 222 |
-
"logits/chosen": -2.
|
| 223 |
-
"logits/rejected": -2.
|
| 224 |
-
"logps/chosen": -308.
|
| 225 |
-
"logps/rejected": -
|
| 226 |
-
"loss": 0.
|
| 227 |
-
"pred_label":
|
| 228 |
-
"rewards/accuracies": 0.
|
| 229 |
-
"rewards/chosen": -0.
|
| 230 |
-
"rewards/margins": 0.
|
| 231 |
-
"rewards/rejected": -0.
|
| 232 |
"step": 130,
|
| 233 |
-
"use_label":
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.15,
|
| 237 |
"learning_rate": 4.743888242142026e-05,
|
| 238 |
-
"logits/chosen": -2.
|
| 239 |
-
"logits/rejected": -2.
|
| 240 |
-
"logps/chosen": -
|
| 241 |
-
"logps/rejected": -
|
| 242 |
-
"loss": 0.
|
| 243 |
-
"pred_label":
|
| 244 |
-
"rewards/accuracies": 0.
|
| 245 |
-
"rewards/chosen": -0.
|
| 246 |
-
"rewards/margins":
|
| 247 |
-
"rewards/rejected": -1.
|
| 248 |
"step": 140,
|
| 249 |
-
"use_label":
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.16,
|
| 253 |
"learning_rate": 4.685681024447032e-05,
|
| 254 |
-
"logits/chosen": -2.
|
| 255 |
-
"logits/rejected": -2.
|
| 256 |
-
"logps/chosen": -
|
| 257 |
-
"logps/rejected": -
|
| 258 |
-
"loss": 0.
|
| 259 |
-
"pred_label":
|
| 260 |
"rewards/accuracies": 0.612500011920929,
|
| 261 |
-
"rewards/chosen": -0.
|
| 262 |
-
"rewards/margins": 0.
|
| 263 |
-
"rewards/rejected": -1.
|
| 264 |
"step": 150,
|
| 265 |
-
"use_label":
|
| 266 |
},
|
| 267 |
{
|
| 268 |
"epoch": 0.17,
|
| 269 |
"learning_rate": 4.6274738067520374e-05,
|
| 270 |
-
"logits/chosen": -2.
|
| 271 |
-
"logits/rejected": -2.
|
| 272 |
-
"logps/chosen": -
|
| 273 |
-
"logps/rejected": -
|
| 274 |
-
"loss": 0.
|
| 275 |
-
"pred_label":
|
| 276 |
-
"rewards/accuracies": 0.
|
| 277 |
-
"rewards/chosen": -0.
|
| 278 |
-
"rewards/margins": 1.
|
| 279 |
-
"rewards/rejected": -1.
|
| 280 |
"step": 160,
|
| 281 |
-
"use_label":
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 0.18,
|
| 285 |
"learning_rate": 4.5692665890570435e-05,
|
| 286 |
-
"logits/chosen": -2.
|
| 287 |
-
"logits/rejected": -2.
|
| 288 |
-
"logps/chosen": -
|
| 289 |
-
"logps/rejected": -
|
| 290 |
-
"loss": 0.
|
| 291 |
-
"pred_label":
|
| 292 |
-
"rewards/accuracies": 0.
|
| 293 |
-
"rewards/chosen": -0.
|
| 294 |
-
"rewards/margins": 1.
|
| 295 |
-
"rewards/rejected": -1.
|
| 296 |
"step": 170,
|
| 297 |
-
"use_label":
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.19,
|
| 301 |
"learning_rate": 4.511059371362049e-05,
|
| 302 |
-
"logits/chosen": -2.
|
| 303 |
-
"logits/rejected": -2.
|
| 304 |
-
"logps/chosen": -
|
| 305 |
-
"logps/rejected": -
|
| 306 |
-
"loss": 0.
|
| 307 |
-
"pred_label":
|
| 308 |
"rewards/accuracies": 0.731249988079071,
|
| 309 |
-
"rewards/chosen": -0.
|
| 310 |
-
"rewards/margins": 1.
|
| 311 |
-
"rewards/rejected": -
|
| 312 |
"step": 180,
|
| 313 |
-
"use_label":
|
| 314 |
},
|
| 315 |
{
|
| 316 |
"epoch": 0.2,
|
| 317 |
"learning_rate": 4.452852153667055e-05,
|
| 318 |
-
"logits/chosen": -2.
|
| 319 |
-
"logits/rejected": -2.
|
| 320 |
-
"logps/chosen": -
|
| 321 |
-
"logps/rejected": -
|
| 322 |
-
"loss": 0.
|
| 323 |
-
"pred_label":
|
| 324 |
-
"rewards/accuracies": 0.
|
| 325 |
-
"rewards/chosen": -0.
|
| 326 |
-
"rewards/margins":
|
| 327 |
-
"rewards/rejected": -
|
| 328 |
"step": 190,
|
| 329 |
-
"use_label":
|
| 330 |
},
|
| 331 |
{
|
| 332 |
"epoch": 0.21,
|
| 333 |
"learning_rate": 4.394644935972061e-05,
|
| 334 |
-
"logits/chosen": -2.
|
| 335 |
-
"logits/rejected": -2.
|
| 336 |
-
"logps/chosen": -
|
| 337 |
-
"logps/rejected": -
|
| 338 |
-
"loss": 0.
|
| 339 |
-
"pred_label":
|
| 340 |
-
"rewards/accuracies": 0.
|
| 341 |
-
"rewards/chosen": -
|
| 342 |
-
"rewards/margins": 1.
|
| 343 |
-
"rewards/rejected": -
|
| 344 |
"step": 200,
|
| 345 |
-
"use_label":
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22,
|
| 349 |
"learning_rate": 4.336437718277067e-05,
|
| 350 |
-
"logits/chosen": -2.
|
| 351 |
-
"logits/rejected": -2.
|
| 352 |
-
"logps/chosen": -
|
| 353 |
-
"logps/rejected": -
|
| 354 |
-
"loss": 0.
|
| 355 |
-
"pred_label":
|
| 356 |
-
"rewards/accuracies": 0.
|
| 357 |
-
"rewards/chosen": -0.
|
| 358 |
-
"rewards/margins": 1.
|
| 359 |
-
"rewards/rejected": -
|
| 360 |
"step": 210,
|
| 361 |
-
"use_label":
|
| 362 |
},
|
| 363 |
{
|
| 364 |
"epoch": 0.23,
|
| 365 |
"learning_rate": 4.278230500582072e-05,
|
| 366 |
-
"logits/chosen": -2.
|
| 367 |
-
"logits/rejected": -2.
|
| 368 |
-
"logps/chosen": -
|
| 369 |
-
"logps/rejected": -
|
| 370 |
-
"loss": 0.
|
| 371 |
-
"pred_label":
|
| 372 |
"rewards/accuracies": 0.75,
|
| 373 |
-
"rewards/chosen": -0.
|
| 374 |
-
"rewards/margins": 1.
|
| 375 |
-
"rewards/rejected": -2.
|
| 376 |
"step": 220,
|
| 377 |
-
"use_label":
|
| 378 |
},
|
| 379 |
{
|
| 380 |
"epoch": 0.24,
|
| 381 |
"learning_rate": 4.220023282887078e-05,
|
| 382 |
-
"logits/chosen": -2.
|
| 383 |
-
"logits/rejected": -2.
|
| 384 |
-
"logps/chosen": -
|
| 385 |
-
"logps/rejected": -
|
| 386 |
-
"loss": 0.
|
| 387 |
-
"pred_label":
|
| 388 |
-
"rewards/accuracies": 0.
|
| 389 |
-
"rewards/chosen": -
|
| 390 |
-
"rewards/margins": 1.
|
| 391 |
-
"rewards/rejected": -
|
| 392 |
"step": 230,
|
| 393 |
-
"use_label":
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 0.25,
|
| 397 |
"learning_rate": 4.161816065192084e-05,
|
| 398 |
-
"logits/chosen": -2.
|
| 399 |
-
"logits/rejected": -2.
|
| 400 |
-
"logps/chosen": -
|
| 401 |
-
"logps/rejected": -
|
| 402 |
-
"loss": 0.
|
| 403 |
-
"pred_label":
|
| 404 |
-
"rewards/accuracies": 0.
|
| 405 |
-
"rewards/chosen": -
|
| 406 |
-
"rewards/margins":
|
| 407 |
-
"rewards/rejected": -
|
| 408 |
"step": 240,
|
| 409 |
-
"use_label":
|
| 410 |
},
|
| 411 |
{
|
| 412 |
"epoch": 0.26,
|
| 413 |
"learning_rate": 4.10360884749709e-05,
|
| 414 |
-
"logits/chosen": -2.
|
| 415 |
-
"logits/rejected": -2.
|
| 416 |
-
"logps/chosen": -
|
| 417 |
-
"logps/rejected": -
|
| 418 |
-
"loss": 0.
|
| 419 |
-
"pred_label":
|
| 420 |
-
"rewards/accuracies": 0.
|
| 421 |
-
"rewards/chosen": -
|
| 422 |
-
"rewards/margins":
|
| 423 |
-
"rewards/rejected": -
|
| 424 |
"step": 250,
|
| 425 |
-
"use_label":
|
| 426 |
},
|
| 427 |
{
|
| 428 |
"epoch": 0.27,
|
| 429 |
"learning_rate": 4.045401629802096e-05,
|
| 430 |
-
"logits/chosen": -2.
|
| 431 |
-
"logits/rejected": -2.
|
| 432 |
-
"logps/chosen": -
|
| 433 |
-
"logps/rejected": -
|
| 434 |
-
"loss": 0.
|
| 435 |
-
"pred_label":
|
| 436 |
-
"rewards/accuracies": 0.
|
| 437 |
-
"rewards/chosen": -
|
| 438 |
-
"rewards/margins": 1.
|
| 439 |
-
"rewards/rejected": -
|
| 440 |
"step": 260,
|
| 441 |
-
"use_label":
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 0.28,
|
| 445 |
"learning_rate": 3.9871944121071014e-05,
|
| 446 |
-
"logits/chosen": -2.
|
| 447 |
-
"logits/rejected": -2.
|
| 448 |
-
"logps/chosen": -
|
| 449 |
-
"logps/rejected": -
|
| 450 |
-
"loss": 0.
|
| 451 |
-
"pred_label":
|
| 452 |
-
"rewards/accuracies": 0.
|
| 453 |
-
"rewards/chosen": -
|
| 454 |
-
"rewards/margins":
|
| 455 |
-
"rewards/rejected": -
|
| 456 |
"step": 270,
|
| 457 |
-
"use_label":
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.29,
|
| 461 |
"learning_rate": 3.928987194412107e-05,
|
| 462 |
-
"logits/chosen": -2.
|
| 463 |
-
"logits/rejected": -2.
|
| 464 |
-
"logps/chosen": -
|
| 465 |
-
"logps/rejected": -
|
| 466 |
-
"loss": 0.
|
| 467 |
-
"pred_label":
|
| 468 |
-
"rewards/accuracies": 0.
|
| 469 |
-
"rewards/chosen": -
|
| 470 |
-
"rewards/margins":
|
| 471 |
-
"rewards/rejected": -
|
| 472 |
"step": 280,
|
| 473 |
-
"use_label":
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 0.3,
|
| 477 |
"learning_rate": 3.870779976717113e-05,
|
| 478 |
-
"logits/chosen": -2.
|
| 479 |
-
"logits/rejected": -2.
|
| 480 |
-
"logps/chosen": -
|
| 481 |
-
"logps/rejected": -
|
| 482 |
-
"loss": 0.
|
| 483 |
-
"pred_label":
|
| 484 |
-
"rewards/accuracies": 0.
|
| 485 |
-
"rewards/chosen": -
|
| 486 |
-
"rewards/margins": 1.
|
| 487 |
-
"rewards/rejected": -
|
| 488 |
"step": 290,
|
| 489 |
-
"use_label":
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.31,
|
| 493 |
"learning_rate": 3.812572759022119e-05,
|
| 494 |
-
"logits/chosen": -2.
|
| 495 |
-
"logits/rejected": -2.
|
| 496 |
-
"logps/chosen": -
|
| 497 |
-
"logps/rejected": -
|
| 498 |
-
"loss": 0.
|
| 499 |
-
"pred_label":
|
| 500 |
-
"rewards/accuracies": 0.
|
| 501 |
-
"rewards/chosen": -
|
| 502 |
-
"rewards/margins":
|
| 503 |
-
"rewards/rejected": -
|
| 504 |
"step": 300,
|
| 505 |
-
"use_label":
|
| 506 |
},
|
| 507 |
{
|
| 508 |
"epoch": 0.32,
|
| 509 |
"learning_rate": 3.7543655413271246e-05,
|
| 510 |
-
"logits/chosen": -2.
|
| 511 |
-
"logits/rejected": -2.
|
| 512 |
-
"logps/chosen": -
|
| 513 |
-
"logps/rejected": -
|
| 514 |
-
"loss": 0.
|
| 515 |
-
"pred_label":
|
| 516 |
-
"rewards/accuracies": 0.
|
| 517 |
-
"rewards/chosen": -
|
| 518 |
-
"rewards/margins":
|
| 519 |
-
"rewards/rejected": -
|
| 520 |
"step": 310,
|
| 521 |
-
"use_label":
|
| 522 |
},
|
| 523 |
{
|
| 524 |
"epoch": 0.33,
|
| 525 |
"learning_rate": 3.696158323632131e-05,
|
| 526 |
-
"logits/chosen": -2.
|
| 527 |
-
"logits/rejected": -2.
|
| 528 |
-
"logps/chosen": -
|
| 529 |
-
"logps/rejected": -
|
| 530 |
-
"loss": 0.
|
| 531 |
-
"pred_label":
|
| 532 |
-
"rewards/accuracies": 0.
|
| 533 |
-
"rewards/chosen": -
|
| 534 |
-
"rewards/margins":
|
| 535 |
-
"rewards/rejected": -
|
| 536 |
"step": 320,
|
| 537 |
-
"use_label":
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 0.35,
|
| 541 |
"learning_rate": 3.637951105937136e-05,
|
| 542 |
-
"logits/chosen": -2.
|
| 543 |
-
"logits/rejected": -2.
|
| 544 |
-
"logps/chosen": -
|
| 545 |
-
"logps/rejected": -
|
| 546 |
-
"loss": 0.
|
| 547 |
-
"pred_label":
|
| 548 |
-
"rewards/accuracies": 0.
|
| 549 |
-
"rewards/chosen": -
|
| 550 |
-
"rewards/margins":
|
| 551 |
-
"rewards/rejected": -
|
| 552 |
"step": 330,
|
| 553 |
-
"use_label":
|
| 554 |
},
|
| 555 |
{
|
| 556 |
"epoch": 0.36,
|
| 557 |
"learning_rate": 3.579743888242142e-05,
|
| 558 |
-
"logits/chosen": -2.
|
| 559 |
-
"logits/rejected": -2.
|
| 560 |
-
"logps/chosen": -
|
| 561 |
-
"logps/rejected": -
|
| 562 |
-
"loss": 0.
|
| 563 |
-
"pred_label":
|
| 564 |
-
"rewards/accuracies": 0.
|
| 565 |
-
"rewards/chosen": -
|
| 566 |
-
"rewards/margins":
|
| 567 |
-
"rewards/rejected": -
|
| 568 |
"step": 340,
|
| 569 |
-
"use_label":
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.37,
|
| 573 |
"learning_rate": 3.5215366705471484e-05,
|
| 574 |
-
"logits/chosen": -2.
|
| 575 |
-
"logits/rejected": -2.
|
| 576 |
-
"logps/chosen": -
|
| 577 |
-
"logps/rejected": -
|
| 578 |
-
"loss": 0.
|
| 579 |
-
"pred_label":
|
| 580 |
-
"rewards/accuracies": 0.
|
| 581 |
-
"rewards/chosen": -
|
| 582 |
-
"rewards/margins":
|
| 583 |
-
"rewards/rejected": -
|
| 584 |
"step": 350,
|
| 585 |
-
"use_label":
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 0.38,
|
| 589 |
"learning_rate": 3.463329452852154e-05,
|
| 590 |
-
"logits/chosen": -2.
|
| 591 |
-
"logits/rejected": -2.
|
| 592 |
-
"logps/chosen": -
|
| 593 |
-
"logps/rejected": -
|
| 594 |
-
"loss": 0.
|
| 595 |
-
"pred_label":
|
| 596 |
-
"rewards/accuracies": 0.
|
| 597 |
-
"rewards/chosen": -
|
| 598 |
-
"rewards/margins":
|
| 599 |
-
"rewards/rejected": -
|
| 600 |
"step": 360,
|
| 601 |
-
"use_label":
|
| 602 |
},
|
| 603 |
{
|
| 604 |
"epoch": 0.39,
|
| 605 |
"learning_rate": 3.40512223515716e-05,
|
| 606 |
-
"logits/chosen": -2.
|
| 607 |
-
"logits/rejected": -2.
|
| 608 |
-
"logps/chosen": -
|
| 609 |
-
"logps/rejected": -
|
| 610 |
-
"loss": 0.
|
| 611 |
-
"pred_label":
|
| 612 |
-
"rewards/accuracies": 0.
|
| 613 |
-
"rewards/chosen": -
|
| 614 |
-
"rewards/margins":
|
| 615 |
-
"rewards/rejected": -
|
| 616 |
"step": 370,
|
| 617 |
-
"use_label":
|
| 618 |
},
|
| 619 |
{
|
| 620 |
"epoch": 0.4,
|
| 621 |
"learning_rate": 3.3469150174621654e-05,
|
| 622 |
-
"logits/chosen": -2.
|
| 623 |
-
"logits/rejected": -2.
|
| 624 |
-
"logps/chosen": -
|
| 625 |
-
"logps/rejected": -
|
| 626 |
-
"loss": 0.
|
| 627 |
-
"pred_label":
|
| 628 |
-
"rewards/accuracies": 0.
|
| 629 |
-
"rewards/chosen": -
|
| 630 |
-
"rewards/margins":
|
| 631 |
-
"rewards/rejected": -
|
| 632 |
"step": 380,
|
| 633 |
-
"use_label":
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 0.41,
|
| 637 |
"learning_rate": 3.288707799767171e-05,
|
| 638 |
-
"logits/chosen": -2.
|
| 639 |
-
"logits/rejected": -2.
|
| 640 |
-
"logps/chosen": -
|
| 641 |
-
"logps/rejected": -
|
| 642 |
-
"loss": 0.
|
| 643 |
-
"pred_label":
|
| 644 |
-
"rewards/accuracies": 0.
|
| 645 |
-
"rewards/chosen": -
|
| 646 |
-
"rewards/margins":
|
| 647 |
-
"rewards/rejected": -
|
| 648 |
"step": 390,
|
| 649 |
-
"use_label":
|
| 650 |
},
|
| 651 |
{
|
| 652 |
"epoch": 0.42,
|
| 653 |
"learning_rate": 3.2305005820721776e-05,
|
| 654 |
-
"logits/chosen": -2.
|
| 655 |
-
"logits/rejected": -2.
|
| 656 |
-
"logps/chosen": -
|
| 657 |
-
"logps/rejected": -
|
| 658 |
-
"loss": 0.
|
| 659 |
-
"pred_label":
|
| 660 |
"rewards/accuracies": 0.71875,
|
| 661 |
-
"rewards/chosen": -
|
| 662 |
-
"rewards/margins":
|
| 663 |
-
"rewards/rejected": -
|
| 664 |
"step": 400,
|
| 665 |
-
"use_label":
|
| 666 |
},
|
| 667 |
{
|
| 668 |
"epoch": 0.43,
|
| 669 |
"learning_rate": 3.172293364377183e-05,
|
| 670 |
-
"logits/chosen": -2.
|
| 671 |
-
"logits/rejected": -2.
|
| 672 |
-
"logps/chosen": -
|
| 673 |
-
"logps/rejected": -
|
| 674 |
-
"loss": 0.
|
| 675 |
-
"pred_label":
|
| 676 |
-
"rewards/accuracies": 0.
|
| 677 |
-
"rewards/chosen": -
|
| 678 |
-
"rewards/margins": 1.
|
| 679 |
-
"rewards/rejected": -
|
| 680 |
"step": 410,
|
| 681 |
-
"use_label":
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44,
|
| 685 |
"learning_rate": 3.1140861466821885e-05,
|
| 686 |
-
"logits/chosen": -2.
|
| 687 |
-
"logits/rejected": -2.
|
| 688 |
-
"logps/chosen": -
|
| 689 |
-
"logps/rejected": -
|
| 690 |
-
"loss": 0.
|
| 691 |
-
"pred_label":
|
| 692 |
"rewards/accuracies": 0.6937500238418579,
|
| 693 |
-
"rewards/chosen": -
|
| 694 |
-
"rewards/margins": 1.
|
| 695 |
-
"rewards/rejected": -
|
| 696 |
"step": 420,
|
| 697 |
-
"use_label":
|
| 698 |
},
|
| 699 |
{
|
| 700 |
"epoch": 0.45,
|
| 701 |
"learning_rate": 3.055878928987195e-05,
|
| 702 |
-
"logits/chosen": -2.
|
| 703 |
-
"logits/rejected": -2.
|
| 704 |
-
"logps/chosen": -
|
| 705 |
-
"logps/rejected": -
|
| 706 |
-
"loss": 0.
|
| 707 |
-
"pred_label":
|
| 708 |
-
"rewards/accuracies": 0.
|
| 709 |
-
"rewards/chosen": -
|
| 710 |
-
"rewards/margins": 1.
|
| 711 |
-
"rewards/rejected": -2.
|
| 712 |
"step": 430,
|
| 713 |
-
"use_label":
|
| 714 |
},
|
| 715 |
{
|
| 716 |
"epoch": 0.46,
|
| 717 |
"learning_rate": 2.9976717112922005e-05,
|
| 718 |
-
"logits/chosen": -2.
|
| 719 |
-
"logits/rejected": -2.
|
| 720 |
-
"logps/chosen": -
|
| 721 |
-
"logps/rejected": -
|
| 722 |
-
"loss": 0.
|
| 723 |
-
"pred_label":
|
| 724 |
-
"rewards/accuracies": 0.
|
| 725 |
-
"rewards/chosen": -
|
| 726 |
-
"rewards/margins": 1.
|
| 727 |
-
"rewards/rejected": -
|
| 728 |
"step": 440,
|
| 729 |
-
"use_label":
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.47,
|
| 733 |
"learning_rate": 2.939464493597206e-05,
|
| 734 |
-
"logits/chosen": -2.
|
| 735 |
-
"logits/rejected": -2.
|
| 736 |
-
"logps/chosen": -
|
| 737 |
-
"logps/rejected": -
|
| 738 |
-
"loss": 0.
|
| 739 |
-
"pred_label":
|
| 740 |
-
"rewards/accuracies": 0.
|
| 741 |
-
"rewards/chosen": -
|
| 742 |
-
"rewards/margins":
|
| 743 |
-
"rewards/rejected": -
|
| 744 |
"step": 450,
|
| 745 |
-
"use_label":
|
| 746 |
},
|
| 747 |
{
|
| 748 |
"epoch": 0.48,
|
| 749 |
"learning_rate": 2.881257275902212e-05,
|
| 750 |
-
"logits/chosen": -2.
|
| 751 |
-
"logits/rejected": -2.
|
| 752 |
-
"logps/chosen": -
|
| 753 |
-
"logps/rejected": -
|
| 754 |
-
"loss": 0.
|
| 755 |
-
"pred_label":
|
| 756 |
-
"rewards/accuracies": 0.
|
| 757 |
-
"rewards/chosen": -
|
| 758 |
-
"rewards/margins": 1.
|
| 759 |
-
"rewards/rejected": -
|
| 760 |
"step": 460,
|
| 761 |
-
"use_label":
|
| 762 |
},
|
| 763 |
{
|
| 764 |
"epoch": 0.49,
|
| 765 |
"learning_rate": 2.8230500582072178e-05,
|
| 766 |
-
"logits/chosen": -2.
|
| 767 |
-
"logits/rejected": -2.
|
| 768 |
-
"logps/chosen": -
|
| 769 |
-
"logps/rejected": -
|
| 770 |
-
"loss": 0.
|
| 771 |
-
"pred_label":
|
| 772 |
-
"rewards/accuracies": 0.
|
| 773 |
-
"rewards/chosen": -
|
| 774 |
-
"rewards/margins":
|
| 775 |
-
"rewards/rejected": -
|
| 776 |
"step": 470,
|
| 777 |
-
"use_label":
|
| 778 |
},
|
| 779 |
{
|
| 780 |
"epoch": 0.5,
|
| 781 |
"learning_rate": 2.7648428405122233e-05,
|
| 782 |
-
"logits/chosen": -2.
|
| 783 |
-
"logits/rejected": -2.
|
| 784 |
-
"logps/chosen": -
|
| 785 |
-
"logps/rejected": -
|
| 786 |
-
"loss": 0.
|
| 787 |
-
"pred_label":
|
| 788 |
-
"rewards/accuracies": 0.
|
| 789 |
-
"rewards/chosen": -
|
| 790 |
-
"rewards/margins":
|
| 791 |
-
"rewards/rejected": -
|
| 792 |
"step": 480,
|
| 793 |
-
"use_label":
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.51,
|
| 797 |
"learning_rate": 2.7066356228172297e-05,
|
| 798 |
-
"logits/chosen": -2.
|
| 799 |
-
"logits/rejected": -2.
|
| 800 |
-
"logps/chosen": -
|
| 801 |
-
"logps/rejected": -
|
| 802 |
-
"loss": 0.
|
| 803 |
-
"pred_label":
|
| 804 |
-
"rewards/accuracies": 0.
|
| 805 |
-
"rewards/chosen": -
|
| 806 |
-
"rewards/margins":
|
| 807 |
-
"rewards/rejected": -
|
| 808 |
"step": 490,
|
| 809 |
-
"use_label":
|
| 810 |
},
|
| 811 |
{
|
| 812 |
"epoch": 0.52,
|
| 813 |
"learning_rate": 2.6484284051222352e-05,
|
| 814 |
-
"logits/chosen": -2.
|
| 815 |
-
"logits/rejected": -2.
|
| 816 |
-
"logps/chosen": -
|
| 817 |
-
"logps/rejected": -
|
| 818 |
-
"loss": 0.
|
| 819 |
-
"pred_label":
|
| 820 |
-
"rewards/accuracies": 0.
|
| 821 |
-
"rewards/chosen": -
|
| 822 |
-
"rewards/margins": 2.
|
| 823 |
-
"rewards/rejected": -
|
| 824 |
"step": 500,
|
| 825 |
-
"use_label":
|
| 826 |
},
|
| 827 |
{
|
| 828 |
"epoch": 0.53,
|
| 829 |
"learning_rate": 2.590221187427241e-05,
|
| 830 |
-
"logits/chosen": -2.
|
| 831 |
-
"logits/rejected": -2.
|
| 832 |
-
"logps/chosen": -
|
| 833 |
-
"logps/rejected": -
|
| 834 |
-
"loss": 0.
|
| 835 |
-
"pred_label":
|
| 836 |
"rewards/accuracies": 0.7749999761581421,
|
| 837 |
-
"rewards/chosen": -
|
| 838 |
-
"rewards/margins":
|
| 839 |
-
"rewards/rejected": -
|
| 840 |
"step": 510,
|
| 841 |
-
"use_label":
|
| 842 |
},
|
| 843 |
{
|
| 844 |
"epoch": 0.54,
|
| 845 |
"learning_rate": 2.532013969732247e-05,
|
| 846 |
-
"logits/chosen": -2.
|
| 847 |
-
"logits/rejected": -2.
|
| 848 |
-
"logps/chosen": -
|
| 849 |
-
"logps/rejected": -
|
| 850 |
-
"loss": 0.
|
| 851 |
-
"pred_label":
|
| 852 |
-
"rewards/accuracies": 0.
|
| 853 |
-
"rewards/chosen": -
|
| 854 |
-
"rewards/margins":
|
| 855 |
-
"rewards/rejected": -
|
| 856 |
"step": 520,
|
| 857 |
-
"use_label":
|
| 858 |
},
|
| 859 |
{
|
| 860 |
"epoch": 0.55,
|
| 861 |
"learning_rate": 2.4738067520372525e-05,
|
| 862 |
-
"logits/chosen": -2.
|
| 863 |
-
"logits/rejected": -2.
|
| 864 |
-
"logps/chosen": -
|
| 865 |
-
"logps/rejected": -
|
| 866 |
-
"loss": 0.
|
| 867 |
-
"pred_label":
|
| 868 |
-
"rewards/accuracies": 0.
|
| 869 |
-
"rewards/chosen": -
|
| 870 |
-
"rewards/margins": 2.
|
| 871 |
-
"rewards/rejected": -
|
| 872 |
"step": 530,
|
| 873 |
-
"use_label":
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 0.57,
|
| 877 |
"learning_rate": 2.4155995343422587e-05,
|
| 878 |
-
"logits/chosen": -2.
|
| 879 |
-
"logits/rejected": -2.
|
| 880 |
-
"logps/chosen": -
|
| 881 |
-
"logps/rejected": -
|
| 882 |
-
"loss": 0.
|
| 883 |
-
"pred_label":
|
| 884 |
-
"rewards/accuracies": 0.
|
| 885 |
-
"rewards/chosen": -
|
| 886 |
-
"rewards/margins":
|
| 887 |
-
"rewards/rejected": -
|
| 888 |
"step": 540,
|
| 889 |
-
"use_label":
|
| 890 |
},
|
| 891 |
{
|
| 892 |
"epoch": 0.58,
|
| 893 |
"learning_rate": 2.3573923166472644e-05,
|
| 894 |
-
"logits/chosen": -2.
|
| 895 |
-
"logits/rejected": -2.
|
| 896 |
-
"logps/chosen": -
|
| 897 |
-
"logps/rejected": -
|
| 898 |
-
"loss": 0.
|
| 899 |
-
"pred_label":
|
| 900 |
-
"rewards/accuracies": 0.
|
| 901 |
-
"rewards/chosen": -
|
| 902 |
-
"rewards/margins":
|
| 903 |
-
"rewards/rejected": -
|
| 904 |
"step": 550,
|
| 905 |
-
"use_label":
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.59,
|
| 909 |
"learning_rate": 2.2991850989522702e-05,
|
| 910 |
-
"logits/chosen": -2.
|
| 911 |
-
"logits/rejected": -2.
|
| 912 |
-
"logps/chosen": -
|
| 913 |
-
"logps/rejected": -
|
| 914 |
-
"loss": 0.
|
| 915 |
-
"pred_label":
|
| 916 |
-
"rewards/accuracies": 0.
|
| 917 |
-
"rewards/chosen": -
|
| 918 |
-
"rewards/margins":
|
| 919 |
-
"rewards/rejected": -
|
| 920 |
"step": 560,
|
| 921 |
-
"use_label":
|
| 922 |
},
|
| 923 |
{
|
| 924 |
"epoch": 0.6,
|
| 925 |
"learning_rate": 2.240977881257276e-05,
|
| 926 |
-
"logits/chosen": -2.
|
| 927 |
-
"logits/rejected": -2.
|
| 928 |
-
"logps/chosen": -
|
| 929 |
-
"logps/rejected": -
|
| 930 |
-
"loss": 0.
|
| 931 |
-
"pred_label":
|
| 932 |
-
"rewards/accuracies": 0.
|
| 933 |
-
"rewards/chosen": -
|
| 934 |
-
"rewards/margins":
|
| 935 |
-
"rewards/rejected": -
|
| 936 |
"step": 570,
|
| 937 |
-
"use_label":
|
| 938 |
},
|
| 939 |
{
|
| 940 |
"epoch": 0.61,
|
| 941 |
"learning_rate": 2.1827706635622818e-05,
|
| 942 |
-
"logits/chosen": -2.
|
| 943 |
-
"logits/rejected": -2.
|
| 944 |
-
"logps/chosen": -
|
| 945 |
-
"logps/rejected": -
|
| 946 |
-
"loss": 0.
|
| 947 |
-
"pred_label":
|
| 948 |
-
"rewards/accuracies": 0.
|
| 949 |
-
"rewards/chosen": -
|
| 950 |
-
"rewards/margins":
|
| 951 |
-
"rewards/rejected": -
|
| 952 |
"step": 580,
|
| 953 |
-
"use_label":
|
| 954 |
},
|
| 955 |
{
|
| 956 |
"epoch": 0.62,
|
| 957 |
"learning_rate": 2.124563445867288e-05,
|
| 958 |
-
"logits/chosen": -2.
|
| 959 |
-
"logits/rejected": -2.
|
| 960 |
-
"logps/chosen": -
|
| 961 |
-
"logps/rejected": -
|
| 962 |
-
"loss": 0.
|
| 963 |
-
"pred_label":
|
| 964 |
"rewards/accuracies": 0.6875,
|
| 965 |
-
"rewards/chosen": -
|
| 966 |
-
"rewards/margins":
|
| 967 |
-
"rewards/rejected": -
|
| 968 |
"step": 590,
|
| 969 |
-
"use_label":
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.63,
|
| 973 |
"learning_rate": 2.0663562281722934e-05,
|
| 974 |
-
"logits/chosen": -2.
|
| 975 |
-
"logits/rejected": -2.
|
| 976 |
-
"logps/chosen": -
|
| 977 |
-
"logps/rejected": -
|
| 978 |
-
"loss": 0.
|
| 979 |
-
"pred_label":
|
| 980 |
-
"rewards/accuracies": 0.
|
| 981 |
-
"rewards/chosen": -
|
| 982 |
-
"rewards/margins":
|
| 983 |
-
"rewards/rejected": -
|
| 984 |
"step": 600,
|
| 985 |
-
"use_label":
|
| 986 |
},
|
| 987 |
{
|
| 988 |
"epoch": 0.64,
|
| 989 |
"learning_rate": 2.0081490104772992e-05,
|
| 990 |
-
"logits/chosen": -2.
|
| 991 |
-
"logits/rejected": -2.
|
| 992 |
-
"logps/chosen": -
|
| 993 |
-
"logps/rejected": -
|
| 994 |
-
"loss": 0.
|
| 995 |
-
"pred_label":
|
| 996 |
-
"rewards/accuracies": 0.
|
| 997 |
-
"rewards/chosen": -
|
| 998 |
-
"rewards/margins":
|
| 999 |
-
"rewards/rejected": -
|
| 1000 |
"step": 610,
|
| 1001 |
-
"use_label":
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
"epoch": 0.65,
|
| 1005 |
"learning_rate": 1.9499417927823053e-05,
|
| 1006 |
-
"logits/chosen": -2.
|
| 1007 |
-
"logits/rejected": -2.
|
| 1008 |
-
"logps/chosen": -
|
| 1009 |
-
"logps/rejected": -
|
| 1010 |
-
"loss": 0.
|
| 1011 |
-
"pred_label":
|
| 1012 |
-
"rewards/accuracies": 0.
|
| 1013 |
-
"rewards/chosen": -
|
| 1014 |
-
"rewards/margins": 2.
|
| 1015 |
-
"rewards/rejected": -
|
| 1016 |
"step": 620,
|
| 1017 |
-
"use_label":
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.66,
|
| 1021 |
"learning_rate": 1.8917345750873107e-05,
|
| 1022 |
-
"logits/chosen": -2.
|
| 1023 |
-
"logits/rejected": -2.
|
| 1024 |
-
"logps/chosen": -
|
| 1025 |
-
"logps/rejected": -
|
| 1026 |
-
"loss": 0.
|
| 1027 |
-
"pred_label":
|
| 1028 |
"rewards/accuracies": 0.7749999761581421,
|
| 1029 |
-
"rewards/chosen": -
|
| 1030 |
-
"rewards/margins":
|
| 1031 |
-
"rewards/rejected": -
|
| 1032 |
"step": 630,
|
| 1033 |
-
"use_label":
|
| 1034 |
},
|
| 1035 |
{
|
| 1036 |
"epoch": 0.67,
|
| 1037 |
"learning_rate": 1.833527357392317e-05,
|
| 1038 |
-
"logits/chosen": -2.
|
| 1039 |
-
"logits/rejected": -2.
|
| 1040 |
-
"logps/chosen": -
|
| 1041 |
-
"logps/rejected": -
|
| 1042 |
-
"loss": 0.
|
| 1043 |
-
"pred_label":
|
| 1044 |
-
"rewards/accuracies": 0.
|
| 1045 |
-
"rewards/chosen": -
|
| 1046 |
-
"rewards/margins":
|
| 1047 |
-
"rewards/rejected": -
|
| 1048 |
"step": 640,
|
| 1049 |
-
"use_label":
|
| 1050 |
},
|
| 1051 |
{
|
| 1052 |
"epoch": 0.68,
|
| 1053 |
"learning_rate": 1.7753201396973227e-05,
|
| 1054 |
-
"logits/chosen": -2.
|
| 1055 |
-
"logits/rejected": -2.
|
| 1056 |
-
"logps/chosen": -
|
| 1057 |
-
"logps/rejected": -
|
| 1058 |
-
"loss": 0.
|
| 1059 |
-
"pred_label":
|
| 1060 |
-
"rewards/accuracies": 0.
|
| 1061 |
-
"rewards/chosen": -
|
| 1062 |
-
"rewards/margins": 2.
|
| 1063 |
-
"rewards/rejected": -
|
| 1064 |
"step": 650,
|
| 1065 |
-
"use_label":
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
"epoch": 0.69,
|
| 1069 |
"learning_rate": 1.717112922002328e-05,
|
| 1070 |
-
"logits/chosen": -2.
|
| 1071 |
-
"logits/rejected": -2.
|
| 1072 |
-
"logps/chosen": -
|
| 1073 |
-
"logps/rejected": -
|
| 1074 |
-
"loss": 0.
|
| 1075 |
-
"pred_label":
|
| 1076 |
-
"rewards/accuracies": 0.
|
| 1077 |
-
"rewards/chosen": -
|
| 1078 |
-
"rewards/margins": 1.
|
| 1079 |
-
"rewards/rejected": -
|
| 1080 |
"step": 660,
|
| 1081 |
-
"use_label":
|
| 1082 |
},
|
| 1083 |
{
|
| 1084 |
"epoch": 0.7,
|
| 1085 |
"learning_rate": 1.6589057043073342e-05,
|
| 1086 |
-
"logits/chosen": -2.
|
| 1087 |
-
"logits/rejected": -2.
|
| 1088 |
-
"logps/chosen": -
|
| 1089 |
-
"logps/rejected": -
|
| 1090 |
-
"loss": 0.
|
| 1091 |
-
"pred_label":
|
| 1092 |
-
"rewards/accuracies": 0.
|
| 1093 |
-
"rewards/chosen": -
|
| 1094 |
-
"rewards/margins":
|
| 1095 |
-
"rewards/rejected": -
|
| 1096 |
"step": 670,
|
| 1097 |
-
"use_label":
|
| 1098 |
},
|
| 1099 |
{
|
| 1100 |
"epoch": 0.71,
|
| 1101 |
"learning_rate": 1.60069848661234e-05,
|
| 1102 |
-
"logits/chosen": -2.
|
| 1103 |
-
"logits/rejected": -2.
|
| 1104 |
-
"logps/chosen": -
|
| 1105 |
-
"logps/rejected": -
|
| 1106 |
-
"loss": 0.
|
| 1107 |
-
"pred_label":
|
| 1108 |
-
"rewards/accuracies": 0.
|
| 1109 |
-
"rewards/chosen": -
|
| 1110 |
-
"rewards/margins": 1.
|
| 1111 |
-
"rewards/rejected": -
|
| 1112 |
"step": 680,
|
| 1113 |
-
"use_label":
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
"epoch": 0.72,
|
| 1117 |
"learning_rate": 1.5424912689173458e-05,
|
| 1118 |
-
"logits/chosen": -2.
|
| 1119 |
-
"logits/rejected": -2.
|
| 1120 |
-
"logps/chosen": -
|
| 1121 |
-
"logps/rejected": -
|
| 1122 |
-
"loss": 0.
|
| 1123 |
-
"pred_label":
|
| 1124 |
"rewards/accuracies": 0.7749999761581421,
|
| 1125 |
-
"rewards/chosen": -
|
| 1126 |
-
"rewards/margins":
|
| 1127 |
-
"rewards/rejected": -
|
| 1128 |
"step": 690,
|
| 1129 |
-
"use_label":
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.73,
|
| 1133 |
"learning_rate": 1.4842840512223516e-05,
|
| 1134 |
-
"logits/chosen": -2.
|
| 1135 |
-
"logits/rejected": -2.
|
| 1136 |
-
"logps/chosen": -
|
| 1137 |
-
"logps/rejected": -
|
| 1138 |
-
"loss": 0.
|
| 1139 |
-
"pred_label":
|
| 1140 |
-
"rewards/accuracies": 0.
|
| 1141 |
-
"rewards/chosen": -1.
|
| 1142 |
-
"rewards/margins":
|
| 1143 |
-
"rewards/rejected": -
|
| 1144 |
"step": 700,
|
| 1145 |
-
"use_label":
|
| 1146 |
},
|
| 1147 |
{
|
| 1148 |
"epoch": 0.74,
|
| 1149 |
"learning_rate": 1.4260768335273575e-05,
|
| 1150 |
-
"logits/chosen": -2.
|
| 1151 |
-
"logits/rejected": -2.
|
| 1152 |
-
"logps/chosen": -
|
| 1153 |
-
"logps/rejected": -
|
| 1154 |
-
"loss": 0.
|
| 1155 |
-
"pred_label":
|
| 1156 |
-
"rewards/accuracies": 0.
|
| 1157 |
-
"rewards/chosen": -
|
| 1158 |
-
"rewards/margins": 2.
|
| 1159 |
-
"rewards/rejected": -
|
| 1160 |
"step": 710,
|
| 1161 |
-
"use_label":
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
"epoch": 0.75,
|
| 1165 |
"learning_rate": 1.3678696158323633e-05,
|
| 1166 |
-
"logits/chosen": -2.
|
| 1167 |
-
"logits/rejected": -2.
|
| 1168 |
-
"logps/chosen": -
|
| 1169 |
-
"logps/rejected": -
|
| 1170 |
-
"loss": 0.
|
| 1171 |
-
"pred_label":
|
| 1172 |
"rewards/accuracies": 0.78125,
|
| 1173 |
-
"rewards/chosen": -1.
|
| 1174 |
-
"rewards/margins":
|
| 1175 |
-
"rewards/rejected": -
|
| 1176 |
"step": 720,
|
| 1177 |
-
"use_label":
|
| 1178 |
},
|
| 1179 |
{
|
| 1180 |
"epoch": 0.76,
|
| 1181 |
"learning_rate": 1.309662398137369e-05,
|
| 1182 |
-
"logits/chosen": -2.
|
| 1183 |
-
"logits/rejected": -2.
|
| 1184 |
-
"logps/chosen": -
|
| 1185 |
-
"logps/rejected": -
|
| 1186 |
-
"loss": 0.
|
| 1187 |
-
"pred_label":
|
| 1188 |
-
"rewards/accuracies": 0.
|
| 1189 |
-
"rewards/chosen": -1.
|
| 1190 |
-
"rewards/margins":
|
| 1191 |
-
"rewards/rejected": -
|
| 1192 |
"step": 730,
|
| 1193 |
-
"use_label":
|
| 1194 |
},
|
| 1195 |
{
|
| 1196 |
"epoch": 0.77,
|
| 1197 |
"learning_rate": 1.2514551804423749e-05,
|
| 1198 |
-
"logits/chosen": -2.
|
| 1199 |
-
"logits/rejected": -2.
|
| 1200 |
-
"logps/chosen": -
|
| 1201 |
-
"logps/rejected": -
|
| 1202 |
-
"loss": 0.
|
| 1203 |
-
"pred_label":
|
| 1204 |
-
"rewards/accuracies": 0.
|
| 1205 |
-
"rewards/chosen": -
|
| 1206 |
-
"rewards/margins":
|
| 1207 |
-
"rewards/rejected": -
|
| 1208 |
"step": 740,
|
| 1209 |
-
"use_label":
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.79,
|
| 1213 |
"learning_rate": 1.1932479627473807e-05,
|
| 1214 |
-
"logits/chosen": -2.
|
| 1215 |
-
"logits/rejected": -2.
|
| 1216 |
-
"logps/chosen": -
|
| 1217 |
-
"logps/rejected": -
|
| 1218 |
-
"loss": 0.
|
| 1219 |
-
"pred_label":
|
| 1220 |
-
"rewards/accuracies": 0.
|
| 1221 |
-
"rewards/chosen": -
|
| 1222 |
-
"rewards/margins":
|
| 1223 |
-
"rewards/rejected": -
|
| 1224 |
"step": 750,
|
| 1225 |
-
"use_label":
|
| 1226 |
},
|
| 1227 |
{
|
| 1228 |
"epoch": 0.8,
|
| 1229 |
"learning_rate": 1.1350407450523866e-05,
|
| 1230 |
-
"logits/chosen": -2.
|
| 1231 |
-
"logits/rejected": -2.
|
| 1232 |
-
"logps/chosen": -
|
| 1233 |
-
"logps/rejected": -
|
| 1234 |
-
"loss": 0.
|
| 1235 |
-
"pred_label":
|
| 1236 |
-
"rewards/accuracies": 0.
|
| 1237 |
-
"rewards/chosen": -1.
|
| 1238 |
-
"rewards/margins":
|
| 1239 |
-
"rewards/rejected": -
|
| 1240 |
"step": 760,
|
| 1241 |
-
"use_label":
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.81,
|
| 1245 |
"learning_rate": 1.0768335273573923e-05,
|
| 1246 |
-
"logits/chosen": -2.
|
| 1247 |
-
"logits/rejected": -2.
|
| 1248 |
-
"logps/chosen": -
|
| 1249 |
-
"logps/rejected": -
|
| 1250 |
-
"loss": 0.
|
| 1251 |
-
"pred_label":
|
| 1252 |
-
"rewards/accuracies": 0.
|
| 1253 |
-
"rewards/chosen": -1.
|
| 1254 |
-
"rewards/margins":
|
| 1255 |
-
"rewards/rejected": -
|
| 1256 |
"step": 770,
|
| 1257 |
-
"use_label":
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
"epoch": 0.82,
|
| 1261 |
"learning_rate": 1.0186263096623982e-05,
|
| 1262 |
-
"logits/chosen": -2.
|
| 1263 |
-
"logits/rejected": -2.
|
| 1264 |
-
"logps/chosen": -
|
| 1265 |
-
"logps/rejected": -
|
| 1266 |
-
"loss": 0.
|
| 1267 |
-
"pred_label":
|
| 1268 |
-
"rewards/accuracies": 0.
|
| 1269 |
-
"rewards/chosen": -
|
| 1270 |
-
"rewards/margins":
|
| 1271 |
-
"rewards/rejected": -
|
| 1272 |
"step": 780,
|
| 1273 |
-
"use_label":
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 0.83,
|
| 1277 |
"learning_rate": 9.60419091967404e-06,
|
| 1278 |
-
"logits/chosen": -2.
|
| 1279 |
-
"logits/rejected": -2.
|
| 1280 |
-
"logps/chosen": -
|
| 1281 |
-
"logps/rejected": -
|
| 1282 |
-
"loss": 0.
|
| 1283 |
-
"pred_label":
|
| 1284 |
-
"rewards/accuracies": 0.
|
| 1285 |
-
"rewards/chosen": -
|
| 1286 |
-
"rewards/margins":
|
| 1287 |
-
"rewards/rejected": -
|
| 1288 |
"step": 790,
|
| 1289 |
-
"use_label":
|
| 1290 |
},
|
| 1291 |
{
|
| 1292 |
"epoch": 0.84,
|
| 1293 |
"learning_rate": 9.022118742724098e-06,
|
| 1294 |
-
"logits/chosen": -2.
|
| 1295 |
-
"logits/rejected": -2.
|
| 1296 |
-
"logps/chosen": -
|
| 1297 |
-
"logps/rejected": -
|
| 1298 |
-
"loss": 0.
|
| 1299 |
-
"pred_label":
|
| 1300 |
-
"rewards/accuracies": 0.
|
| 1301 |
-
"rewards/chosen": -
|
| 1302 |
-
"rewards/margins":
|
| 1303 |
-
"rewards/rejected": -
|
| 1304 |
"step": 800,
|
| 1305 |
-
"use_label":
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
"epoch": 0.85,
|
| 1309 |
"learning_rate": 8.440046565774158e-06,
|
| 1310 |
-
"logits/chosen": -2.
|
| 1311 |
-
"logits/rejected": -2.
|
| 1312 |
-
"logps/chosen": -
|
| 1313 |
-
"logps/rejected": -
|
| 1314 |
-
"loss": 0.
|
| 1315 |
-
"pred_label":
|
| 1316 |
-
"rewards/accuracies": 0.
|
| 1317 |
-
"rewards/chosen": -
|
| 1318 |
-
"rewards/margins":
|
| 1319 |
-
"rewards/rejected": -
|
| 1320 |
"step": 810,
|
| 1321 |
-
"use_label":
|
| 1322 |
},
|
| 1323 |
{
|
| 1324 |
"epoch": 0.86,
|
| 1325 |
"learning_rate": 7.857974388824214e-06,
|
| 1326 |
-
"logits/chosen": -2.
|
| 1327 |
-
"logits/rejected": -2.
|
| 1328 |
-
"logps/chosen": -
|
| 1329 |
-
"logps/rejected": -
|
| 1330 |
-
"loss": 0.
|
| 1331 |
-
"pred_label":
|
| 1332 |
-
"rewards/accuracies": 0.
|
| 1333 |
-
"rewards/chosen": -
|
| 1334 |
-
"rewards/margins":
|
| 1335 |
-
"rewards/rejected": -
|
| 1336 |
"step": 820,
|
| 1337 |
-
"use_label":
|
| 1338 |
},
|
| 1339 |
{
|
| 1340 |
"epoch": 0.87,
|
| 1341 |
"learning_rate": 7.275902211874273e-06,
|
| 1342 |
-
"logits/chosen": -2.
|
| 1343 |
-
"logits/rejected": -2.
|
| 1344 |
-
"logps/chosen": -
|
| 1345 |
-
"logps/rejected": -
|
| 1346 |
-
"loss": 0.
|
| 1347 |
-
"pred_label":
|
| 1348 |
-
"rewards/accuracies": 0.
|
| 1349 |
-
"rewards/chosen": -
|
| 1350 |
-
"rewards/margins":
|
| 1351 |
-
"rewards/rejected": -
|
| 1352 |
"step": 830,
|
| 1353 |
-
"use_label":
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.88,
|
| 1357 |
"learning_rate": 6.693830034924331e-06,
|
| 1358 |
-
"logits/chosen": -2.
|
| 1359 |
-
"logits/rejected": -2.
|
| 1360 |
-
"logps/chosen": -
|
| 1361 |
-
"logps/rejected": -
|
| 1362 |
-
"loss": 0.
|
| 1363 |
-
"pred_label":
|
| 1364 |
-
"rewards/accuracies": 0.
|
| 1365 |
-
"rewards/chosen": -
|
| 1366 |
-
"rewards/margins":
|
| 1367 |
-
"rewards/rejected": -
|
| 1368 |
"step": 840,
|
| 1369 |
-
"use_label":
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
"epoch": 0.89,
|
| 1373 |
"learning_rate": 6.111757857974389e-06,
|
| 1374 |
-
"logits/chosen": -2.
|
| 1375 |
-
"logits/rejected": -2.
|
| 1376 |
-
"logps/chosen": -
|
| 1377 |
-
"logps/rejected": -
|
| 1378 |
-
"loss": 0.
|
| 1379 |
-
"pred_label":
|
| 1380 |
"rewards/accuracies": 0.7749999761581421,
|
| 1381 |
-
"rewards/chosen": -
|
| 1382 |
-
"rewards/margins":
|
| 1383 |
-
"rewards/rejected": -
|
| 1384 |
"step": 850,
|
| 1385 |
-
"use_label":
|
| 1386 |
},
|
| 1387 |
{
|
| 1388 |
"epoch": 0.9,
|
| 1389 |
"learning_rate": 5.529685681024447e-06,
|
| 1390 |
-
"logits/chosen": -2.
|
| 1391 |
-
"logits/rejected": -2.
|
| 1392 |
-
"logps/chosen": -
|
| 1393 |
-
"logps/rejected": -
|
| 1394 |
-
"loss": 0.
|
| 1395 |
-
"pred_label":
|
| 1396 |
-
"rewards/accuracies": 0.
|
| 1397 |
-
"rewards/chosen": -
|
| 1398 |
-
"rewards/margins":
|
| 1399 |
-
"rewards/rejected": -
|
| 1400 |
"step": 860,
|
| 1401 |
-
"use_label":
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
"epoch": 0.91,
|
| 1405 |
"learning_rate": 4.947613504074506e-06,
|
| 1406 |
-
"logits/chosen": -2.
|
| 1407 |
-
"logits/rejected": -2.
|
| 1408 |
-
"logps/chosen": -
|
| 1409 |
-
"logps/rejected": -
|
| 1410 |
-
"loss": 0.
|
| 1411 |
-
"pred_label":
|
| 1412 |
-
"rewards/accuracies": 0.
|
| 1413 |
-
"rewards/chosen": -
|
| 1414 |
-
"rewards/margins":
|
| 1415 |
-
"rewards/rejected": -
|
| 1416 |
"step": 870,
|
| 1417 |
-
"use_label":
|
| 1418 |
},
|
| 1419 |
{
|
| 1420 |
"epoch": 0.92,
|
| 1421 |
"learning_rate": 4.3655413271245635e-06,
|
| 1422 |
-
"logits/chosen": -2.
|
| 1423 |
-
"logits/rejected": -2.
|
| 1424 |
-
"logps/chosen": -
|
| 1425 |
-
"logps/rejected": -
|
| 1426 |
-
"loss": 0.
|
| 1427 |
-
"pred_label":
|
| 1428 |
-
"rewards/accuracies": 0.
|
| 1429 |
-
"rewards/chosen": -
|
| 1430 |
-
"rewards/margins": 2.
|
| 1431 |
-
"rewards/rejected": -
|
| 1432 |
"step": 880,
|
| 1433 |
-
"use_label":
|
| 1434 |
},
|
| 1435 |
{
|
| 1436 |
"epoch": 0.93,
|
| 1437 |
"learning_rate": 3.7834691501746217e-06,
|
| 1438 |
-
"logits/chosen": -2.
|
| 1439 |
-
"logits/rejected": -2.
|
| 1440 |
-
"logps/chosen": -
|
| 1441 |
-
"logps/rejected": -
|
| 1442 |
-
"loss": 0.
|
| 1443 |
-
"pred_label":
|
| 1444 |
-
"rewards/accuracies": 0.
|
| 1445 |
-
"rewards/chosen": -
|
| 1446 |
-
"rewards/margins":
|
| 1447 |
-
"rewards/rejected": -
|
| 1448 |
"step": 890,
|
| 1449 |
-
"use_label":
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
"epoch": 0.94,
|
| 1453 |
"learning_rate": 3.2013969732246805e-06,
|
| 1454 |
-
"logits/chosen": -2.
|
| 1455 |
-
"logits/rejected": -2.
|
| 1456 |
-
"logps/chosen": -
|
| 1457 |
-
"logps/rejected": -
|
| 1458 |
-
"loss": 0.
|
| 1459 |
-
"pred_label":
|
| 1460 |
-
"rewards/accuracies": 0.
|
| 1461 |
-
"rewards/chosen": -
|
| 1462 |
-
"rewards/margins":
|
| 1463 |
-
"rewards/rejected": -
|
| 1464 |
"step": 900,
|
| 1465 |
-
"use_label":
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.95,
|
| 1469 |
"learning_rate": 2.6193247962747383e-06,
|
| 1470 |
-
"logits/chosen": -2.
|
| 1471 |
-
"logits/rejected": -2.
|
| 1472 |
-
"logps/chosen": -
|
| 1473 |
-
"logps/rejected": -
|
| 1474 |
-
"loss": 0.
|
| 1475 |
-
"pred_label":
|
| 1476 |
-
"rewards/accuracies": 0.
|
| 1477 |
-
"rewards/chosen": -
|
| 1478 |
-
"rewards/margins":
|
| 1479 |
-
"rewards/rejected": -
|
| 1480 |
"step": 910,
|
| 1481 |
-
"use_label":
|
| 1482 |
},
|
| 1483 |
{
|
| 1484 |
"epoch": 0.96,
|
| 1485 |
"learning_rate": 2.037252619324796e-06,
|
| 1486 |
-
"logits/chosen": -2.
|
| 1487 |
-
"logits/rejected": -2.
|
| 1488 |
-
"logps/chosen": -
|
| 1489 |
-
"logps/rejected": -
|
| 1490 |
-
"loss": 0.
|
| 1491 |
-
"pred_label":
|
| 1492 |
-
"rewards/accuracies": 0.
|
| 1493 |
-
"rewards/chosen": -
|
| 1494 |
-
"rewards/margins":
|
| 1495 |
-
"rewards/rejected": -
|
| 1496 |
"step": 920,
|
| 1497 |
-
"use_label":
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
"epoch": 0.97,
|
| 1501 |
"learning_rate": 1.4551804423748545e-06,
|
| 1502 |
-
"logits/chosen": -2.
|
| 1503 |
-
"logits/rejected": -2.
|
| 1504 |
-
"logps/chosen": -
|
| 1505 |
-
"logps/rejected": -
|
| 1506 |
-
"loss": 0.
|
| 1507 |
-
"pred_label":
|
| 1508 |
-
"rewards/accuracies": 0.
|
| 1509 |
-
"rewards/chosen": -
|
| 1510 |
-
"rewards/margins":
|
| 1511 |
-
"rewards/rejected": -
|
| 1512 |
"step": 930,
|
| 1513 |
-
"use_label":
|
| 1514 |
},
|
| 1515 |
{
|
| 1516 |
"epoch": 0.98,
|
| 1517 |
"learning_rate": 8.731082654249127e-07,
|
| 1518 |
-
"logits/chosen": -2.
|
| 1519 |
-
"logits/rejected": -2.
|
| 1520 |
-
"logps/chosen": -
|
| 1521 |
-
"logps/rejected": -
|
| 1522 |
-
"loss": 0.
|
| 1523 |
-
"pred_label":
|
| 1524 |
-
"rewards/accuracies": 0.
|
| 1525 |
-
"rewards/chosen": -
|
| 1526 |
-
"rewards/margins":
|
| 1527 |
-
"rewards/rejected": -
|
| 1528 |
"step": 940,
|
| 1529 |
-
"use_label":
|
| 1530 |
},
|
| 1531 |
{
|
| 1532 |
"epoch": 0.99,
|
| 1533 |
"learning_rate": 2.910360884749709e-07,
|
| 1534 |
-
"logits/chosen": -2.
|
| 1535 |
-
"logits/rejected": -2.
|
| 1536 |
-
"logps/chosen": -
|
| 1537 |
-
"logps/rejected": -
|
| 1538 |
-
"loss": 0.
|
| 1539 |
-
"pred_label":
|
| 1540 |
-
"rewards/accuracies": 0.
|
| 1541 |
-
"rewards/chosen": -
|
| 1542 |
-
"rewards/margins":
|
| 1543 |
-
"rewards/rejected": -
|
| 1544 |
"step": 950,
|
| 1545 |
-
"use_label":
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
"epoch": 1.0,
|
| 1549 |
-
"eval_logits/chosen": -2.
|
| 1550 |
-
"eval_logits/rejected": -2.
|
| 1551 |
-
"eval_logps/chosen": -
|
| 1552 |
-
"eval_logps/rejected": -
|
| 1553 |
-
"eval_loss": 0.
|
| 1554 |
-
"eval_pred_label":
|
| 1555 |
-
"eval_rewards/accuracies": 0.
|
| 1556 |
-
"eval_rewards/chosen": -
|
| 1557 |
-
"eval_rewards/margins":
|
| 1558 |
-
"eval_rewards/rejected": -
|
| 1559 |
-
"eval_runtime":
|
| 1560 |
-
"eval_samples_per_second": 4.
|
| 1561 |
-
"eval_steps_per_second": 0.
|
| 1562 |
-
"eval_use_label":
|
| 1563 |
"step": 955
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 1.0,
|
| 1567 |
"step": 955,
|
| 1568 |
"total_flos": 0.0,
|
| 1569 |
-
"train_loss": 0.
|
| 1570 |
-
"train_runtime":
|
| 1571 |
-
"train_samples_per_second": 2.
|
| 1572 |
"train_steps_per_second": 0.038
|
| 1573 |
}
|
| 1574 |
],
|
|
|
|
| 80 |
"logps/chosen": -281.32928466796875,
|
| 81 |
"logps/rejected": -277.8607482910156,
|
| 82 |
"loss": 0.6339,
|
| 83 |
+
"pred_label": 0.0,
|
| 84 |
"rewards/accuracies": 0.6875,
|
| 85 |
"rewards/chosen": 0.02641097828745842,
|
| 86 |
"rewards/margins": 0.2079576551914215,
|
| 87 |
"rewards/rejected": -0.1815466731786728,
|
| 88 |
"step": 40,
|
| 89 |
+
"use_label": 562.0
|
| 90 |
},
|
| 91 |
{
|
| 92 |
"epoch": 0.05,
|
|
|
|
| 96 |
"logps/chosen": -266.80517578125,
|
| 97 |
"logps/rejected": -261.9176025390625,
|
| 98 |
"loss": 0.5914,
|
| 99 |
+
"pred_label": 1.2999999523162842,
|
| 100 |
"rewards/accuracies": 0.643750011920929,
|
| 101 |
"rewards/chosen": -0.06968289613723755,
|
| 102 |
"rewards/margins": 0.33043327927589417,
|
| 103 |
"rewards/rejected": -0.4001162648200989,
|
| 104 |
"step": 50,
|
| 105 |
+
"use_label": 720.7000122070312
|
| 106 |
},
|
| 107 |
{
|
| 108 |
"epoch": 0.06,
|
| 109 |
"learning_rate": 3.125e-05,
|
| 110 |
+
"logits/chosen": -2.818674087524414,
|
| 111 |
+
"logits/rejected": -2.8088903427124023,
|
| 112 |
+
"logps/chosen": -300.7886962890625,
|
| 113 |
+
"logps/rejected": -290.3478698730469,
|
| 114 |
+
"loss": 0.6053,
|
| 115 |
+
"pred_label": 5.875,
|
| 116 |
"rewards/accuracies": 0.699999988079071,
|
| 117 |
+
"rewards/chosen": -0.03734510391950607,
|
| 118 |
+
"rewards/margins": 0.38690200448036194,
|
| 119 |
+
"rewards/rejected": -0.4242470860481262,
|
| 120 |
"step": 60,
|
| 121 |
+
"use_label": 876.125
|
| 122 |
},
|
| 123 |
{
|
| 124 |
"epoch": 0.07,
|
| 125 |
"learning_rate": 3.6458333333333336e-05,
|
| 126 |
+
"logits/chosen": -2.8236336708068848,
|
| 127 |
+
"logits/rejected": -2.818040132522583,
|
| 128 |
+
"logps/chosen": -294.469970703125,
|
| 129 |
+
"logps/rejected": -278.6357421875,
|
| 130 |
+
"loss": 0.535,
|
| 131 |
+
"pred_label": 13.5,
|
| 132 |
+
"rewards/accuracies": 0.737500011920929,
|
| 133 |
+
"rewards/chosen": -0.015645451843738556,
|
| 134 |
+
"rewards/margins": 0.5513776540756226,
|
| 135 |
+
"rewards/rejected": -0.5670231580734253,
|
| 136 |
"step": 70,
|
| 137 |
+
"use_label": 1028.5
|
| 138 |
},
|
| 139 |
{
|
| 140 |
"epoch": 0.08,
|
| 141 |
"learning_rate": 4.166666666666667e-05,
|
| 142 |
+
"logits/chosen": -2.770120143890381,
|
| 143 |
+
"logits/rejected": -2.756917953491211,
|
| 144 |
+
"logps/chosen": -306.6744384765625,
|
| 145 |
+
"logps/rejected": -288.44793701171875,
|
| 146 |
+
"loss": 0.4709,
|
| 147 |
+
"pred_label": 36.79999923706055,
|
| 148 |
+
"rewards/accuracies": 0.762499988079071,
|
| 149 |
+
"rewards/chosen": -0.13690751791000366,
|
| 150 |
+
"rewards/margins": 0.8066266179084778,
|
| 151 |
+
"rewards/rejected": -0.9435340762138367,
|
| 152 |
"step": 80,
|
| 153 |
+
"use_label": 1165.199951171875
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 0.09,
|
| 157 |
"learning_rate": 4.6875e-05,
|
| 158 |
+
"logits/chosen": -2.698793411254883,
|
| 159 |
+
"logits/rejected": -2.75233793258667,
|
| 160 |
+
"logps/chosen": -295.5422668457031,
|
| 161 |
+
"logps/rejected": -274.40032958984375,
|
| 162 |
+
"loss": 0.5055,
|
| 163 |
+
"pred_label": 72.92500305175781,
|
| 164 |
+
"rewards/accuracies": 0.71875,
|
| 165 |
+
"rewards/chosen": 0.02409188821911812,
|
| 166 |
+
"rewards/margins": 0.8366730809211731,
|
| 167 |
+
"rewards/rejected": -0.8125811815261841,
|
| 168 |
"step": 90,
|
| 169 |
+
"use_label": 1289.074951171875
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 0.1,
|
| 173 |
"learning_rate": 4.976717112922003e-05,
|
| 174 |
+
"logits/chosen": -2.719881057739258,
|
| 175 |
+
"logits/rejected": -2.7187676429748535,
|
| 176 |
+
"logps/chosen": -276.2088317871094,
|
| 177 |
+
"logps/rejected": -304.5025329589844,
|
| 178 |
+
"loss": 0.5147,
|
| 179 |
+
"pred_label": 106.19999694824219,
|
| 180 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 181 |
+
"rewards/chosen": -0.20771054923534393,
|
| 182 |
+
"rewards/margins": 0.8888995051383972,
|
| 183 |
+
"rewards/rejected": -1.096610188484192,
|
| 184 |
"step": 100,
|
| 185 |
+
"use_label": 1415.800048828125
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 0.12,
|
| 189 |
"learning_rate": 4.918509895227008e-05,
|
| 190 |
+
"logits/chosen": -2.7200231552124023,
|
| 191 |
+
"logits/rejected": -2.6798348426818848,
|
| 192 |
+
"logps/chosen": -266.4808044433594,
|
| 193 |
+
"logps/rejected": -259.0126037597656,
|
| 194 |
+
"loss": 0.5078,
|
| 195 |
+
"pred_label": 140.47500610351562,
|
| 196 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 197 |
+
"rewards/chosen": -0.18062053620815277,
|
| 198 |
+
"rewards/margins": 0.8032993078231812,
|
| 199 |
+
"rewards/rejected": -0.9839197993278503,
|
| 200 |
"step": 110,
|
| 201 |
+
"use_label": 1541.5250244140625
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 0.13,
|
| 205 |
"learning_rate": 4.860302677532014e-05,
|
| 206 |
+
"logits/chosen": -2.7260735034942627,
|
| 207 |
+
"logits/rejected": -2.729832172393799,
|
| 208 |
+
"logps/chosen": -280.414306640625,
|
| 209 |
+
"logps/rejected": -256.8193664550781,
|
| 210 |
+
"loss": 0.5022,
|
| 211 |
+
"pred_label": 172.5,
|
| 212 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 213 |
+
"rewards/chosen": -0.09014640003442764,
|
| 214 |
+
"rewards/margins": 0.8234178423881531,
|
| 215 |
+
"rewards/rejected": -0.9135642051696777,
|
| 216 |
"step": 120,
|
| 217 |
+
"use_label": 1669.5
|
| 218 |
},
|
| 219 |
{
|
| 220 |
"epoch": 0.14,
|
| 221 |
"learning_rate": 4.80209545983702e-05,
|
| 222 |
+
"logits/chosen": -2.7872612476348877,
|
| 223 |
+
"logits/rejected": -2.748222589492798,
|
| 224 |
+
"logps/chosen": -308.61865234375,
|
| 225 |
+
"logps/rejected": -281.79461669921875,
|
| 226 |
+
"loss": 0.5032,
|
| 227 |
+
"pred_label": 206.75,
|
| 228 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 229 |
+
"rewards/chosen": -0.1351226270198822,
|
| 230 |
+
"rewards/margins": 0.7332055568695068,
|
| 231 |
+
"rewards/rejected": -0.8683282136917114,
|
| 232 |
"step": 130,
|
| 233 |
+
"use_label": 1795.25
|
| 234 |
},
|
| 235 |
{
|
| 236 |
"epoch": 0.15,
|
| 237 |
"learning_rate": 4.743888242142026e-05,
|
| 238 |
+
"logits/chosen": -2.794041872024536,
|
| 239 |
+
"logits/rejected": -2.7909293174743652,
|
| 240 |
+
"logps/chosen": -260.35333251953125,
|
| 241 |
+
"logps/rejected": -247.11325073242188,
|
| 242 |
+
"loss": 0.4325,
|
| 243 |
+
"pred_label": 246.6999969482422,
|
| 244 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 245 |
+
"rewards/chosen": -0.22465598583221436,
|
| 246 |
+
"rewards/margins": 0.984777569770813,
|
| 247 |
+
"rewards/rejected": -1.209433674812317,
|
| 248 |
"step": 140,
|
| 249 |
+
"use_label": 1915.300048828125
|
| 250 |
},
|
| 251 |
{
|
| 252 |
"epoch": 0.16,
|
| 253 |
"learning_rate": 4.685681024447032e-05,
|
| 254 |
+
"logits/chosen": -2.705472946166992,
|
| 255 |
+
"logits/rejected": -2.7284083366394043,
|
| 256 |
+
"logps/chosen": -300.3094787597656,
|
| 257 |
+
"logps/rejected": -294.97125244140625,
|
| 258 |
+
"loss": 0.4364,
|
| 259 |
+
"pred_label": 291.7250061035156,
|
| 260 |
"rewards/accuracies": 0.612500011920929,
|
| 261 |
+
"rewards/chosen": -0.5423473119735718,
|
| 262 |
+
"rewards/margins": 0.7800337076187134,
|
| 263 |
+
"rewards/rejected": -1.3223809003829956,
|
| 264 |
"step": 150,
|
| 265 |
+
"use_label": 2030.2750244140625
|
| 266 |
},
|
| 267 |
{
|
| 268 |
"epoch": 0.17,
|
| 269 |
"learning_rate": 4.6274738067520374e-05,
|
| 270 |
+
"logits/chosen": -2.743900775909424,
|
| 271 |
+
"logits/rejected": -2.688366174697876,
|
| 272 |
+
"logps/chosen": -263.2331237792969,
|
| 273 |
+
"logps/rejected": -281.53582763671875,
|
| 274 |
+
"loss": 0.4222,
|
| 275 |
+
"pred_label": 345.32501220703125,
|
| 276 |
+
"rewards/accuracies": 0.699999988079071,
|
| 277 |
+
"rewards/chosen": -0.0851084440946579,
|
| 278 |
+
"rewards/margins": 1.1642277240753174,
|
| 279 |
+
"rewards/rejected": -1.2493362426757812,
|
| 280 |
"step": 160,
|
| 281 |
+
"use_label": 2136.675048828125
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"epoch": 0.18,
|
| 285 |
"learning_rate": 4.5692665890570435e-05,
|
| 286 |
+
"logits/chosen": -2.7768936157226562,
|
| 287 |
+
"logits/rejected": -2.779693126678467,
|
| 288 |
+
"logps/chosen": -297.79376220703125,
|
| 289 |
+
"logps/rejected": -261.3406982421875,
|
| 290 |
+
"loss": 0.4307,
|
| 291 |
+
"pred_label": 394.0,
|
| 292 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 293 |
+
"rewards/chosen": -0.13176263868808746,
|
| 294 |
+
"rewards/margins": 1.0377919673919678,
|
| 295 |
+
"rewards/rejected": -1.169554591178894,
|
| 296 |
"step": 170,
|
| 297 |
+
"use_label": 2248.0
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"epoch": 0.19,
|
| 301 |
"learning_rate": 4.511059371362049e-05,
|
| 302 |
+
"logits/chosen": -2.6910805702209473,
|
| 303 |
+
"logits/rejected": -2.6632163524627686,
|
| 304 |
+
"logps/chosen": -259.5965270996094,
|
| 305 |
+
"logps/rejected": -289.4101257324219,
|
| 306 |
+
"loss": 0.4077,
|
| 307 |
+
"pred_label": 464.125,
|
| 308 |
"rewards/accuracies": 0.731249988079071,
|
| 309 |
+
"rewards/chosen": -0.3667333722114563,
|
| 310 |
+
"rewards/margins": 1.5060479640960693,
|
| 311 |
+
"rewards/rejected": -1.8727811574935913,
|
| 312 |
"step": 180,
|
| 313 |
+
"use_label": 2337.875
|
| 314 |
},
|
| 315 |
{
|
| 316 |
"epoch": 0.2,
|
| 317 |
"learning_rate": 4.452852153667055e-05,
|
| 318 |
+
"logits/chosen": -2.7862701416015625,
|
| 319 |
+
"logits/rejected": -2.747239589691162,
|
| 320 |
+
"logps/chosen": -308.29376220703125,
|
| 321 |
+
"logps/rejected": -280.5757751464844,
|
| 322 |
+
"loss": 0.3987,
|
| 323 |
+
"pred_label": 532.6500244140625,
|
| 324 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 325 |
+
"rewards/chosen": -0.10377927869558334,
|
| 326 |
+
"rewards/margins": 1.8495817184448242,
|
| 327 |
+
"rewards/rejected": -1.953360915184021,
|
| 328 |
"step": 190,
|
| 329 |
+
"use_label": 2429.35009765625
|
| 330 |
},
|
| 331 |
{
|
| 332 |
"epoch": 0.21,
|
| 333 |
"learning_rate": 4.394644935972061e-05,
|
| 334 |
+
"logits/chosen": -2.787761688232422,
|
| 335 |
+
"logits/rejected": -2.7670867443084717,
|
| 336 |
+
"logps/chosen": -267.0201110839844,
|
| 337 |
+
"logps/rejected": -287.81341552734375,
|
| 338 |
+
"loss": 0.4013,
|
| 339 |
+
"pred_label": 604.4000244140625,
|
| 340 |
+
"rewards/accuracies": 0.737500011920929,
|
| 341 |
+
"rewards/chosen": -0.5691341757774353,
|
| 342 |
+
"rewards/margins": 1.3881704807281494,
|
| 343 |
+
"rewards/rejected": -1.95730459690094,
|
| 344 |
"step": 200,
|
| 345 |
+
"use_label": 2517.60009765625
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 0.22,
|
| 349 |
"learning_rate": 4.336437718277067e-05,
|
| 350 |
+
"logits/chosen": -2.7742085456848145,
|
| 351 |
+
"logits/rejected": -2.794970989227295,
|
| 352 |
+
"logps/chosen": -306.9640808105469,
|
| 353 |
+
"logps/rejected": -271.60906982421875,
|
| 354 |
+
"loss": 0.4304,
|
| 355 |
+
"pred_label": 661.4000244140625,
|
| 356 |
+
"rewards/accuracies": 0.762499988079071,
|
| 357 |
+
"rewards/chosen": -0.5716164708137512,
|
| 358 |
+
"rewards/margins": 1.1703290939331055,
|
| 359 |
+
"rewards/rejected": -1.7419456243515015,
|
| 360 |
"step": 210,
|
| 361 |
+
"use_label": 2620.60009765625
|
| 362 |
},
|
| 363 |
{
|
| 364 |
"epoch": 0.23,
|
| 365 |
"learning_rate": 4.278230500582072e-05,
|
| 366 |
+
"logits/chosen": -2.7252678871154785,
|
| 367 |
+
"logits/rejected": -2.740291118621826,
|
| 368 |
+
"logps/chosen": -280.5372009277344,
|
| 369 |
+
"logps/rejected": -290.718994140625,
|
| 370 |
+
"loss": 0.3818,
|
| 371 |
+
"pred_label": 722.1749877929688,
|
| 372 |
"rewards/accuracies": 0.75,
|
| 373 |
+
"rewards/chosen": -0.5545529723167419,
|
| 374 |
+
"rewards/margins": 1.4516589641571045,
|
| 375 |
+
"rewards/rejected": -2.006211996078491,
|
| 376 |
"step": 220,
|
| 377 |
+
"use_label": 2719.824951171875
|
| 378 |
},
|
| 379 |
{
|
| 380 |
"epoch": 0.24,
|
| 381 |
"learning_rate": 4.220023282887078e-05,
|
| 382 |
+
"logits/chosen": -2.7117373943328857,
|
| 383 |
+
"logits/rejected": -2.7296900749206543,
|
| 384 |
+
"logps/chosen": -316.2425231933594,
|
| 385 |
+
"logps/rejected": -287.38421630859375,
|
| 386 |
+
"loss": 0.3674,
|
| 387 |
+
"pred_label": 794.0499877929688,
|
| 388 |
+
"rewards/accuracies": 0.75,
|
| 389 |
+
"rewards/chosen": -0.8069397211074829,
|
| 390 |
+
"rewards/margins": 1.5205641984939575,
|
| 391 |
+
"rewards/rejected": -2.3275039196014404,
|
| 392 |
"step": 230,
|
| 393 |
+
"use_label": 2807.949951171875
|
| 394 |
},
|
| 395 |
{
|
| 396 |
"epoch": 0.25,
|
| 397 |
"learning_rate": 4.161816065192084e-05,
|
| 398 |
+
"logits/chosen": -2.7252871990203857,
|
| 399 |
+
"logits/rejected": -2.7065608501434326,
|
| 400 |
+
"logps/chosen": -281.2114562988281,
|
| 401 |
+
"logps/rejected": -245.62576293945312,
|
| 402 |
+
"loss": 0.4159,
|
| 403 |
+
"pred_label": 862.6500244140625,
|
| 404 |
+
"rewards/accuracies": 0.706250011920929,
|
| 405 |
+
"rewards/chosen": -0.6634035706520081,
|
| 406 |
+
"rewards/margins": 1.5971952676773071,
|
| 407 |
+
"rewards/rejected": -2.260598659515381,
|
| 408 |
"step": 240,
|
| 409 |
+
"use_label": 2899.35009765625
|
| 410 |
},
|
| 411 |
{
|
| 412 |
"epoch": 0.26,
|
| 413 |
"learning_rate": 4.10360884749709e-05,
|
| 414 |
+
"logits/chosen": -2.7199912071228027,
|
| 415 |
+
"logits/rejected": -2.709681987762451,
|
| 416 |
+
"logps/chosen": -296.9573974609375,
|
| 417 |
+
"logps/rejected": -291.0780334472656,
|
| 418 |
+
"loss": 0.4335,
|
| 419 |
+
"pred_label": 924.7000122070312,
|
| 420 |
+
"rewards/accuracies": 0.737500011920929,
|
| 421 |
+
"rewards/chosen": -0.63172847032547,
|
| 422 |
+
"rewards/margins": 1.5341920852661133,
|
| 423 |
+
"rewards/rejected": -2.1659207344055176,
|
| 424 |
"step": 250,
|
| 425 |
+
"use_label": 2997.300048828125
|
| 426 |
},
|
| 427 |
{
|
| 428 |
"epoch": 0.27,
|
| 429 |
"learning_rate": 4.045401629802096e-05,
|
| 430 |
+
"logits/chosen": -2.7235524654388428,
|
| 431 |
+
"logits/rejected": -2.7239880561828613,
|
| 432 |
+
"logps/chosen": -256.81951904296875,
|
| 433 |
+
"logps/rejected": -270.71209716796875,
|
| 434 |
+
"loss": 0.4011,
|
| 435 |
+
"pred_label": 982.9749755859375,
|
| 436 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 437 |
+
"rewards/chosen": -0.8210113644599915,
|
| 438 |
+
"rewards/margins": 1.2212928533554077,
|
| 439 |
+
"rewards/rejected": -2.042304277420044,
|
| 440 |
"step": 260,
|
| 441 |
+
"use_label": 3099.02490234375
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 0.28,
|
| 445 |
"learning_rate": 3.9871944121071014e-05,
|
| 446 |
+
"logits/chosen": -2.7312769889831543,
|
| 447 |
+
"logits/rejected": -2.739067554473877,
|
| 448 |
+
"logps/chosen": -298.7057800292969,
|
| 449 |
+
"logps/rejected": -300.0126953125,
|
| 450 |
+
"loss": 0.4119,
|
| 451 |
+
"pred_label": 1049.5,
|
| 452 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 453 |
+
"rewards/chosen": -0.6834469437599182,
|
| 454 |
+
"rewards/margins": 1.8426272869110107,
|
| 455 |
+
"rewards/rejected": -2.5260744094848633,
|
| 456 |
"step": 270,
|
| 457 |
+
"use_label": 3192.5
|
| 458 |
},
|
| 459 |
{
|
| 460 |
"epoch": 0.29,
|
| 461 |
"learning_rate": 3.928987194412107e-05,
|
| 462 |
+
"logits/chosen": -2.807619333267212,
|
| 463 |
+
"logits/rejected": -2.8070969581604004,
|
| 464 |
+
"logps/chosen": -298.39654541015625,
|
| 465 |
+
"logps/rejected": -282.5412902832031,
|
| 466 |
+
"loss": 0.401,
|
| 467 |
+
"pred_label": 1114.199951171875,
|
| 468 |
+
"rewards/accuracies": 0.78125,
|
| 469 |
+
"rewards/chosen": -0.44325321912765503,
|
| 470 |
+
"rewards/margins": 1.5879560708999634,
|
| 471 |
+
"rewards/rejected": -2.0312094688415527,
|
| 472 |
"step": 280,
|
| 473 |
+
"use_label": 3287.800048828125
|
| 474 |
},
|
| 475 |
{
|
| 476 |
"epoch": 0.3,
|
| 477 |
"learning_rate": 3.870779976717113e-05,
|
| 478 |
+
"logits/chosen": -2.7824301719665527,
|
| 479 |
+
"logits/rejected": -2.762150764465332,
|
| 480 |
+
"logps/chosen": -300.687255859375,
|
| 481 |
+
"logps/rejected": -250.01663208007812,
|
| 482 |
+
"loss": 0.3923,
|
| 483 |
+
"pred_label": 1174.4749755859375,
|
| 484 |
+
"rewards/accuracies": 0.706250011920929,
|
| 485 |
+
"rewards/chosen": -0.7764869928359985,
|
| 486 |
+
"rewards/margins": 1.2035253047943115,
|
| 487 |
+
"rewards/rejected": -1.9800125360488892,
|
| 488 |
"step": 290,
|
| 489 |
+
"use_label": 3387.52490234375
|
| 490 |
},
|
| 491 |
{
|
| 492 |
"epoch": 0.31,
|
| 493 |
"learning_rate": 3.812572759022119e-05,
|
| 494 |
+
"logits/chosen": -2.769491195678711,
|
| 495 |
+
"logits/rejected": -2.732734203338623,
|
| 496 |
+
"logps/chosen": -271.75091552734375,
|
| 497 |
+
"logps/rejected": -263.69464111328125,
|
| 498 |
+
"loss": 0.4009,
|
| 499 |
+
"pred_label": 1240.0250244140625,
|
| 500 |
+
"rewards/accuracies": 0.699999988079071,
|
| 501 |
+
"rewards/chosen": -0.8072065114974976,
|
| 502 |
+
"rewards/margins": 1.5778390169143677,
|
| 503 |
+
"rewards/rejected": -2.3850455284118652,
|
| 504 |
"step": 300,
|
| 505 |
+
"use_label": 3481.97509765625
|
| 506 |
},
|
| 507 |
{
|
| 508 |
"epoch": 0.32,
|
| 509 |
"learning_rate": 3.7543655413271246e-05,
|
| 510 |
+
"logits/chosen": -2.7556052207946777,
|
| 511 |
+
"logits/rejected": -2.6895577907562256,
|
| 512 |
+
"logps/chosen": -273.4451904296875,
|
| 513 |
+
"logps/rejected": -317.7626953125,
|
| 514 |
+
"loss": 0.3834,
|
| 515 |
+
"pred_label": 1312.050048828125,
|
| 516 |
+
"rewards/accuracies": 0.800000011920929,
|
| 517 |
+
"rewards/chosen": -0.6513568162918091,
|
| 518 |
+
"rewards/margins": 2.270569324493408,
|
| 519 |
+
"rewards/rejected": -2.921926259994507,
|
| 520 |
"step": 310,
|
| 521 |
+
"use_label": 3569.949951171875
|
| 522 |
},
|
| 523 |
{
|
| 524 |
"epoch": 0.33,
|
| 525 |
"learning_rate": 3.696158323632131e-05,
|
| 526 |
+
"logits/chosen": -2.73972749710083,
|
| 527 |
+
"logits/rejected": -2.7072513103485107,
|
| 528 |
+
"logps/chosen": -303.25665283203125,
|
| 529 |
+
"logps/rejected": -305.5859069824219,
|
| 530 |
+
"loss": 0.3917,
|
| 531 |
+
"pred_label": 1381.9000244140625,
|
| 532 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 533 |
+
"rewards/chosen": -0.7276700735092163,
|
| 534 |
+
"rewards/margins": 1.7501161098480225,
|
| 535 |
+
"rewards/rejected": -2.4777863025665283,
|
| 536 |
"step": 320,
|
| 537 |
+
"use_label": 3660.10009765625
|
| 538 |
},
|
| 539 |
{
|
| 540 |
"epoch": 0.35,
|
| 541 |
"learning_rate": 3.637951105937136e-05,
|
| 542 |
+
"logits/chosen": -2.793794631958008,
|
| 543 |
+
"logits/rejected": -2.7443809509277344,
|
| 544 |
+
"logps/chosen": -277.2012634277344,
|
| 545 |
+
"logps/rejected": -254.2499542236328,
|
| 546 |
+
"loss": 0.4169,
|
| 547 |
+
"pred_label": 1455.949951171875,
|
| 548 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 549 |
+
"rewards/chosen": -0.6252767443656921,
|
| 550 |
+
"rewards/margins": 1.3715183734893799,
|
| 551 |
+
"rewards/rejected": -1.9967950582504272,
|
| 552 |
"step": 330,
|
| 553 |
+
"use_label": 3746.050048828125
|
| 554 |
},
|
| 555 |
{
|
| 556 |
"epoch": 0.36,
|
| 557 |
"learning_rate": 3.579743888242142e-05,
|
| 558 |
+
"logits/chosen": -2.747743606567383,
|
| 559 |
+
"logits/rejected": -2.763869524002075,
|
| 560 |
+
"logps/chosen": -294.94415283203125,
|
| 561 |
+
"logps/rejected": -256.500244140625,
|
| 562 |
+
"loss": 0.4603,
|
| 563 |
+
"pred_label": 1514.300048828125,
|
| 564 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 565 |
+
"rewards/chosen": -0.5764142870903015,
|
| 566 |
+
"rewards/margins": 0.9885671734809875,
|
| 567 |
+
"rewards/rejected": -1.564981460571289,
|
| 568 |
"step": 340,
|
| 569 |
+
"use_label": 3847.699951171875
|
| 570 |
},
|
| 571 |
{
|
| 572 |
"epoch": 0.37,
|
| 573 |
"learning_rate": 3.5215366705471484e-05,
|
| 574 |
+
"logits/chosen": -2.730821132659912,
|
| 575 |
+
"logits/rejected": -2.7214157581329346,
|
| 576 |
+
"logps/chosen": -302.2289733886719,
|
| 577 |
+
"logps/rejected": -301.8661804199219,
|
| 578 |
+
"loss": 0.3947,
|
| 579 |
+
"pred_label": 1581.550048828125,
|
| 580 |
+
"rewards/accuracies": 0.793749988079071,
|
| 581 |
+
"rewards/chosen": -0.503420352935791,
|
| 582 |
+
"rewards/margins": 1.9169738292694092,
|
| 583 |
+
"rewards/rejected": -2.4203941822052,
|
| 584 |
"step": 350,
|
| 585 |
+
"use_label": 3940.449951171875
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"epoch": 0.38,
|
| 589 |
"learning_rate": 3.463329452852154e-05,
|
| 590 |
+
"logits/chosen": -2.804471969604492,
|
| 591 |
+
"logits/rejected": -2.8228957653045654,
|
| 592 |
+
"logps/chosen": -281.3868103027344,
|
| 593 |
+
"logps/rejected": -285.85137939453125,
|
| 594 |
+
"loss": 0.4073,
|
| 595 |
+
"pred_label": 1646.2750244140625,
|
| 596 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 597 |
+
"rewards/chosen": -0.8592750430107117,
|
| 598 |
+
"rewards/margins": 1.2057772874832153,
|
| 599 |
+
"rewards/rejected": -2.0650525093078613,
|
| 600 |
"step": 360,
|
| 601 |
+
"use_label": 4035.72509765625
|
| 602 |
},
|
| 603 |
{
|
| 604 |
"epoch": 0.39,
|
| 605 |
"learning_rate": 3.40512223515716e-05,
|
| 606 |
+
"logits/chosen": -2.7621943950653076,
|
| 607 |
+
"logits/rejected": -2.719569683074951,
|
| 608 |
+
"logps/chosen": -283.62945556640625,
|
| 609 |
+
"logps/rejected": -303.4671936035156,
|
| 610 |
+
"loss": 0.4147,
|
| 611 |
+
"pred_label": 1718.800048828125,
|
| 612 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 613 |
+
"rewards/chosen": -0.7426743507385254,
|
| 614 |
+
"rewards/margins": 2.031805992126465,
|
| 615 |
+
"rewards/rejected": -2.7744803428649902,
|
| 616 |
"step": 370,
|
| 617 |
+
"use_label": 4123.2001953125
|
| 618 |
},
|
| 619 |
{
|
| 620 |
"epoch": 0.4,
|
| 621 |
"learning_rate": 3.3469150174621654e-05,
|
| 622 |
+
"logits/chosen": -2.7821342945098877,
|
| 623 |
+
"logits/rejected": -2.7678303718566895,
|
| 624 |
+
"logps/chosen": -292.40106201171875,
|
| 625 |
+
"logps/rejected": -284.78363037109375,
|
| 626 |
+
"loss": 0.3521,
|
| 627 |
+
"pred_label": 1789.875,
|
| 628 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 629 |
+
"rewards/chosen": -0.7170799374580383,
|
| 630 |
+
"rewards/margins": 1.70342218875885,
|
| 631 |
+
"rewards/rejected": -2.420502185821533,
|
| 632 |
"step": 380,
|
| 633 |
+
"use_label": 4212.125
|
| 634 |
},
|
| 635 |
{
|
| 636 |
"epoch": 0.41,
|
| 637 |
"learning_rate": 3.288707799767171e-05,
|
| 638 |
+
"logits/chosen": -2.673171281814575,
|
| 639 |
+
"logits/rejected": -2.665820837020874,
|
| 640 |
+
"logps/chosen": -289.74224853515625,
|
| 641 |
+
"logps/rejected": -273.6763916015625,
|
| 642 |
+
"loss": 0.3695,
|
| 643 |
+
"pred_label": 1873.8499755859375,
|
| 644 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 645 |
+
"rewards/chosen": -0.6950327157974243,
|
| 646 |
+
"rewards/margins": 2.244393825531006,
|
| 647 |
+
"rewards/rejected": -2.939426898956299,
|
| 648 |
"step": 390,
|
| 649 |
+
"use_label": 4288.14990234375
|
| 650 |
},
|
| 651 |
{
|
| 652 |
"epoch": 0.42,
|
| 653 |
"learning_rate": 3.2305005820721776e-05,
|
| 654 |
+
"logits/chosen": -2.809072971343994,
|
| 655 |
+
"logits/rejected": -2.8203823566436768,
|
| 656 |
+
"logps/chosen": -312.44891357421875,
|
| 657 |
+
"logps/rejected": -279.95147705078125,
|
| 658 |
+
"loss": 0.3548,
|
| 659 |
+
"pred_label": 1947.875,
|
| 660 |
"rewards/accuracies": 0.71875,
|
| 661 |
+
"rewards/chosen": -0.5397918224334717,
|
| 662 |
+
"rewards/margins": 1.5214416980743408,
|
| 663 |
+
"rewards/rejected": -2.0612335205078125,
|
| 664 |
"step": 400,
|
| 665 |
+
"use_label": 4374.125
|
| 666 |
},
|
| 667 |
{
|
| 668 |
"epoch": 0.43,
|
| 669 |
"learning_rate": 3.172293364377183e-05,
|
| 670 |
+
"logits/chosen": -2.7990567684173584,
|
| 671 |
+
"logits/rejected": -2.7979464530944824,
|
| 672 |
+
"logps/chosen": -306.15771484375,
|
| 673 |
+
"logps/rejected": -280.4829406738281,
|
| 674 |
+
"loss": 0.3551,
|
| 675 |
+
"pred_label": 2012.9749755859375,
|
| 676 |
+
"rewards/accuracies": 0.675000011920929,
|
| 677 |
+
"rewards/chosen": -0.806006133556366,
|
| 678 |
+
"rewards/margins": 1.3874499797821045,
|
| 679 |
+
"rewards/rejected": -2.1934561729431152,
|
| 680 |
"step": 410,
|
| 681 |
+
"use_label": 4469.02490234375
|
| 682 |
},
|
| 683 |
{
|
| 684 |
"epoch": 0.44,
|
| 685 |
"learning_rate": 3.1140861466821885e-05,
|
| 686 |
+
"logits/chosen": -2.7010891437530518,
|
| 687 |
+
"logits/rejected": -2.6862704753875732,
|
| 688 |
+
"logps/chosen": -249.2625732421875,
|
| 689 |
+
"logps/rejected": -274.73675537109375,
|
| 690 |
+
"loss": 0.3416,
|
| 691 |
+
"pred_label": 2088.074951171875,
|
| 692 |
"rewards/accuracies": 0.6937500238418579,
|
| 693 |
+
"rewards/chosen": -0.8682168126106262,
|
| 694 |
+
"rewards/margins": 1.693968415260315,
|
| 695 |
+
"rewards/rejected": -2.562185049057007,
|
| 696 |
"step": 420,
|
| 697 |
+
"use_label": 4553.9248046875
|
| 698 |
},
|
| 699 |
{
|
| 700 |
"epoch": 0.45,
|
| 701 |
"learning_rate": 3.055878928987195e-05,
|
| 702 |
+
"logits/chosen": -2.734596014022827,
|
| 703 |
+
"logits/rejected": -2.7307052612304688,
|
| 704 |
+
"logps/chosen": -285.8783264160156,
|
| 705 |
+
"logps/rejected": -281.76055908203125,
|
| 706 |
+
"loss": 0.384,
|
| 707 |
+
"pred_label": 2167.85009765625,
|
| 708 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 709 |
+
"rewards/chosen": -0.7977985739707947,
|
| 710 |
+
"rewards/margins": 1.426555871963501,
|
| 711 |
+
"rewards/rejected": -2.2243542671203613,
|
| 712 |
"step": 430,
|
| 713 |
+
"use_label": 4634.14990234375
|
| 714 |
},
|
| 715 |
{
|
| 716 |
"epoch": 0.46,
|
| 717 |
"learning_rate": 2.9976717112922005e-05,
|
| 718 |
+
"logits/chosen": -2.7989730834960938,
|
| 719 |
+
"logits/rejected": -2.8120064735412598,
|
| 720 |
+
"logps/chosen": -291.6087646484375,
|
| 721 |
+
"logps/rejected": -275.2794494628906,
|
| 722 |
+
"loss": 0.3699,
|
| 723 |
+
"pred_label": 2237.574951171875,
|
| 724 |
+
"rewards/accuracies": 0.6875,
|
| 725 |
+
"rewards/chosen": -0.8338401913642883,
|
| 726 |
+
"rewards/margins": 1.3944308757781982,
|
| 727 |
+
"rewards/rejected": -2.228271007537842,
|
| 728 |
"step": 440,
|
| 729 |
+
"use_label": 4724.4248046875
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"epoch": 0.47,
|
| 733 |
"learning_rate": 2.939464493597206e-05,
|
| 734 |
+
"logits/chosen": -2.5998082160949707,
|
| 735 |
+
"logits/rejected": -2.586780071258545,
|
| 736 |
+
"logps/chosen": -238.42855834960938,
|
| 737 |
+
"logps/rejected": -255.28445434570312,
|
| 738 |
+
"loss": 0.392,
|
| 739 |
+
"pred_label": 2305.5,
|
| 740 |
+
"rewards/accuracies": 0.731249988079071,
|
| 741 |
+
"rewards/chosen": -0.8709820508956909,
|
| 742 |
+
"rewards/margins": 1.692291021347046,
|
| 743 |
+
"rewards/rejected": -2.5632731914520264,
|
| 744 |
"step": 450,
|
| 745 |
+
"use_label": 4816.5
|
| 746 |
},
|
| 747 |
{
|
| 748 |
"epoch": 0.48,
|
| 749 |
"learning_rate": 2.881257275902212e-05,
|
| 750 |
+
"logits/chosen": -2.800619125366211,
|
| 751 |
+
"logits/rejected": -2.7941062450408936,
|
| 752 |
+
"logps/chosen": -308.8480529785156,
|
| 753 |
+
"logps/rejected": -291.00689697265625,
|
| 754 |
+
"loss": 0.3698,
|
| 755 |
+
"pred_label": 2380.89990234375,
|
| 756 |
+
"rewards/accuracies": 0.706250011920929,
|
| 757 |
+
"rewards/chosen": -0.704586386680603,
|
| 758 |
+
"rewards/margins": 1.4961796998977661,
|
| 759 |
+
"rewards/rejected": -2.200766086578369,
|
| 760 |
"step": 460,
|
| 761 |
+
"use_label": 4901.10009765625
|
| 762 |
},
|
| 763 |
{
|
| 764 |
"epoch": 0.49,
|
| 765 |
"learning_rate": 2.8230500582072178e-05,
|
| 766 |
+
"logits/chosen": -2.731538772583008,
|
| 767 |
+
"logits/rejected": -2.70729660987854,
|
| 768 |
+
"logps/chosen": -277.3286437988281,
|
| 769 |
+
"logps/rejected": -282.3288269042969,
|
| 770 |
+
"loss": 0.364,
|
| 771 |
+
"pred_label": 2453.449951171875,
|
| 772 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 773 |
+
"rewards/chosen": -1.0228168964385986,
|
| 774 |
+
"rewards/margins": 1.525810956954956,
|
| 775 |
+
"rewards/rejected": -2.5486273765563965,
|
| 776 |
"step": 470,
|
| 777 |
+
"use_label": 4988.5498046875
|
| 778 |
},
|
| 779 |
{
|
| 780 |
"epoch": 0.5,
|
| 781 |
"learning_rate": 2.7648428405122233e-05,
|
| 782 |
+
"logits/chosen": -2.6976053714752197,
|
| 783 |
+
"logits/rejected": -2.6677191257476807,
|
| 784 |
+
"logps/chosen": -255.345703125,
|
| 785 |
+
"logps/rejected": -273.9075012207031,
|
| 786 |
+
"loss": 0.3902,
|
| 787 |
+
"pred_label": 2528.52490234375,
|
| 788 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 789 |
+
"rewards/chosen": -0.9771437644958496,
|
| 790 |
+
"rewards/margins": 1.5616614818572998,
|
| 791 |
+
"rewards/rejected": -2.5388054847717285,
|
| 792 |
"step": 480,
|
| 793 |
+
"use_label": 5073.47509765625
|
| 794 |
},
|
| 795 |
{
|
| 796 |
"epoch": 0.51,
|
| 797 |
"learning_rate": 2.7066356228172297e-05,
|
| 798 |
+
"logits/chosen": -2.7429401874542236,
|
| 799 |
+
"logits/rejected": -2.7300381660461426,
|
| 800 |
+
"logps/chosen": -275.7494201660156,
|
| 801 |
+
"logps/rejected": -261.87689208984375,
|
| 802 |
+
"loss": 0.3725,
|
| 803 |
+
"pred_label": 2596.800048828125,
|
| 804 |
+
"rewards/accuracies": 0.71875,
|
| 805 |
+
"rewards/chosen": -0.8720412254333496,
|
| 806 |
+
"rewards/margins": 1.576229453086853,
|
| 807 |
+
"rewards/rejected": -2.448270559310913,
|
| 808 |
"step": 490,
|
| 809 |
+
"use_label": 5165.2001953125
|
| 810 |
},
|
| 811 |
{
|
| 812 |
"epoch": 0.52,
|
| 813 |
"learning_rate": 2.6484284051222352e-05,
|
| 814 |
+
"logits/chosen": -2.7378077507019043,
|
| 815 |
+
"logits/rejected": -2.7030460834503174,
|
| 816 |
+
"logps/chosen": -297.50439453125,
|
| 817 |
+
"logps/rejected": -287.47821044921875,
|
| 818 |
+
"loss": 0.3548,
|
| 819 |
+
"pred_label": 2678.425048828125,
|
| 820 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 821 |
+
"rewards/chosen": -0.8047893643379211,
|
| 822 |
+
"rewards/margins": 2.116807222366333,
|
| 823 |
+
"rewards/rejected": -2.9215962886810303,
|
| 824 |
"step": 500,
|
| 825 |
+
"use_label": 5243.5751953125
|
| 826 |
},
|
| 827 |
{
|
| 828 |
"epoch": 0.53,
|
| 829 |
"learning_rate": 2.590221187427241e-05,
|
| 830 |
+
"logits/chosen": -2.7614152431488037,
|
| 831 |
+
"logits/rejected": -2.730996608734131,
|
| 832 |
+
"logps/chosen": -285.0855407714844,
|
| 833 |
+
"logps/rejected": -265.8699951171875,
|
| 834 |
+
"loss": 0.3337,
|
| 835 |
+
"pred_label": 2751.699951171875,
|
| 836 |
"rewards/accuracies": 0.7749999761581421,
|
| 837 |
+
"rewards/chosen": -0.9748345613479614,
|
| 838 |
+
"rewards/margins": 2.262641429901123,
|
| 839 |
+
"rewards/rejected": -3.237475872039795,
|
| 840 |
"step": 510,
|
| 841 |
+
"use_label": 5330.2998046875
|
| 842 |
},
|
| 843 |
{
|
| 844 |
"epoch": 0.54,
|
| 845 |
"learning_rate": 2.532013969732247e-05,
|
| 846 |
+
"logits/chosen": -2.777317523956299,
|
| 847 |
+
"logits/rejected": -2.761885404586792,
|
| 848 |
+
"logps/chosen": -297.0881042480469,
|
| 849 |
+
"logps/rejected": -294.2591247558594,
|
| 850 |
+
"loss": 0.391,
|
| 851 |
+
"pred_label": 2820.324951171875,
|
| 852 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 853 |
+
"rewards/chosen": -0.5134023427963257,
|
| 854 |
+
"rewards/margins": 1.941035270690918,
|
| 855 |
+
"rewards/rejected": -2.454437732696533,
|
| 856 |
"step": 520,
|
| 857 |
+
"use_label": 5421.6748046875
|
| 858 |
},
|
| 859 |
{
|
| 860 |
"epoch": 0.55,
|
| 861 |
"learning_rate": 2.4738067520372525e-05,
|
| 862 |
+
"logits/chosen": -2.779775381088257,
|
| 863 |
+
"logits/rejected": -2.7358291149139404,
|
| 864 |
+
"logps/chosen": -308.1401062011719,
|
| 865 |
+
"logps/rejected": -322.45196533203125,
|
| 866 |
+
"loss": 0.3465,
|
| 867 |
+
"pred_label": 2893.199951171875,
|
| 868 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 869 |
+
"rewards/chosen": -0.7824206948280334,
|
| 870 |
+
"rewards/margins": 2.0653204917907715,
|
| 871 |
+
"rewards/rejected": -2.84774112701416,
|
| 872 |
"step": 530,
|
| 873 |
+
"use_label": 5508.7998046875
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"epoch": 0.57,
|
| 877 |
"learning_rate": 2.4155995343422587e-05,
|
| 878 |
+
"logits/chosen": -2.791782855987549,
|
| 879 |
+
"logits/rejected": -2.7862627506256104,
|
| 880 |
+
"logps/chosen": -296.45611572265625,
|
| 881 |
+
"logps/rejected": -298.04638671875,
|
| 882 |
+
"loss": 0.3581,
|
| 883 |
+
"pred_label": 2980.824951171875,
|
| 884 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 885 |
+
"rewards/chosen": -0.7919517755508423,
|
| 886 |
+
"rewards/margins": 2.415773868560791,
|
| 887 |
+
"rewards/rejected": -3.2077255249023438,
|
| 888 |
"step": 540,
|
| 889 |
+
"use_label": 5581.1748046875
|
| 890 |
},
|
| 891 |
{
|
| 892 |
"epoch": 0.58,
|
| 893 |
"learning_rate": 2.3573923166472644e-05,
|
| 894 |
+
"logits/chosen": -2.761887311935425,
|
| 895 |
+
"logits/rejected": -2.735725164413452,
|
| 896 |
+
"logps/chosen": -296.20660400390625,
|
| 897 |
+
"logps/rejected": -290.74456787109375,
|
| 898 |
+
"loss": 0.3915,
|
| 899 |
+
"pred_label": 3058.625,
|
| 900 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 901 |
+
"rewards/chosen": -0.7003430128097534,
|
| 902 |
+
"rewards/margins": 2.179884433746338,
|
| 903 |
+
"rewards/rejected": -2.880227565765381,
|
| 904 |
"step": 550,
|
| 905 |
+
"use_label": 5663.375
|
| 906 |
},
|
| 907 |
{
|
| 908 |
"epoch": 0.59,
|
| 909 |
"learning_rate": 2.2991850989522702e-05,
|
| 910 |
+
"logits/chosen": -2.7279725074768066,
|
| 911 |
+
"logits/rejected": -2.742814540863037,
|
| 912 |
+
"logps/chosen": -283.19342041015625,
|
| 913 |
+
"logps/rejected": -241.63967895507812,
|
| 914 |
+
"loss": 0.3821,
|
| 915 |
+
"pred_label": 3127.85009765625,
|
| 916 |
+
"rewards/accuracies": 0.737500011920929,
|
| 917 |
+
"rewards/chosen": -0.8789056539535522,
|
| 918 |
+
"rewards/margins": 1.5364501476287842,
|
| 919 |
+
"rewards/rejected": -2.415355920791626,
|
| 920 |
"step": 560,
|
| 921 |
+
"use_label": 5754.14990234375
|
| 922 |
},
|
| 923 |
{
|
| 924 |
"epoch": 0.6,
|
| 925 |
"learning_rate": 2.240977881257276e-05,
|
| 926 |
+
"logits/chosen": -2.757524251937866,
|
| 927 |
+
"logits/rejected": -2.7482261657714844,
|
| 928 |
+
"logps/chosen": -276.85345458984375,
|
| 929 |
+
"logps/rejected": -283.73046875,
|
| 930 |
+
"loss": 0.4022,
|
| 931 |
+
"pred_label": 3200.25,
|
| 932 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 933 |
+
"rewards/chosen": -0.736481785774231,
|
| 934 |
+
"rewards/margins": 1.4332786798477173,
|
| 935 |
+
"rewards/rejected": -2.1697604656219482,
|
| 936 |
"step": 570,
|
| 937 |
+
"use_label": 5841.75
|
| 938 |
},
|
| 939 |
{
|
| 940 |
"epoch": 0.61,
|
| 941 |
"learning_rate": 2.1827706635622818e-05,
|
| 942 |
+
"logits/chosen": -2.7960715293884277,
|
| 943 |
+
"logits/rejected": -2.7829785346984863,
|
| 944 |
+
"logps/chosen": -282.49871826171875,
|
| 945 |
+
"logps/rejected": -286.5024719238281,
|
| 946 |
+
"loss": 0.404,
|
| 947 |
+
"pred_label": 3271.77490234375,
|
| 948 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 949 |
+
"rewards/chosen": -0.4390278458595276,
|
| 950 |
+
"rewards/margins": 1.7644996643066406,
|
| 951 |
+
"rewards/rejected": -2.2035276889801025,
|
| 952 |
"step": 580,
|
| 953 |
+
"use_label": 5930.22509765625
|
| 954 |
},
|
| 955 |
{
|
| 956 |
"epoch": 0.62,
|
| 957 |
"learning_rate": 2.124563445867288e-05,
|
| 958 |
+
"logits/chosen": -2.7188072204589844,
|
| 959 |
+
"logits/rejected": -2.70890474319458,
|
| 960 |
+
"logps/chosen": -278.3678894042969,
|
| 961 |
+
"logps/rejected": -304.47906494140625,
|
| 962 |
+
"loss": 0.3951,
|
| 963 |
+
"pred_label": 3332.175048828125,
|
| 964 |
"rewards/accuracies": 0.6875,
|
| 965 |
+
"rewards/chosen": -0.6770004034042358,
|
| 966 |
+
"rewards/margins": 1.6928961277008057,
|
| 967 |
+
"rewards/rejected": -2.369896411895752,
|
| 968 |
"step": 590,
|
| 969 |
+
"use_label": 6029.8251953125
|
| 970 |
},
|
| 971 |
{
|
| 972 |
"epoch": 0.63,
|
| 973 |
"learning_rate": 2.0663562281722934e-05,
|
| 974 |
+
"logits/chosen": -2.788299322128296,
|
| 975 |
+
"logits/rejected": -2.7893383502960205,
|
| 976 |
+
"logps/chosen": -314.5467529296875,
|
| 977 |
+
"logps/rejected": -325.0448913574219,
|
| 978 |
+
"loss": 0.3942,
|
| 979 |
+
"pred_label": 3400.875,
|
| 980 |
+
"rewards/accuracies": 0.762499988079071,
|
| 981 |
+
"rewards/chosen": -0.11838535964488983,
|
| 982 |
+
"rewards/margins": 2.110067129135132,
|
| 983 |
+
"rewards/rejected": -2.228452205657959,
|
| 984 |
"step": 600,
|
| 985 |
+
"use_label": 6121.125
|
| 986 |
},
|
| 987 |
{
|
| 988 |
"epoch": 0.64,
|
| 989 |
"learning_rate": 2.0081490104772992e-05,
|
| 990 |
+
"logits/chosen": -2.7528319358825684,
|
| 991 |
+
"logits/rejected": -2.7486462593078613,
|
| 992 |
+
"logps/chosen": -311.63922119140625,
|
| 993 |
+
"logps/rejected": -263.623046875,
|
| 994 |
+
"loss": 0.3407,
|
| 995 |
+
"pred_label": 3483.35009765625,
|
| 996 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 997 |
+
"rewards/chosen": -0.5729966163635254,
|
| 998 |
+
"rewards/margins": 1.8266725540161133,
|
| 999 |
+
"rewards/rejected": -2.3996691703796387,
|
| 1000 |
"step": 610,
|
| 1001 |
+
"use_label": 6198.64990234375
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
"epoch": 0.65,
|
| 1005 |
"learning_rate": 1.9499417927823053e-05,
|
| 1006 |
+
"logits/chosen": -2.756082057952881,
|
| 1007 |
+
"logits/rejected": -2.706552028656006,
|
| 1008 |
+
"logps/chosen": -288.4559631347656,
|
| 1009 |
+
"logps/rejected": -281.10784912109375,
|
| 1010 |
+
"loss": 0.3528,
|
| 1011 |
+
"pred_label": 3560.199951171875,
|
| 1012 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1013 |
+
"rewards/chosen": -0.847445011138916,
|
| 1014 |
+
"rewards/margins": 2.0444235801696777,
|
| 1015 |
+
"rewards/rejected": -2.891869068145752,
|
| 1016 |
"step": 620,
|
| 1017 |
+
"use_label": 6281.7998046875
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
"epoch": 0.66,
|
| 1021 |
"learning_rate": 1.8917345750873107e-05,
|
| 1022 |
+
"logits/chosen": -2.754945993423462,
|
| 1023 |
+
"logits/rejected": -2.764570951461792,
|
| 1024 |
+
"logps/chosen": -290.9334411621094,
|
| 1025 |
+
"logps/rejected": -289.19891357421875,
|
| 1026 |
+
"loss": 0.3691,
|
| 1027 |
+
"pred_label": 3639.375,
|
| 1028 |
"rewards/accuracies": 0.7749999761581421,
|
| 1029 |
+
"rewards/chosen": -0.5813703536987305,
|
| 1030 |
+
"rewards/margins": 2.221205949783325,
|
| 1031 |
+
"rewards/rejected": -2.8025763034820557,
|
| 1032 |
"step": 630,
|
| 1033 |
+
"use_label": 6362.625
|
| 1034 |
},
|
| 1035 |
{
|
| 1036 |
"epoch": 0.67,
|
| 1037 |
"learning_rate": 1.833527357392317e-05,
|
| 1038 |
+
"logits/chosen": -2.7366719245910645,
|
| 1039 |
+
"logits/rejected": -2.7163639068603516,
|
| 1040 |
+
"logps/chosen": -257.1309509277344,
|
| 1041 |
+
"logps/rejected": -274.4014892578125,
|
| 1042 |
+
"loss": 0.3542,
|
| 1043 |
+
"pred_label": 3707.27490234375,
|
| 1044 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1045 |
+
"rewards/chosen": -0.3939729332923889,
|
| 1046 |
+
"rewards/margins": 2.188854217529297,
|
| 1047 |
+
"rewards/rejected": -2.582827091217041,
|
| 1048 |
"step": 640,
|
| 1049 |
+
"use_label": 6454.72509765625
|
| 1050 |
},
|
| 1051 |
{
|
| 1052 |
"epoch": 0.68,
|
| 1053 |
"learning_rate": 1.7753201396973227e-05,
|
| 1054 |
+
"logits/chosen": -2.7500858306884766,
|
| 1055 |
+
"logits/rejected": -2.72772216796875,
|
| 1056 |
+
"logps/chosen": -308.90020751953125,
|
| 1057 |
+
"logps/rejected": -307.0010070800781,
|
| 1058 |
+
"loss": 0.3495,
|
| 1059 |
+
"pred_label": 3787.10009765625,
|
| 1060 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1061 |
+
"rewards/chosen": -0.6951545476913452,
|
| 1062 |
+
"rewards/margins": 2.2066218852996826,
|
| 1063 |
+
"rewards/rejected": -2.9017765522003174,
|
| 1064 |
"step": 650,
|
| 1065 |
+
"use_label": 6534.89990234375
|
| 1066 |
},
|
| 1067 |
{
|
| 1068 |
"epoch": 0.69,
|
| 1069 |
"learning_rate": 1.717112922002328e-05,
|
| 1070 |
+
"logits/chosen": -2.798872232437134,
|
| 1071 |
+
"logits/rejected": -2.7752318382263184,
|
| 1072 |
+
"logps/chosen": -304.62677001953125,
|
| 1073 |
+
"logps/rejected": -275.2629089355469,
|
| 1074 |
+
"loss": 0.3558,
|
| 1075 |
+
"pred_label": 3870.0,
|
| 1076 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1077 |
+
"rewards/chosen": -0.9263059496879578,
|
| 1078 |
+
"rewards/margins": 1.612428069114685,
|
| 1079 |
+
"rewards/rejected": -2.538734197616577,
|
| 1080 |
"step": 660,
|
| 1081 |
+
"use_label": 6612.0
|
| 1082 |
},
|
| 1083 |
{
|
| 1084 |
"epoch": 0.7,
|
| 1085 |
"learning_rate": 1.6589057043073342e-05,
|
| 1086 |
+
"logits/chosen": -2.7448298931121826,
|
| 1087 |
+
"logits/rejected": -2.7179744243621826,
|
| 1088 |
+
"logps/chosen": -289.99267578125,
|
| 1089 |
+
"logps/rejected": -273.59368896484375,
|
| 1090 |
+
"loss": 0.3439,
|
| 1091 |
+
"pred_label": 3954.050048828125,
|
| 1092 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1093 |
+
"rewards/chosen": -0.8919633030891418,
|
| 1094 |
+
"rewards/margins": 2.3276610374450684,
|
| 1095 |
+
"rewards/rejected": -3.2196242809295654,
|
| 1096 |
"step": 670,
|
| 1097 |
+
"use_label": 6687.9501953125
|
| 1098 |
},
|
| 1099 |
{
|
| 1100 |
"epoch": 0.71,
|
| 1101 |
"learning_rate": 1.60069848661234e-05,
|
| 1102 |
+
"logits/chosen": -2.7539925575256348,
|
| 1103 |
+
"logits/rejected": -2.757519006729126,
|
| 1104 |
+
"logps/chosen": -295.6678771972656,
|
| 1105 |
+
"logps/rejected": -280.3077392578125,
|
| 1106 |
+
"loss": 0.3793,
|
| 1107 |
+
"pred_label": 4029.449951171875,
|
| 1108 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1109 |
+
"rewards/chosen": -0.8732484579086304,
|
| 1110 |
+
"rewards/margins": 1.334531545639038,
|
| 1111 |
+
"rewards/rejected": -2.207780122756958,
|
| 1112 |
"step": 680,
|
| 1113 |
+
"use_label": 6772.5498046875
|
| 1114 |
},
|
| 1115 |
{
|
| 1116 |
"epoch": 0.72,
|
| 1117 |
"learning_rate": 1.5424912689173458e-05,
|
| 1118 |
+
"logits/chosen": -2.736889123916626,
|
| 1119 |
+
"logits/rejected": -2.7280757427215576,
|
| 1120 |
+
"logps/chosen": -257.803466796875,
|
| 1121 |
+
"logps/rejected": -238.80056762695312,
|
| 1122 |
+
"loss": 0.334,
|
| 1123 |
+
"pred_label": 4103.75,
|
| 1124 |
"rewards/accuracies": 0.7749999761581421,
|
| 1125 |
+
"rewards/chosen": -0.7419306039810181,
|
| 1126 |
+
"rewards/margins": 2.2096810340881348,
|
| 1127 |
+
"rewards/rejected": -2.9516117572784424,
|
| 1128 |
"step": 690,
|
| 1129 |
+
"use_label": 6858.25
|
| 1130 |
},
|
| 1131 |
{
|
| 1132 |
"epoch": 0.73,
|
| 1133 |
"learning_rate": 1.4842840512223516e-05,
|
| 1134 |
+
"logits/chosen": -2.7456727027893066,
|
| 1135 |
+
"logits/rejected": -2.7135043144226074,
|
| 1136 |
+
"logps/chosen": -286.5834655761719,
|
| 1137 |
+
"logps/rejected": -281.8011169433594,
|
| 1138 |
+
"loss": 0.323,
|
| 1139 |
+
"pred_label": 4186.8251953125,
|
| 1140 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1141 |
+
"rewards/chosen": -1.0893428325653076,
|
| 1142 |
+
"rewards/margins": 2.2629120349884033,
|
| 1143 |
+
"rewards/rejected": -3.352254867553711,
|
| 1144 |
"step": 700,
|
| 1145 |
+
"use_label": 6935.1748046875
|
| 1146 |
},
|
| 1147 |
{
|
| 1148 |
"epoch": 0.74,
|
| 1149 |
"learning_rate": 1.4260768335273575e-05,
|
| 1150 |
+
"logits/chosen": -2.7708964347839355,
|
| 1151 |
+
"logits/rejected": -2.745072603225708,
|
| 1152 |
+
"logps/chosen": -298.2213134765625,
|
| 1153 |
+
"logps/rejected": -270.7209167480469,
|
| 1154 |
+
"loss": 0.3476,
|
| 1155 |
+
"pred_label": 4265.2998046875,
|
| 1156 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1157 |
+
"rewards/chosen": -0.8646717071533203,
|
| 1158 |
+
"rewards/margins": 2.080548048019409,
|
| 1159 |
+
"rewards/rejected": -2.9452195167541504,
|
| 1160 |
"step": 710,
|
| 1161 |
+
"use_label": 7016.7001953125
|
| 1162 |
},
|
| 1163 |
{
|
| 1164 |
"epoch": 0.75,
|
| 1165 |
"learning_rate": 1.3678696158323633e-05,
|
| 1166 |
+
"logits/chosen": -2.7377257347106934,
|
| 1167 |
+
"logits/rejected": -2.7467942237854004,
|
| 1168 |
+
"logps/chosen": -283.3219909667969,
|
| 1169 |
+
"logps/rejected": -291.44952392578125,
|
| 1170 |
+
"loss": 0.2997,
|
| 1171 |
+
"pred_label": 4351.2998046875,
|
| 1172 |
"rewards/accuracies": 0.78125,
|
| 1173 |
+
"rewards/chosen": -1.1869792938232422,
|
| 1174 |
+
"rewards/margins": 2.617690086364746,
|
| 1175 |
+
"rewards/rejected": -3.8046698570251465,
|
| 1176 |
"step": 720,
|
| 1177 |
+
"use_label": 7090.7001953125
|
| 1178 |
},
|
| 1179 |
{
|
| 1180 |
"epoch": 0.76,
|
| 1181 |
"learning_rate": 1.309662398137369e-05,
|
| 1182 |
+
"logits/chosen": -2.7232143878936768,
|
| 1183 |
+
"logits/rejected": -2.7313239574432373,
|
| 1184 |
+
"logps/chosen": -309.1606750488281,
|
| 1185 |
+
"logps/rejected": -270.8785095214844,
|
| 1186 |
+
"loss": 0.3276,
|
| 1187 |
+
"pred_label": 4437.2998046875,
|
| 1188 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1189 |
+
"rewards/chosen": -1.171083927154541,
|
| 1190 |
+
"rewards/margins": 1.749921202659607,
|
| 1191 |
+
"rewards/rejected": -2.9210047721862793,
|
| 1192 |
"step": 730,
|
| 1193 |
+
"use_label": 7164.7001953125
|
| 1194 |
},
|
| 1195 |
{
|
| 1196 |
"epoch": 0.77,
|
| 1197 |
"learning_rate": 1.2514551804423749e-05,
|
| 1198 |
+
"logits/chosen": -2.7109317779541016,
|
| 1199 |
+
"logits/rejected": -2.722676992416382,
|
| 1200 |
+
"logps/chosen": -320.5135192871094,
|
| 1201 |
+
"logps/rejected": -325.51531982421875,
|
| 1202 |
+
"loss": 0.3404,
|
| 1203 |
+
"pred_label": 4520.64990234375,
|
| 1204 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1205 |
+
"rewards/chosen": -1.3141722679138184,
|
| 1206 |
+
"rewards/margins": 2.5884618759155273,
|
| 1207 |
+
"rewards/rejected": -3.902634382247925,
|
| 1208 |
"step": 740,
|
| 1209 |
+
"use_label": 7241.35009765625
|
| 1210 |
},
|
| 1211 |
{
|
| 1212 |
"epoch": 0.79,
|
| 1213 |
"learning_rate": 1.1932479627473807e-05,
|
| 1214 |
+
"logits/chosen": -2.71235990524292,
|
| 1215 |
+
"logits/rejected": -2.6860625743865967,
|
| 1216 |
+
"logps/chosen": -276.7895202636719,
|
| 1217 |
+
"logps/rejected": -302.22442626953125,
|
| 1218 |
+
"loss": 0.3014,
|
| 1219 |
+
"pred_label": 4610.0751953125,
|
| 1220 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1221 |
+
"rewards/chosen": -1.6301307678222656,
|
| 1222 |
+
"rewards/margins": 2.539576292037964,
|
| 1223 |
+
"rewards/rejected": -4.169707298278809,
|
| 1224 |
"step": 750,
|
| 1225 |
+
"use_label": 7311.9248046875
|
| 1226 |
},
|
| 1227 |
{
|
| 1228 |
"epoch": 0.8,
|
| 1229 |
"learning_rate": 1.1350407450523866e-05,
|
| 1230 |
+
"logits/chosen": -2.7227749824523926,
|
| 1231 |
+
"logits/rejected": -2.76944899559021,
|
| 1232 |
+
"logps/chosen": -306.2732238769531,
|
| 1233 |
+
"logps/rejected": -296.8711242675781,
|
| 1234 |
+
"loss": 0.323,
|
| 1235 |
+
"pred_label": 4696.0751953125,
|
| 1236 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1237 |
+
"rewards/chosen": -1.0810390710830688,
|
| 1238 |
+
"rewards/margins": 2.3947594165802,
|
| 1239 |
+
"rewards/rejected": -3.4757981300354004,
|
| 1240 |
"step": 760,
|
| 1241 |
+
"use_label": 7385.9248046875
|
| 1242 |
},
|
| 1243 |
{
|
| 1244 |
"epoch": 0.81,
|
| 1245 |
"learning_rate": 1.0768335273573923e-05,
|
| 1246 |
+
"logits/chosen": -2.7785048484802246,
|
| 1247 |
+
"logits/rejected": -2.7769179344177246,
|
| 1248 |
+
"logps/chosen": -313.1297607421875,
|
| 1249 |
+
"logps/rejected": -317.2922058105469,
|
| 1250 |
+
"loss": 0.3118,
|
| 1251 |
+
"pred_label": 4793.8251953125,
|
| 1252 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1253 |
+
"rewards/chosen": -1.0833934545516968,
|
| 1254 |
+
"rewards/margins": 2.1734704971313477,
|
| 1255 |
+
"rewards/rejected": -3.256863832473755,
|
| 1256 |
"step": 770,
|
| 1257 |
+
"use_label": 7448.1748046875
|
| 1258 |
},
|
| 1259 |
{
|
| 1260 |
"epoch": 0.82,
|
| 1261 |
"learning_rate": 1.0186263096623982e-05,
|
| 1262 |
+
"logits/chosen": -2.7547028064727783,
|
| 1263 |
+
"logits/rejected": -2.7470812797546387,
|
| 1264 |
+
"logps/chosen": -294.8021545410156,
|
| 1265 |
+
"logps/rejected": -284.2214050292969,
|
| 1266 |
+
"loss": 0.3235,
|
| 1267 |
+
"pred_label": 4885.39990234375,
|
| 1268 |
+
"rewards/accuracies": 0.71875,
|
| 1269 |
+
"rewards/chosen": -1.3608357906341553,
|
| 1270 |
+
"rewards/margins": 2.3477611541748047,
|
| 1271 |
+
"rewards/rejected": -3.708597183227539,
|
| 1272 |
"step": 780,
|
| 1273 |
+
"use_label": 7516.60009765625
|
| 1274 |
},
|
| 1275 |
{
|
| 1276 |
"epoch": 0.83,
|
| 1277 |
"learning_rate": 9.60419091967404e-06,
|
| 1278 |
+
"logits/chosen": -2.7538862228393555,
|
| 1279 |
+
"logits/rejected": -2.7440714836120605,
|
| 1280 |
+
"logps/chosen": -283.4197998046875,
|
| 1281 |
+
"logps/rejected": -299.2707214355469,
|
| 1282 |
+
"loss": 0.335,
|
| 1283 |
+
"pred_label": 4973.375,
|
| 1284 |
+
"rewards/accuracies": 0.71875,
|
| 1285 |
+
"rewards/chosen": -1.5610417127609253,
|
| 1286 |
+
"rewards/margins": 2.2930331230163574,
|
| 1287 |
+
"rewards/rejected": -3.8540749549865723,
|
| 1288 |
"step": 790,
|
| 1289 |
+
"use_label": 7588.625
|
| 1290 |
},
|
| 1291 |
{
|
| 1292 |
"epoch": 0.84,
|
| 1293 |
"learning_rate": 9.022118742724098e-06,
|
| 1294 |
+
"logits/chosen": -2.6376984119415283,
|
| 1295 |
+
"logits/rejected": -2.625143051147461,
|
| 1296 |
+
"logps/chosen": -329.54010009765625,
|
| 1297 |
+
"logps/rejected": -318.21673583984375,
|
| 1298 |
+
"loss": 0.3064,
|
| 1299 |
+
"pred_label": 5072.7998046875,
|
| 1300 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1301 |
+
"rewards/chosen": -1.5068175792694092,
|
| 1302 |
+
"rewards/margins": 2.708643913269043,
|
| 1303 |
+
"rewards/rejected": -4.215461254119873,
|
| 1304 |
"step": 800,
|
| 1305 |
+
"use_label": 7649.2001953125
|
| 1306 |
},
|
| 1307 |
{
|
| 1308 |
"epoch": 0.85,
|
| 1309 |
"learning_rate": 8.440046565774158e-06,
|
| 1310 |
+
"logits/chosen": -2.6940813064575195,
|
| 1311 |
+
"logits/rejected": -2.6780383586883545,
|
| 1312 |
+
"logps/chosen": -297.8315124511719,
|
| 1313 |
+
"logps/rejected": -264.34185791015625,
|
| 1314 |
+
"loss": 0.3022,
|
| 1315 |
+
"pred_label": 5162.1748046875,
|
| 1316 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1317 |
+
"rewards/chosen": -1.6807565689086914,
|
| 1318 |
+
"rewards/margins": 1.872127890586853,
|
| 1319 |
+
"rewards/rejected": -3.552884578704834,
|
| 1320 |
"step": 810,
|
| 1321 |
+
"use_label": 7719.8251953125
|
| 1322 |
},
|
| 1323 |
{
|
| 1324 |
"epoch": 0.86,
|
| 1325 |
"learning_rate": 7.857974388824214e-06,
|
| 1326 |
+
"logits/chosen": -2.7031378746032715,
|
| 1327 |
+
"logits/rejected": -2.714179277420044,
|
| 1328 |
+
"logps/chosen": -305.0399169921875,
|
| 1329 |
+
"logps/rejected": -282.8117980957031,
|
| 1330 |
+
"loss": 0.3364,
|
| 1331 |
+
"pred_label": 5257.125,
|
| 1332 |
+
"rewards/accuracies": 0.78125,
|
| 1333 |
+
"rewards/chosen": -1.1721713542938232,
|
| 1334 |
+
"rewards/margins": 2.952846050262451,
|
| 1335 |
+
"rewards/rejected": -4.125017166137695,
|
| 1336 |
"step": 820,
|
| 1337 |
+
"use_label": 7784.875
|
| 1338 |
},
|
| 1339 |
{
|
| 1340 |
"epoch": 0.87,
|
| 1341 |
"learning_rate": 7.275902211874273e-06,
|
| 1342 |
+
"logits/chosen": -2.7300140857696533,
|
| 1343 |
+
"logits/rejected": -2.728454113006592,
|
| 1344 |
+
"logps/chosen": -278.2652282714844,
|
| 1345 |
+
"logps/rejected": -319.8196716308594,
|
| 1346 |
+
"loss": 0.3174,
|
| 1347 |
+
"pred_label": 5347.625,
|
| 1348 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1349 |
+
"rewards/chosen": -1.5451847314834595,
|
| 1350 |
+
"rewards/margins": 2.416550636291504,
|
| 1351 |
+
"rewards/rejected": -3.961735486984253,
|
| 1352 |
"step": 830,
|
| 1353 |
+
"use_label": 7854.375
|
| 1354 |
},
|
| 1355 |
{
|
| 1356 |
"epoch": 0.88,
|
| 1357 |
"learning_rate": 6.693830034924331e-06,
|
| 1358 |
+
"logits/chosen": -2.716839551925659,
|
| 1359 |
+
"logits/rejected": -2.7346701622009277,
|
| 1360 |
+
"logps/chosen": -283.65179443359375,
|
| 1361 |
+
"logps/rejected": -295.53912353515625,
|
| 1362 |
+
"loss": 0.3542,
|
| 1363 |
+
"pred_label": 5434.0751953125,
|
| 1364 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1365 |
+
"rewards/chosen": -1.5634657144546509,
|
| 1366 |
+
"rewards/margins": 1.5420477390289307,
|
| 1367 |
+
"rewards/rejected": -3.105513334274292,
|
| 1368 |
"step": 840,
|
| 1369 |
+
"use_label": 7927.9248046875
|
| 1370 |
},
|
| 1371 |
{
|
| 1372 |
"epoch": 0.89,
|
| 1373 |
"learning_rate": 6.111757857974389e-06,
|
| 1374 |
+
"logits/chosen": -2.722932815551758,
|
| 1375 |
+
"logits/rejected": -2.772580623626709,
|
| 1376 |
+
"logps/chosen": -308.41778564453125,
|
| 1377 |
+
"logps/rejected": -278.7804260253906,
|
| 1378 |
+
"loss": 0.3299,
|
| 1379 |
+
"pred_label": 5507.02490234375,
|
| 1380 |
"rewards/accuracies": 0.7749999761581421,
|
| 1381 |
+
"rewards/chosen": -1.3390918970108032,
|
| 1382 |
+
"rewards/margins": 2.2028300762176514,
|
| 1383 |
+
"rewards/rejected": -3.541922092437744,
|
| 1384 |
"step": 850,
|
| 1385 |
+
"use_label": 8014.97509765625
|
| 1386 |
},
|
| 1387 |
{
|
| 1388 |
"epoch": 0.9,
|
| 1389 |
"learning_rate": 5.529685681024447e-06,
|
| 1390 |
+
"logits/chosen": -2.7165842056274414,
|
| 1391 |
+
"logits/rejected": -2.7188987731933594,
|
| 1392 |
+
"logps/chosen": -300.7062683105469,
|
| 1393 |
+
"logps/rejected": -310.7410888671875,
|
| 1394 |
+
"loss": 0.3504,
|
| 1395 |
+
"pred_label": 5592.375,
|
| 1396 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1397 |
+
"rewards/chosen": -1.582289457321167,
|
| 1398 |
+
"rewards/margins": 2.2427265644073486,
|
| 1399 |
+
"rewards/rejected": -3.8250160217285156,
|
| 1400 |
"step": 860,
|
| 1401 |
+
"use_label": 8089.625
|
| 1402 |
},
|
| 1403 |
{
|
| 1404 |
"epoch": 0.91,
|
| 1405 |
"learning_rate": 4.947613504074506e-06,
|
| 1406 |
+
"logits/chosen": -2.7352089881896973,
|
| 1407 |
+
"logits/rejected": -2.74865984916687,
|
| 1408 |
+
"logps/chosen": -255.16845703125,
|
| 1409 |
+
"logps/rejected": -277.4416198730469,
|
| 1410 |
+
"loss": 0.3527,
|
| 1411 |
+
"pred_label": 5676.5,
|
| 1412 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1413 |
+
"rewards/chosen": -1.218096137046814,
|
| 1414 |
+
"rewards/margins": 2.5011773109436035,
|
| 1415 |
+
"rewards/rejected": -3.719273328781128,
|
| 1416 |
"step": 870,
|
| 1417 |
+
"use_label": 8165.5
|
| 1418 |
},
|
| 1419 |
{
|
| 1420 |
"epoch": 0.92,
|
| 1421 |
"learning_rate": 4.3655413271245635e-06,
|
| 1422 |
+
"logits/chosen": -2.754638195037842,
|
| 1423 |
+
"logits/rejected": -2.7097582817077637,
|
| 1424 |
+
"logps/chosen": -298.7574157714844,
|
| 1425 |
+
"logps/rejected": -294.0702209472656,
|
| 1426 |
+
"loss": 0.3122,
|
| 1427 |
+
"pred_label": 5764.7001953125,
|
| 1428 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1429 |
+
"rewards/chosen": -1.216820240020752,
|
| 1430 |
+
"rewards/margins": 2.025149345397949,
|
| 1431 |
+
"rewards/rejected": -3.241969585418701,
|
| 1432 |
"step": 880,
|
| 1433 |
+
"use_label": 8237.2998046875
|
| 1434 |
},
|
| 1435 |
{
|
| 1436 |
"epoch": 0.93,
|
| 1437 |
"learning_rate": 3.7834691501746217e-06,
|
| 1438 |
+
"logits/chosen": -2.761935234069824,
|
| 1439 |
+
"logits/rejected": -2.767670154571533,
|
| 1440 |
+
"logps/chosen": -291.3049011230469,
|
| 1441 |
+
"logps/rejected": -307.02984619140625,
|
| 1442 |
+
"loss": 0.3179,
|
| 1443 |
+
"pred_label": 5856.4248046875,
|
| 1444 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1445 |
+
"rewards/chosen": -1.0020384788513184,
|
| 1446 |
+
"rewards/margins": 2.4981894493103027,
|
| 1447 |
+
"rewards/rejected": -3.5002281665802,
|
| 1448 |
"step": 890,
|
| 1449 |
+
"use_label": 8305.5751953125
|
| 1450 |
},
|
| 1451 |
{
|
| 1452 |
"epoch": 0.94,
|
| 1453 |
"learning_rate": 3.2013969732246805e-06,
|
| 1454 |
+
"logits/chosen": -2.6637260913848877,
|
| 1455 |
+
"logits/rejected": -2.6392276287078857,
|
| 1456 |
+
"logps/chosen": -291.3471374511719,
|
| 1457 |
+
"logps/rejected": -295.1965637207031,
|
| 1458 |
+
"loss": 0.3437,
|
| 1459 |
+
"pred_label": 5942.14990234375,
|
| 1460 |
+
"rewards/accuracies": 0.78125,
|
| 1461 |
+
"rewards/chosen": -1.2209066152572632,
|
| 1462 |
+
"rewards/margins": 2.44693660736084,
|
| 1463 |
+
"rewards/rejected": -3.6678433418273926,
|
| 1464 |
"step": 900,
|
| 1465 |
+
"use_label": 8379.849609375
|
| 1466 |
},
|
| 1467 |
{
|
| 1468 |
"epoch": 0.95,
|
| 1469 |
"learning_rate": 2.6193247962747383e-06,
|
| 1470 |
+
"logits/chosen": -2.7372214794158936,
|
| 1471 |
+
"logits/rejected": -2.737588405609131,
|
| 1472 |
+
"logps/chosen": -308.132568359375,
|
| 1473 |
+
"logps/rejected": -295.11187744140625,
|
| 1474 |
+
"loss": 0.3504,
|
| 1475 |
+
"pred_label": 6017.4501953125,
|
| 1476 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1477 |
+
"rewards/chosen": -1.17165207862854,
|
| 1478 |
+
"rewards/margins": 2.026106834411621,
|
| 1479 |
+
"rewards/rejected": -3.1977591514587402,
|
| 1480 |
"step": 910,
|
| 1481 |
+
"use_label": 8464.5498046875
|
| 1482 |
},
|
| 1483 |
{
|
| 1484 |
"epoch": 0.96,
|
| 1485 |
"learning_rate": 2.037252619324796e-06,
|
| 1486 |
+
"logits/chosen": -2.6864240169525146,
|
| 1487 |
+
"logits/rejected": -2.693215847015381,
|
| 1488 |
+
"logps/chosen": -308.95989990234375,
|
| 1489 |
+
"logps/rejected": -296.0351867675781,
|
| 1490 |
+
"loss": 0.3214,
|
| 1491 |
+
"pred_label": 6108.375,
|
| 1492 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1493 |
+
"rewards/chosen": -1.2072199583053589,
|
| 1494 |
+
"rewards/margins": 2.437824249267578,
|
| 1495 |
+
"rewards/rejected": -3.6450438499450684,
|
| 1496 |
"step": 920,
|
| 1497 |
+
"use_label": 8533.625
|
| 1498 |
},
|
| 1499 |
{
|
| 1500 |
"epoch": 0.97,
|
| 1501 |
"learning_rate": 1.4551804423748545e-06,
|
| 1502 |
+
"logits/chosen": -2.7318947315216064,
|
| 1503 |
+
"logits/rejected": -2.6931326389312744,
|
| 1504 |
+
"logps/chosen": -274.75,
|
| 1505 |
+
"logps/rejected": -280.5450134277344,
|
| 1506 |
+
"loss": 0.3077,
|
| 1507 |
+
"pred_label": 6194.10009765625,
|
| 1508 |
+
"rewards/accuracies": 0.824999988079071,
|
| 1509 |
+
"rewards/chosen": -0.6536897420883179,
|
| 1510 |
+
"rewards/margins": 3.221295118331909,
|
| 1511 |
+
"rewards/rejected": -3.8749847412109375,
|
| 1512 |
"step": 930,
|
| 1513 |
+
"use_label": 8607.900390625
|
| 1514 |
},
|
| 1515 |
{
|
| 1516 |
"epoch": 0.98,
|
| 1517 |
"learning_rate": 8.731082654249127e-07,
|
| 1518 |
+
"logits/chosen": -2.764519214630127,
|
| 1519 |
+
"logits/rejected": -2.7424118518829346,
|
| 1520 |
+
"logps/chosen": -302.68206787109375,
|
| 1521 |
+
"logps/rejected": -300.8379821777344,
|
| 1522 |
+
"loss": 0.3294,
|
| 1523 |
+
"pred_label": 6278.64990234375,
|
| 1524 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1525 |
+
"rewards/chosen": -1.528961181640625,
|
| 1526 |
+
"rewards/margins": 1.8663454055786133,
|
| 1527 |
+
"rewards/rejected": -3.3953068256378174,
|
| 1528 |
"step": 940,
|
| 1529 |
+
"use_label": 8683.349609375
|
| 1530 |
},
|
| 1531 |
{
|
| 1532 |
"epoch": 0.99,
|
| 1533 |
"learning_rate": 2.910360884749709e-07,
|
| 1534 |
+
"logits/chosen": -2.7359890937805176,
|
| 1535 |
+
"logits/rejected": -2.679305076599121,
|
| 1536 |
+
"logps/chosen": -283.5569152832031,
|
| 1537 |
+
"logps/rejected": -291.2298889160156,
|
| 1538 |
+
"loss": 0.333,
|
| 1539 |
+
"pred_label": 6368.97509765625,
|
| 1540 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1541 |
+
"rewards/chosen": -1.5102250576019287,
|
| 1542 |
+
"rewards/margins": 2.3179523944854736,
|
| 1543 |
+
"rewards/rejected": -3.8281772136688232,
|
| 1544 |
"step": 950,
|
| 1545 |
+
"use_label": 8753.025390625
|
| 1546 |
},
|
| 1547 |
{
|
| 1548 |
"epoch": 1.0,
|
| 1549 |
+
"eval_logits/chosen": -2.742107391357422,
|
| 1550 |
+
"eval_logits/rejected": -2.739009141921997,
|
| 1551 |
+
"eval_logps/chosen": -295.89080810546875,
|
| 1552 |
+
"eval_logps/rejected": -294.5900573730469,
|
| 1553 |
+
"eval_loss": 0.3439472019672394,
|
| 1554 |
+
"eval_pred_label": 6581.2001953125,
|
| 1555 |
+
"eval_rewards/accuracies": 0.7419999837875366,
|
| 1556 |
+
"eval_rewards/chosen": -1.163341999053955,
|
| 1557 |
+
"eval_rewards/margins": 2.3656928539276123,
|
| 1558 |
+
"eval_rewards/rejected": -3.5290346145629883,
|
| 1559 |
+
"eval_runtime": 460.9775,
|
| 1560 |
+
"eval_samples_per_second": 4.339,
|
| 1561 |
+
"eval_steps_per_second": 0.271,
|
| 1562 |
+
"eval_use_label": 8950.7998046875,
|
| 1563 |
"step": 955
|
| 1564 |
},
|
| 1565 |
{
|
| 1566 |
"epoch": 1.0,
|
| 1567 |
"step": 955,
|
| 1568 |
"total_flos": 0.0,
|
| 1569 |
+
"train_loss": 0.39464539333163756,
|
| 1570 |
+
"train_runtime": 25354.6696,
|
| 1571 |
+
"train_samples_per_second": 2.411,
|
| 1572 |
"train_steps_per_second": 0.038
|
| 1573 |
}
|
| 1574 |
],
|