Model save
Browse files- README.md +21 -1
- all_results.json +5 -5
- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- train_results.json +5 -5
- trainer_state.json +1256 -456
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -16,6 +16,16 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
# zephyr-7b-dpo-full
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
@@ -37,7 +47,7 @@ The following hyperparameters were used during training:
|
|
| 37 |
- learning_rate: 5e-07
|
| 38 |
- train_batch_size: 8
|
| 39 |
- eval_batch_size: 8
|
| 40 |
-
- seed:
|
| 41 |
- distributed_type: multi-GPU
|
| 42 |
- num_devices: 8
|
| 43 |
- gradient_accumulation_steps: 2
|
|
@@ -50,6 +60,16 @@ The following hyperparameters were used during training:
|
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
|
| 55 |
### Framework versions
|
|
|
|
| 16 |
# zephyr-7b-dpo-full
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0870
|
| 21 |
+
- Rewards/chosen: -2.7029
|
| 22 |
+
- Rewards/rejected: -3.6206
|
| 23 |
+
- Rewards/accuracies: 0.7461
|
| 24 |
+
- Rewards/margins: 0.9178
|
| 25 |
+
- Logps/rejected: -619.4178
|
| 26 |
+
- Logps/chosen: -527.3273
|
| 27 |
+
- Logits/rejected: -1.9016
|
| 28 |
+
- Logits/chosen: -1.9549
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
|
|
| 47 |
- learning_rate: 5e-07
|
| 48 |
- train_batch_size: 8
|
| 49 |
- eval_batch_size: 8
|
| 50 |
+
- seed: 1
|
| 51 |
- distributed_type: multi-GPU
|
| 52 |
- num_devices: 8
|
| 53 |
- gradient_accumulation_steps: 2
|
|
|
|
| 60 |
|
| 61 |
### Training results
|
| 62 |
|
| 63 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 64 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 65 |
+
| 0.2588 | 0.11 | 100 | 0.2530 | -0.6480 | -0.9151 | 0.6406 | 0.2671 | -348.8655 | -321.8432 | -2.5122 | -2.5274 |
|
| 66 |
+
| 0.1405 | 0.23 | 200 | 0.1423 | -1.6364 | -2.1668 | 0.7070 | 0.5303 | -474.0294 | -420.6826 | -2.1132 | -2.1455 |
|
| 67 |
+
| 0.0841 | 0.34 | 300 | 0.1030 | -2.2868 | -3.0540 | 0.7383 | 0.7672 | -562.7563 | -485.7206 | -1.9654 | -2.0047 |
|
| 68 |
+
| 0.0916 | 0.45 | 400 | 0.1080 | -2.2956 | -3.0938 | 0.7344 | 0.7982 | -566.7339 | -486.5965 | -1.9394 | -1.9812 |
|
| 69 |
+
| 0.0864 | 0.57 | 500 | 0.0956 | -2.3099 | -3.1316 | 0.7461 | 0.8217 | -570.5160 | -488.0345 | -1.9095 | -1.9560 |
|
| 70 |
+
| 0.065 | 0.68 | 600 | 0.0849 | -2.8564 | -3.7576 | 0.7266 | 0.9012 | -633.1135 | -542.6826 | -1.8868 | -1.9405 |
|
| 71 |
+
| 0.0663 | 0.79 | 700 | 0.0840 | -2.8127 | -3.7340 | 0.7383 | 0.9213 | -630.7556 | -538.3111 | -1.9100 | -1.9627 |
|
| 72 |
+
| 0.0663 | 0.91 | 800 | 0.0870 | -2.7029 | -3.6206 | 0.7461 | 0.9178 | -619.4178 | -527.3273 | -1.9016 | -1.9549 |
|
| 73 |
|
| 74 |
|
| 75 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.11349766382510908,
|
| 4 |
+
"train_runtime": 8005.048,
|
| 5 |
+
"train_samples": 113028,
|
| 6 |
+
"train_samples_per_second": 14.12,
|
| 7 |
+
"train_steps_per_second": 0.11
|
| 8 |
}
|
model-00001-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4943162336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d0100db022e508c7165c36f69d0b9b8bb891630b5098773fb118894879c59a83
|
| 3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999819336
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d1b514fc11019e3edb6963595ffc32c3f463b576c83ad6e179c0fb4d3ecdc9c7
|
| 3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4540516344
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a87d655aa114c81d612aa0a478ca3c21caf414fcf7afc6ba84774275794d4706
|
| 3 |
size 4540516344
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.11349766382510908,
|
| 4 |
+
"train_runtime": 8005.048,
|
| 5 |
+
"train_samples": 113028,
|
| 6 |
+
"train_samples_per_second": 14.12,
|
| 7 |
+
"train_steps_per_second": 0.11
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,21 +1,21 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
-
"eval_steps":
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
-
"learning_rate":
|
| 14 |
-
"logits/chosen": -2.
|
| 15 |
-
"logits/rejected": -2.
|
| 16 |
-
"logps/chosen": -
|
| 17 |
-
"logps/rejected": -
|
| 18 |
-
"loss": 0.
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
| 21 |
"rewards/margins": 0.0,
|
|
@@ -23,579 +23,1379 @@
|
|
| 23 |
"step": 1
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"learning_rate":
|
| 28 |
-
"logits/chosen": -2.
|
| 29 |
-
"logits/rejected": -2.
|
| 30 |
-
"logps/chosen": -
|
| 31 |
-
"logps/rejected": -
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"rewards/accuracies": 0.
|
| 34 |
-
"rewards/chosen":
|
| 35 |
-
"rewards/margins":
|
| 36 |
-
"rewards/rejected": -
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"learning_rate":
|
| 42 |
-
"logits/chosen": -2.
|
| 43 |
-
"logits/rejected": -2.
|
| 44 |
-
"logps/chosen": -
|
| 45 |
-
"logps/rejected": -
|
| 46 |
-
"loss": 0.
|
| 47 |
-
"rewards/accuracies": 0.
|
| 48 |
-
"rewards/chosen": 0.
|
| 49 |
-
"rewards/margins": 0.
|
| 50 |
-
"rewards/rejected": -0.
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"logits/chosen": -2.
|
| 57 |
-
"logits/rejected": -2.
|
| 58 |
-
"logps/chosen": -
|
| 59 |
-
"logps/rejected": -
|
| 60 |
-
"loss": 0.
|
| 61 |
-
"rewards/accuracies": 0.
|
| 62 |
-
"rewards/chosen": 0.
|
| 63 |
-
"rewards/margins": 0.
|
| 64 |
-
"rewards/rejected": -0.
|
| 65 |
"step": 30
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch": 0.
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"logits/chosen": -2.
|
| 71 |
-
"logits/rejected": -2.
|
| 72 |
-
"logps/chosen": -
|
| 73 |
-
"logps/rejected": -
|
| 74 |
-
"loss": 0.
|
| 75 |
-
"rewards/accuracies": 0.
|
| 76 |
-
"rewards/chosen": 0.
|
| 77 |
-
"rewards/margins": 0.
|
| 78 |
-
"rewards/rejected": -0.
|
| 79 |
"step": 40
|
| 80 |
},
|
| 81 |
{
|
| 82 |
-
"epoch": 0.
|
| 83 |
-
"learning_rate":
|
| 84 |
-
"logits/chosen": -2.
|
| 85 |
-
"logits/rejected": -2.
|
| 86 |
-
"logps/chosen": -
|
| 87 |
-
"logps/rejected": -
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"rewards/accuracies": 0.
|
| 90 |
-
"rewards/chosen":
|
| 91 |
-
"rewards/margins": 0.
|
| 92 |
-
"rewards/rejected": -0.
|
| 93 |
"step": 50
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate":
|
| 98 |
-
"logits/chosen": -2.
|
| 99 |
-
"logits/rejected": -2.
|
| 100 |
-
"logps/chosen": -
|
| 101 |
-
"logps/rejected": -
|
| 102 |
-
"loss": 0.
|
| 103 |
-
"rewards/accuracies": 0.
|
| 104 |
-
"rewards/chosen": -0.
|
| 105 |
-
"rewards/margins": 0.
|
| 106 |
-
"rewards/rejected": -
|
| 107 |
"step": 60
|
| 108 |
},
|
| 109 |
{
|
| 110 |
-
"epoch": 0.
|
| 111 |
-
"learning_rate":
|
| 112 |
-
"logits/chosen": -2.
|
| 113 |
-
"logits/rejected": -2.
|
| 114 |
-
"logps/chosen": -
|
| 115 |
-
"logps/rejected": -
|
| 116 |
-
"loss": 0.
|
| 117 |
-
"rewards/accuracies": 0.
|
| 118 |
-
"rewards/chosen": -0.
|
| 119 |
-
"rewards/margins":
|
| 120 |
-
"rewards/rejected": -
|
| 121 |
"step": 70
|
| 122 |
},
|
| 123 |
{
|
| 124 |
-
"epoch": 0.
|
| 125 |
-
"learning_rate": 4.
|
| 126 |
-
"logits/chosen": -2.
|
| 127 |
-
"logits/rejected": -2.
|
| 128 |
-
"logps/chosen": -
|
| 129 |
-
"logps/rejected": -
|
| 130 |
-
"loss": 0.
|
| 131 |
-
"rewards/accuracies": 0.
|
| 132 |
-
"rewards/chosen": -0.
|
| 133 |
-
"rewards/margins":
|
| 134 |
-
"rewards/rejected": -
|
| 135 |
"step": 80
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"learning_rate": 4.
|
| 140 |
-
"logits/chosen": -2.
|
| 141 |
-
"logits/rejected": -2.
|
| 142 |
-
"logps/chosen": -
|
| 143 |
-
"logps/rejected": -
|
| 144 |
-
"loss": 0.
|
| 145 |
-
"rewards/accuracies": 0.
|
| 146 |
-
"rewards/chosen": -0.
|
| 147 |
-
"rewards/margins":
|
| 148 |
-
"rewards/rejected": -
|
| 149 |
"step": 90
|
| 150 |
},
|
| 151 |
{
|
| 152 |
-
"epoch": 0.
|
| 153 |
-
"learning_rate": 4.
|
| 154 |
-
"logits/chosen": -2.
|
| 155 |
-
"logits/rejected": -2.
|
| 156 |
-
"logps/chosen": -
|
| 157 |
-
"logps/rejected": -
|
| 158 |
-
"loss": 0.
|
| 159 |
-
"rewards/accuracies": 0.
|
| 160 |
-
"rewards/chosen": -0.
|
| 161 |
-
"rewards/margins":
|
| 162 |
-
"rewards/rejected": -
|
| 163 |
"step": 100
|
| 164 |
},
|
| 165 |
{
|
| 166 |
-
"epoch": 0.
|
| 167 |
-
"
|
| 168 |
-
"
|
| 169 |
-
"
|
| 170 |
-
"
|
| 171 |
-
"
|
| 172 |
-
"
|
| 173 |
-
"
|
| 174 |
-
"
|
| 175 |
-
"
|
| 176 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 177 |
"step": 110
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
-
"learning_rate": 4.
|
| 182 |
-
"logits/chosen": -2.
|
| 183 |
-
"logits/rejected": -2.
|
| 184 |
-
"logps/chosen": -
|
| 185 |
-
"logps/rejected": -
|
| 186 |
-
"loss": 0.
|
| 187 |
-
"rewards/accuracies": 0.
|
| 188 |
-
"rewards/chosen": -0.
|
| 189 |
-
"rewards/margins":
|
| 190 |
-
"rewards/rejected": -
|
| 191 |
"step": 120
|
| 192 |
},
|
| 193 |
{
|
| 194 |
-
"epoch": 0.
|
| 195 |
-
"learning_rate": 4.
|
| 196 |
-
"logits/chosen": -2.
|
| 197 |
-
"logits/rejected": -2.
|
| 198 |
-
"logps/chosen": -
|
| 199 |
-
"logps/rejected": -
|
| 200 |
-
"loss": 0.
|
| 201 |
-
"rewards/accuracies": 0.
|
| 202 |
-
"rewards/chosen": -0.
|
| 203 |
-
"rewards/margins":
|
| 204 |
-
"rewards/rejected": -
|
| 205 |
"step": 130
|
| 206 |
},
|
| 207 |
{
|
| 208 |
-
"epoch": 0.
|
| 209 |
-
"learning_rate": 4.
|
| 210 |
-
"logits/chosen": -2.
|
| 211 |
-
"logits/rejected": -2.
|
| 212 |
-
"logps/chosen": -
|
| 213 |
-
"logps/rejected": -
|
| 214 |
-
"loss": 0.
|
| 215 |
-
"rewards/accuracies": 0.
|
| 216 |
-
"rewards/chosen": -0.
|
| 217 |
-
"rewards/margins":
|
| 218 |
-
"rewards/rejected": -
|
| 219 |
"step": 140
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
-
"learning_rate":
|
| 224 |
-
"logits/chosen": -2.
|
| 225 |
-
"logits/rejected": -2.
|
| 226 |
-
"logps/chosen": -
|
| 227 |
-
"logps/rejected": -
|
| 228 |
-
"loss": 0.
|
| 229 |
-
"rewards/accuracies": 0.
|
| 230 |
-
"rewards/chosen": -
|
| 231 |
-
"rewards/margins":
|
| 232 |
-
"rewards/rejected": -2.
|
| 233 |
"step": 150
|
| 234 |
},
|
| 235 |
{
|
| 236 |
-
"epoch": 0.
|
| 237 |
-
"learning_rate":
|
| 238 |
-
"logits/chosen": -2.
|
| 239 |
-
"logits/rejected": -2.
|
| 240 |
-
"logps/chosen": -
|
| 241 |
-
"logps/rejected": -
|
| 242 |
-
"loss": 0.
|
| 243 |
-
"rewards/accuracies": 0.
|
| 244 |
-
"rewards/chosen": -
|
| 245 |
-
"rewards/margins":
|
| 246 |
-
"rewards/rejected": -2.
|
| 247 |
"step": 160
|
| 248 |
},
|
| 249 |
{
|
| 250 |
-
"epoch": 0.
|
| 251 |
-
"learning_rate":
|
| 252 |
-
"logits/chosen": -2.
|
| 253 |
-
"logits/rejected": -2.
|
| 254 |
-
"logps/chosen": -
|
| 255 |
-
"logps/rejected": -
|
| 256 |
-
"loss": 0.
|
| 257 |
-
"rewards/accuracies": 0.
|
| 258 |
-
"rewards/chosen": -1.
|
| 259 |
-
"rewards/margins":
|
| 260 |
-
"rewards/rejected": -
|
| 261 |
"step": 170
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
-
"learning_rate":
|
| 266 |
-
"logits/chosen": -2.
|
| 267 |
-
"logits/rejected": -2.
|
| 268 |
-
"logps/chosen": -
|
| 269 |
-
"logps/rejected": -
|
| 270 |
-
"loss": 0.
|
| 271 |
-
"rewards/accuracies": 0.
|
| 272 |
-
"rewards/chosen": -
|
| 273 |
-
"rewards/margins":
|
| 274 |
-
"rewards/rejected": -
|
| 275 |
"step": 180
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 0.
|
| 279 |
-
"learning_rate":
|
| 280 |
-
"logits/chosen": -2.
|
| 281 |
-
"logits/rejected": -2.
|
| 282 |
-
"logps/chosen": -
|
| 283 |
-
"logps/rejected": -
|
| 284 |
-
"loss": 0.
|
| 285 |
-
"rewards/accuracies": 0.
|
| 286 |
-
"rewards/chosen": -
|
| 287 |
-
"rewards/margins":
|
| 288 |
-
"rewards/rejected": -
|
| 289 |
"step": 190
|
| 290 |
},
|
| 291 |
{
|
| 292 |
-
"epoch": 0.
|
| 293 |
-
"learning_rate":
|
| 294 |
-
"logits/chosen": -2.
|
| 295 |
-
"logits/rejected": -2.
|
| 296 |
-
"logps/chosen": -
|
| 297 |
-
"logps/rejected": -
|
| 298 |
-
"loss": 0.
|
| 299 |
-
"rewards/accuracies": 0.
|
| 300 |
-
"rewards/chosen": -
|
| 301 |
-
"rewards/margins":
|
| 302 |
-
"rewards/rejected": -
|
| 303 |
"step": 200
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 0.
|
| 307 |
-
"
|
| 308 |
-
"
|
| 309 |
-
"
|
| 310 |
-
"
|
| 311 |
-
"
|
| 312 |
-
"
|
| 313 |
-
"
|
| 314 |
-
"
|
| 315 |
-
"
|
| 316 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 317 |
"step": 210
|
| 318 |
},
|
| 319 |
{
|
| 320 |
-
"epoch": 0.
|
| 321 |
-
"learning_rate":
|
| 322 |
-
"logits/chosen": -2.
|
| 323 |
-
"logits/rejected": -2.
|
| 324 |
-
"logps/chosen": -
|
| 325 |
-
"logps/rejected": -
|
| 326 |
-
"loss": 0.
|
| 327 |
-
"rewards/accuracies": 0.
|
| 328 |
-
"rewards/chosen": -1.
|
| 329 |
-
"rewards/margins":
|
| 330 |
-
"rewards/rejected": -
|
| 331 |
"step": 220
|
| 332 |
},
|
| 333 |
{
|
| 334 |
-
"epoch": 0.
|
| 335 |
-
"learning_rate":
|
| 336 |
-
"logits/chosen": -2.
|
| 337 |
-
"logits/rejected": -2.
|
| 338 |
-
"logps/chosen": -
|
| 339 |
-
"logps/rejected": -
|
| 340 |
-
"loss": 0.
|
| 341 |
-
"rewards/accuracies": 0.
|
| 342 |
-
"rewards/chosen": -1.
|
| 343 |
-
"rewards/margins": 1.
|
| 344 |
-
"rewards/rejected": -
|
| 345 |
"step": 230
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 0.
|
| 349 |
-
"learning_rate":
|
| 350 |
-
"logits/chosen": -2.
|
| 351 |
-
"logits/rejected": -
|
| 352 |
-
"logps/chosen": -
|
| 353 |
-
"logps/rejected": -
|
| 354 |
-
"loss": 0.
|
| 355 |
-
"rewards/accuracies": 0.
|
| 356 |
-
"rewards/chosen": -
|
| 357 |
-
"rewards/margins":
|
| 358 |
-
"rewards/rejected": -3.
|
| 359 |
"step": 240
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch": 0.
|
| 363 |
-
"learning_rate":
|
| 364 |
-
"logits/chosen": -2.
|
| 365 |
-
"logits/rejected": -
|
| 366 |
-
"logps/chosen": -
|
| 367 |
-
"logps/rejected": -
|
| 368 |
-
"loss": 0.
|
| 369 |
-
"rewards/accuracies": 0.
|
| 370 |
-
"rewards/chosen": -
|
| 371 |
-
"rewards/margins":
|
| 372 |
-
"rewards/rejected": -
|
| 373 |
"step": 250
|
| 374 |
},
|
| 375 |
{
|
| 376 |
-
"epoch": 0.
|
| 377 |
-
"learning_rate":
|
| 378 |
-
"logits/chosen": -2.
|
| 379 |
-
"logits/rejected": -2.
|
| 380 |
-
"logps/chosen": -
|
| 381 |
-
"logps/rejected": -
|
| 382 |
-
"loss": 0.
|
| 383 |
-
"rewards/accuracies": 0.
|
| 384 |
-
"rewards/chosen": -
|
| 385 |
-
"rewards/margins":
|
| 386 |
-
"rewards/rejected": -
|
| 387 |
"step": 260
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch": 0.
|
| 391 |
-
"learning_rate":
|
| 392 |
-
"logits/chosen": -2.
|
| 393 |
-
"logits/rejected": -
|
| 394 |
-
"logps/chosen": -
|
| 395 |
-
"logps/rejected": -
|
| 396 |
-
"loss": 0.
|
| 397 |
-
"rewards/accuracies": 0.
|
| 398 |
-
"rewards/chosen": -
|
| 399 |
-
"rewards/margins":
|
| 400 |
-
"rewards/rejected": -
|
| 401 |
"step": 270
|
| 402 |
},
|
| 403 |
{
|
| 404 |
-
"epoch": 0.
|
| 405 |
-
"learning_rate":
|
| 406 |
-
"logits/chosen": -2.
|
| 407 |
-
"logits/rejected": -
|
| 408 |
-
"logps/chosen": -
|
| 409 |
-
"logps/rejected": -
|
| 410 |
-
"loss": 0.
|
| 411 |
-
"rewards/accuracies": 0.
|
| 412 |
-
"rewards/chosen": -1.
|
| 413 |
-
"rewards/margins":
|
| 414 |
-
"rewards/rejected": -
|
| 415 |
"step": 280
|
| 416 |
},
|
| 417 |
{
|
| 418 |
-
"epoch": 0.
|
| 419 |
-
"learning_rate":
|
| 420 |
-
"logits/chosen": -2.
|
| 421 |
-
"logits/rejected": -
|
| 422 |
-
"logps/chosen": -
|
| 423 |
-
"logps/rejected": -
|
| 424 |
-
"loss": 0.
|
| 425 |
-
"rewards/accuracies": 0.
|
| 426 |
-
"rewards/chosen": -1.
|
| 427 |
-
"rewards/margins":
|
| 428 |
-
"rewards/rejected": -
|
| 429 |
"step": 290
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch": 0.
|
| 433 |
-
"learning_rate":
|
| 434 |
-
"logits/chosen": -
|
| 435 |
-
"logits/rejected": -
|
| 436 |
-
"logps/chosen": -
|
| 437 |
-
"logps/rejected": -
|
| 438 |
-
"loss": 0.
|
| 439 |
-
"rewards/accuracies": 0.
|
| 440 |
-
"rewards/chosen": -
|
| 441 |
-
"rewards/margins":
|
| 442 |
-
"rewards/rejected": -3.
|
| 443 |
"step": 300
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch": 0.
|
| 447 |
-
"
|
| 448 |
-
"
|
| 449 |
-
"
|
| 450 |
-
"
|
| 451 |
-
"
|
| 452 |
-
"
|
| 453 |
-
"
|
| 454 |
-
"
|
| 455 |
-
"
|
| 456 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 457 |
"step": 310
|
| 458 |
},
|
| 459 |
{
|
| 460 |
-
"epoch": 0.
|
| 461 |
-
"learning_rate":
|
| 462 |
-
"logits/chosen": -2.
|
| 463 |
-
"logits/rejected": -
|
| 464 |
-
"logps/chosen": -
|
| 465 |
-
"logps/rejected": -
|
| 466 |
-
"loss": 0.
|
| 467 |
-
"rewards/accuracies": 0.
|
| 468 |
-
"rewards/chosen": -1.
|
| 469 |
-
"rewards/margins":
|
| 470 |
-
"rewards/rejected": -3.
|
| 471 |
"step": 320
|
| 472 |
},
|
| 473 |
{
|
| 474 |
-
"epoch": 0.
|
| 475 |
-
"learning_rate":
|
| 476 |
-
"logits/chosen": -2.
|
| 477 |
-
"logits/rejected": -
|
| 478 |
-
"logps/chosen": -
|
| 479 |
-
"logps/rejected": -
|
| 480 |
-
"loss": 0.
|
| 481 |
-
"rewards/accuracies": 0.
|
| 482 |
-
"rewards/chosen": -
|
| 483 |
-
"rewards/margins":
|
| 484 |
-
"rewards/rejected": -3.
|
| 485 |
"step": 330
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch": 0.
|
| 489 |
-
"learning_rate": 3.
|
| 490 |
-
"logits/chosen": -
|
| 491 |
-
"logits/rejected": -
|
| 492 |
-
"logps/chosen": -
|
| 493 |
-
"logps/rejected": -
|
| 494 |
-
"loss": 0.
|
| 495 |
-
"rewards/accuracies": 0.
|
| 496 |
-
"rewards/chosen": -1.
|
| 497 |
-
"rewards/margins":
|
| 498 |
-
"rewards/rejected": -
|
| 499 |
"step": 340
|
| 500 |
},
|
| 501 |
{
|
| 502 |
-
"epoch": 0.
|
| 503 |
-
"learning_rate":
|
| 504 |
-
"logits/chosen": -2.
|
| 505 |
-
"logits/rejected": -
|
| 506 |
-
"logps/chosen": -
|
| 507 |
-
"logps/rejected": -
|
| 508 |
-
"loss": 0.
|
| 509 |
-
"rewards/accuracies": 0.
|
| 510 |
-
"rewards/chosen": -
|
| 511 |
-
"rewards/margins":
|
| 512 |
-
"rewards/rejected": -3.
|
| 513 |
"step": 350
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch": 0.
|
| 517 |
-
"learning_rate":
|
| 518 |
-
"logits/chosen": -2.
|
| 519 |
-
"logits/rejected": -
|
| 520 |
-
"logps/chosen": -
|
| 521 |
-
"logps/rejected": -
|
| 522 |
-
"loss": 0.
|
| 523 |
-
"rewards/accuracies": 0.
|
| 524 |
-
"rewards/chosen": -1.
|
| 525 |
-
"rewards/margins":
|
| 526 |
-
"rewards/rejected": -
|
| 527 |
"step": 360
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch": 0.
|
| 531 |
-
"learning_rate":
|
| 532 |
-
"logits/chosen": -
|
| 533 |
-
"logits/rejected": -
|
| 534 |
-
"logps/chosen": -
|
| 535 |
-
"logps/rejected": -
|
| 536 |
-
"loss": 0.
|
| 537 |
-
"rewards/accuracies": 0.
|
| 538 |
-
"rewards/chosen": -1.
|
| 539 |
-
"rewards/margins":
|
| 540 |
-
"rewards/rejected": -
|
| 541 |
"step": 370
|
| 542 |
},
|
| 543 |
{
|
| 544 |
-
"epoch": 0.
|
| 545 |
-
"learning_rate":
|
| 546 |
-
"logits/chosen": -
|
| 547 |
-
"logits/rejected": -
|
| 548 |
-
"logps/chosen": -
|
| 549 |
-
"logps/rejected": -
|
| 550 |
-
"loss": 0.
|
| 551 |
-
"rewards/accuracies": 0.
|
| 552 |
-
"rewards/chosen": -
|
| 553 |
-
"rewards/margins":
|
| 554 |
-
"rewards/rejected": -3.
|
| 555 |
"step": 380
|
| 556 |
},
|
| 557 |
{
|
| 558 |
-
"epoch": 0.
|
| 559 |
-
"learning_rate":
|
| 560 |
-
"logits/chosen": -
|
| 561 |
-
"logits/rejected": -
|
| 562 |
-
"logps/chosen": -
|
| 563 |
-
"logps/rejected": -
|
| 564 |
-
"loss": 0.
|
| 565 |
-
"rewards/accuracies": 0.
|
| 566 |
-
"rewards/chosen": -1.
|
| 567 |
-
"rewards/margins":
|
| 568 |
-
"rewards/rejected": -3.
|
| 569 |
"step": 390
|
| 570 |
},
|
| 571 |
{
|
| 572 |
-
"epoch": 0.
|
| 573 |
-
"learning_rate":
|
| 574 |
-
"logits/chosen": -2.
|
| 575 |
-
"logits/rejected": -
|
| 576 |
-
"logps/chosen": -
|
| 577 |
-
"logps/rejected": -
|
| 578 |
-
"loss": 0.
|
| 579 |
-
"rewards/accuracies": 0.
|
| 580 |
-
"rewards/chosen": -1.
|
| 581 |
-
"rewards/margins":
|
| 582 |
-
"rewards/rejected": -
|
| 583 |
"step": 400
|
| 584 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 585 |
{
|
| 586 |
"epoch": 1.0,
|
| 587 |
-
"step":
|
| 588 |
"total_flos": 0.0,
|
| 589 |
-
"train_loss": 0.
|
| 590 |
-
"train_runtime":
|
| 591 |
-
"train_samples_per_second":
|
| 592 |
-
"train_steps_per_second": 0.
|
| 593 |
}
|
| 594 |
],
|
| 595 |
"logging_steps": 10,
|
| 596 |
-
"max_steps":
|
| 597 |
"num_train_epochs": 1,
|
| 598 |
-
"save_steps":
|
| 599 |
"total_flos": 0.0,
|
| 600 |
"trial_name": null,
|
| 601 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994340690435767,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 883,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 5.617977528089887e-09,
|
| 14 |
+
"logits/chosen": -2.763059616088867,
|
| 15 |
+
"logits/rejected": -2.7395401000976562,
|
| 16 |
+
"logps/chosen": -322.45367431640625,
|
| 17 |
+
"logps/rejected": -273.0731506347656,
|
| 18 |
+
"loss": 0.3632,
|
| 19 |
"rewards/accuracies": 0.0,
|
| 20 |
"rewards/chosen": 0.0,
|
| 21 |
"rewards/margins": 0.0,
|
|
|
|
| 23 |
"step": 1
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 5.617977528089887e-08,
|
| 28 |
+
"logits/chosen": -2.7944271564483643,
|
| 29 |
+
"logits/rejected": -2.7713630199432373,
|
| 30 |
+
"logps/chosen": -334.48004150390625,
|
| 31 |
+
"logps/rejected": -186.60906982421875,
|
| 32 |
+
"loss": 0.3527,
|
| 33 |
+
"rewards/accuracies": 0.4375,
|
| 34 |
+
"rewards/chosen": 4.6880424633855e-05,
|
| 35 |
+
"rewards/margins": 2.8881140679004602e-05,
|
| 36 |
+
"rewards/rejected": 1.7999276678892784e-05,
|
| 37 |
"step": 10
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"learning_rate": 1.1235955056179774e-07,
|
| 42 |
+
"logits/chosen": -2.8209891319274902,
|
| 43 |
+
"logits/rejected": -2.8004016876220703,
|
| 44 |
+
"logps/chosen": -334.3337707519531,
|
| 45 |
+
"logps/rejected": -174.12008666992188,
|
| 46 |
+
"loss": 0.3507,
|
| 47 |
+
"rewards/accuracies": 0.65625,
|
| 48 |
+
"rewards/chosen": 0.0022166508715599775,
|
| 49 |
+
"rewards/margins": 0.003457559272646904,
|
| 50 |
+
"rewards/rejected": -0.001240908750332892,
|
| 51 |
"step": 20
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"learning_rate": 1.6853932584269663e-07,
|
| 56 |
+
"logits/chosen": -2.7569632530212402,
|
| 57 |
+
"logits/rejected": -2.7449216842651367,
|
| 58 |
+
"logps/chosen": -318.72857666015625,
|
| 59 |
+
"logps/rejected": -187.9712371826172,
|
| 60 |
+
"loss": 0.3448,
|
| 61 |
+
"rewards/accuracies": 0.637499988079071,
|
| 62 |
+
"rewards/chosen": 0.008794652298092842,
|
| 63 |
+
"rewards/margins": 0.01837952807545662,
|
| 64 |
+
"rewards/rejected": -0.009584875777363777,
|
| 65 |
"step": 30
|
| 66 |
},
|
| 67 |
{
|
| 68 |
+
"epoch": 0.05,
|
| 69 |
+
"learning_rate": 2.2471910112359549e-07,
|
| 70 |
+
"logits/chosen": -2.7657992839813232,
|
| 71 |
+
"logits/rejected": -2.7251369953155518,
|
| 72 |
+
"logps/chosen": -361.417236328125,
|
| 73 |
+
"logps/rejected": -208.3668975830078,
|
| 74 |
+
"loss": 0.3535,
|
| 75 |
+
"rewards/accuracies": 0.668749988079071,
|
| 76 |
+
"rewards/chosen": 0.02502177096903324,
|
| 77 |
+
"rewards/margins": 0.06268725544214249,
|
| 78 |
+
"rewards/rejected": -0.037665486335754395,
|
| 79 |
"step": 40
|
| 80 |
},
|
| 81 |
{
|
| 82 |
+
"epoch": 0.06,
|
| 83 |
+
"learning_rate": 2.8089887640449437e-07,
|
| 84 |
+
"logits/chosen": -2.672497272491455,
|
| 85 |
+
"logits/rejected": -2.6617045402526855,
|
| 86 |
+
"logps/chosen": -290.23529052734375,
|
| 87 |
+
"logps/rejected": -174.43826293945312,
|
| 88 |
+
"loss": 0.3593,
|
| 89 |
+
"rewards/accuracies": 0.6875,
|
| 90 |
+
"rewards/chosen": 0.020228449255228043,
|
| 91 |
+
"rewards/margins": 0.11420907080173492,
|
| 92 |
+
"rewards/rejected": -0.09398062527179718,
|
| 93 |
"step": 50
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 3.3707865168539325e-07,
|
| 98 |
+
"logits/chosen": -2.5486276149749756,
|
| 99 |
+
"logits/rejected": -2.540907621383667,
|
| 100 |
+
"logps/chosen": -329.50079345703125,
|
| 101 |
+
"logps/rejected": -229.45065307617188,
|
| 102 |
+
"loss": 0.3584,
|
| 103 |
+
"rewards/accuracies": 0.643750011920929,
|
| 104 |
+
"rewards/chosen": -0.030313704162836075,
|
| 105 |
+
"rewards/margins": 0.17366722226142883,
|
| 106 |
+
"rewards/rejected": -0.2039809226989746,
|
| 107 |
"step": 60
|
| 108 |
},
|
| 109 |
{
|
| 110 |
+
"epoch": 0.08,
|
| 111 |
+
"learning_rate": 3.9325842696629214e-07,
|
| 112 |
+
"logits/chosen": -2.5654215812683105,
|
| 113 |
+
"logits/rejected": -2.5420703887939453,
|
| 114 |
+
"logps/chosen": -344.0436096191406,
|
| 115 |
+
"logps/rejected": -257.1449890136719,
|
| 116 |
+
"loss": 0.3532,
|
| 117 |
+
"rewards/accuracies": 0.668749988079071,
|
| 118 |
+
"rewards/chosen": -0.13107505440711975,
|
| 119 |
+
"rewards/margins": 0.29146069288253784,
|
| 120 |
+
"rewards/rejected": -0.4225357472896576,
|
| 121 |
"step": 70
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"epoch": 0.09,
|
| 125 |
+
"learning_rate": 4.4943820224719097e-07,
|
| 126 |
+
"logits/chosen": -2.470280170440674,
|
| 127 |
+
"logits/rejected": -2.4680607318878174,
|
| 128 |
+
"logps/chosen": -402.6678161621094,
|
| 129 |
+
"logps/rejected": -240.67776489257812,
|
| 130 |
+
"loss": 0.3141,
|
| 131 |
+
"rewards/accuracies": 0.65625,
|
| 132 |
+
"rewards/chosen": -0.19497649371623993,
|
| 133 |
+
"rewards/margins": 0.45426544547080994,
|
| 134 |
+
"rewards/rejected": -0.6492420434951782,
|
| 135 |
"step": 80
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 4.999980431020109e-07,
|
| 140 |
+
"logits/chosen": -2.4644992351531982,
|
| 141 |
+
"logits/rejected": -2.436156749725342,
|
| 142 |
+
"logps/chosen": -366.55352783203125,
|
| 143 |
+
"logps/rejected": -270.9005432128906,
|
| 144 |
+
"loss": 0.2824,
|
| 145 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 146 |
+
"rewards/chosen": -0.16562625765800476,
|
| 147 |
+
"rewards/margins": 0.6153150796890259,
|
| 148 |
+
"rewards/rejected": -0.780941367149353,
|
| 149 |
"step": 90
|
| 150 |
},
|
| 151 |
{
|
| 152 |
+
"epoch": 0.11,
|
| 153 |
+
"learning_rate": 4.997632524101301e-07,
|
| 154 |
+
"logits/chosen": -2.5054173469543457,
|
| 155 |
+
"logits/rejected": -2.465770959854126,
|
| 156 |
+
"logps/chosen": -371.3381042480469,
|
| 157 |
+
"logps/rejected": -287.1797790527344,
|
| 158 |
+
"loss": 0.2588,
|
| 159 |
+
"rewards/accuracies": 0.71875,
|
| 160 |
+
"rewards/chosen": -0.4102350175380707,
|
| 161 |
+
"rewards/margins": 0.5942500829696655,
|
| 162 |
+
"rewards/rejected": -1.0044851303100586,
|
| 163 |
"step": 100
|
| 164 |
},
|
| 165 |
{
|
| 166 |
+
"epoch": 0.11,
|
| 167 |
+
"eval_logits/chosen": -2.527438163757324,
|
| 168 |
+
"eval_logits/rejected": -2.5121896266937256,
|
| 169 |
+
"eval_logps/chosen": -321.84320068359375,
|
| 170 |
+
"eval_logps/rejected": -348.865478515625,
|
| 171 |
+
"eval_loss": 0.25303778052330017,
|
| 172 |
+
"eval_rewards/accuracies": 0.640625,
|
| 173 |
+
"eval_rewards/chosen": -0.6480357646942139,
|
| 174 |
+
"eval_rewards/margins": 0.2670864164829254,
|
| 175 |
+
"eval_rewards/rejected": -0.9151221513748169,
|
| 176 |
+
"eval_runtime": 53.627,
|
| 177 |
+
"eval_samples_per_second": 37.295,
|
| 178 |
+
"eval_steps_per_second": 0.597,
|
| 179 |
+
"step": 100
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.12,
|
| 183 |
+
"learning_rate": 4.991375032514749e-07,
|
| 184 |
+
"logits/chosen": -2.4942123889923096,
|
| 185 |
+
"logits/rejected": -2.4537932872772217,
|
| 186 |
+
"logps/chosen": -366.0484924316406,
|
| 187 |
+
"logps/rejected": -300.3791198730469,
|
| 188 |
+
"loss": 0.218,
|
| 189 |
+
"rewards/accuracies": 0.699999988079071,
|
| 190 |
+
"rewards/chosen": -0.7532877326011658,
|
| 191 |
+
"rewards/margins": 0.5014825463294983,
|
| 192 |
+
"rewards/rejected": -1.254770278930664,
|
| 193 |
"step": 110
|
| 194 |
},
|
| 195 |
{
|
| 196 |
+
"epoch": 0.14,
|
| 197 |
+
"learning_rate": 4.98121775121344e-07,
|
| 198 |
+
"logits/chosen": -2.4438962936401367,
|
| 199 |
+
"logits/rejected": -2.413790702819824,
|
| 200 |
+
"logps/chosen": -402.3663635253906,
|
| 201 |
+
"logps/rejected": -359.6197509765625,
|
| 202 |
+
"loss": 0.1769,
|
| 203 |
+
"rewards/accuracies": 0.768750011920929,
|
| 204 |
+
"rewards/chosen": -0.6919637322425842,
|
| 205 |
+
"rewards/margins": 0.7776089906692505,
|
| 206 |
+
"rewards/rejected": -1.46957266330719,
|
| 207 |
"step": 120
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.15,
|
| 211 |
+
"learning_rate": 4.96717657955441e-07,
|
| 212 |
+
"logits/chosen": -2.395176649093628,
|
| 213 |
+
"logits/rejected": -2.347350597381592,
|
| 214 |
+
"logps/chosen": -435.8388671875,
|
| 215 |
+
"logps/rejected": -366.80767822265625,
|
| 216 |
+
"loss": 0.157,
|
| 217 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 218 |
+
"rewards/chosen": -0.8537490963935852,
|
| 219 |
+
"rewards/margins": 0.8641785383224487,
|
| 220 |
+
"rewards/rejected": -1.7179279327392578,
|
| 221 |
"step": 130
|
| 222 |
},
|
| 223 |
{
|
| 224 |
+
"epoch": 0.16,
|
| 225 |
+
"learning_rate": 4.949273496411216e-07,
|
| 226 |
+
"logits/chosen": -2.370129346847534,
|
| 227 |
+
"logits/rejected": -2.345968246459961,
|
| 228 |
+
"logps/chosen": -416.231689453125,
|
| 229 |
+
"logps/rejected": -379.8551330566406,
|
| 230 |
+
"loss": 0.1355,
|
| 231 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 232 |
+
"rewards/chosen": -0.9431791305541992,
|
| 233 |
+
"rewards/margins": 0.9395672678947449,
|
| 234 |
+
"rewards/rejected": -1.8827464580535889,
|
| 235 |
"step": 140
|
| 236 |
},
|
| 237 |
{
|
| 238 |
+
"epoch": 0.17,
|
| 239 |
+
"learning_rate": 4.927536525770046e-07,
|
| 240 |
+
"logits/chosen": -2.2446436882019043,
|
| 241 |
+
"logits/rejected": -2.176954507827759,
|
| 242 |
+
"logps/chosen": -439.6024475097656,
|
| 243 |
+
"logps/rejected": -401.255126953125,
|
| 244 |
+
"loss": 0.1276,
|
| 245 |
+
"rewards/accuracies": 0.75,
|
| 246 |
+
"rewards/chosen": -1.096846580505371,
|
| 247 |
+
"rewards/margins": 0.9727560877799988,
|
| 248 |
+
"rewards/rejected": -2.0696027278900146,
|
| 249 |
"step": 150
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.18,
|
| 253 |
+
"learning_rate": 4.901999692863326e-07,
|
| 254 |
+
"logits/chosen": -2.210822582244873,
|
| 255 |
+
"logits/rejected": -2.1662418842315674,
|
| 256 |
+
"logps/chosen": -469.39190673828125,
|
| 257 |
+
"logps/rejected": -381.78662109375,
|
| 258 |
+
"loss": 0.1318,
|
| 259 |
+
"rewards/accuracies": 0.737500011920929,
|
| 260 |
+
"rewards/chosen": -1.149139165878296,
|
| 261 |
+
"rewards/margins": 0.8712302446365356,
|
| 262 |
+
"rewards/rejected": -2.020369291305542,
|
| 263 |
"step": 160
|
| 264 |
},
|
| 265 |
{
|
| 266 |
+
"epoch": 0.19,
|
| 267 |
+
"learning_rate": 4.872702970909464e-07,
|
| 268 |
+
"logits/chosen": -2.2581284046173096,
|
| 269 |
+
"logits/rejected": -2.154148817062378,
|
| 270 |
+
"logps/chosen": -467.570556640625,
|
| 271 |
+
"logps/rejected": -428.92559814453125,
|
| 272 |
+
"loss": 0.1324,
|
| 273 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 274 |
+
"rewards/chosen": -1.0990054607391357,
|
| 275 |
+
"rewards/margins": 1.353104829788208,
|
| 276 |
+
"rewards/rejected": -2.4521100521087646,
|
| 277 |
"step": 170
|
| 278 |
},
|
| 279 |
{
|
| 280 |
+
"epoch": 0.2,
|
| 281 |
+
"learning_rate": 4.839692218542131e-07,
|
| 282 |
+
"logits/chosen": -2.1558468341827393,
|
| 283 |
+
"logits/rejected": -2.092869520187378,
|
| 284 |
+
"logps/chosen": -422.3897399902344,
|
| 285 |
+
"logps/rejected": -405.66558837890625,
|
| 286 |
+
"loss": 0.1499,
|
| 287 |
+
"rewards/accuracies": 0.800000011920929,
|
| 288 |
+
"rewards/chosen": -0.9482347369194031,
|
| 289 |
+
"rewards/margins": 1.135727047920227,
|
| 290 |
+
"rewards/rejected": -2.0839619636535645,
|
| 291 |
"step": 180
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.22,
|
| 295 |
+
"learning_rate": 4.803019108026997e-07,
|
| 296 |
+
"logits/chosen": -2.2160496711730957,
|
| 297 |
+
"logits/rejected": -2.139484405517578,
|
| 298 |
+
"logps/chosen": -448.75732421875,
|
| 299 |
+
"logps/rejected": -376.4322204589844,
|
| 300 |
+
"loss": 0.1547,
|
| 301 |
+
"rewards/accuracies": 0.800000011920929,
|
| 302 |
+
"rewards/chosen": -0.621215283870697,
|
| 303 |
+
"rewards/margins": 1.1775684356689453,
|
| 304 |
+
"rewards/rejected": -1.7987838983535767,
|
| 305 |
"step": 190
|
| 306 |
},
|
| 307 |
{
|
| 308 |
+
"epoch": 0.23,
|
| 309 |
+
"learning_rate": 4.7627410443782887e-07,
|
| 310 |
+
"logits/chosen": -2.189039468765259,
|
| 311 |
+
"logits/rejected": -2.098072052001953,
|
| 312 |
+
"logps/chosen": -424.6959533691406,
|
| 313 |
+
"logps/rejected": -381.8597106933594,
|
| 314 |
+
"loss": 0.1405,
|
| 315 |
+
"rewards/accuracies": 0.737500011920929,
|
| 316 |
+
"rewards/chosen": -1.002191424369812,
|
| 317 |
+
"rewards/margins": 0.9783417582511902,
|
| 318 |
+
"rewards/rejected": -1.980533242225647,
|
| 319 |
"step": 200
|
| 320 |
},
|
| 321 |
{
|
| 322 |
+
"epoch": 0.23,
|
| 323 |
+
"eval_logits/chosen": -2.1455271244049072,
|
| 324 |
+
"eval_logits/rejected": -2.11315655708313,
|
| 325 |
+
"eval_logps/chosen": -420.6826171875,
|
| 326 |
+
"eval_logps/rejected": -474.0294189453125,
|
| 327 |
+
"eval_loss": 0.14227548241615295,
|
| 328 |
+
"eval_rewards/accuracies": 0.70703125,
|
| 329 |
+
"eval_rewards/chosen": -1.6364303827285767,
|
| 330 |
+
"eval_rewards/margins": 0.530331015586853,
|
| 331 |
+
"eval_rewards/rejected": -2.1667611598968506,
|
| 332 |
+
"eval_runtime": 53.4112,
|
| 333 |
+
"eval_samples_per_second": 37.445,
|
| 334 |
+
"eval_steps_per_second": 0.599,
|
| 335 |
+
"step": 200
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.24,
|
| 339 |
+
"learning_rate": 4.7189210755018034e-07,
|
| 340 |
+
"logits/chosen": -2.1224522590637207,
|
| 341 |
+
"logits/rejected": -2.068533182144165,
|
| 342 |
+
"logps/chosen": -491.70062255859375,
|
| 343 |
+
"logps/rejected": -460.06390380859375,
|
| 344 |
+
"loss": 0.1106,
|
| 345 |
+
"rewards/accuracies": 0.768750011920929,
|
| 346 |
+
"rewards/chosen": -1.3628227710723877,
|
| 347 |
+
"rewards/margins": 1.2269973754882812,
|
| 348 |
+
"rewards/rejected": -2.589820146560669,
|
| 349 |
"step": 210
|
| 350 |
},
|
| 351 |
{
|
| 352 |
+
"epoch": 0.25,
|
| 353 |
+
"learning_rate": 4.671627793504988e-07,
|
| 354 |
+
"logits/chosen": -2.1610889434814453,
|
| 355 |
+
"logits/rejected": -2.081235408782959,
|
| 356 |
+
"logps/chosen": -464.84039306640625,
|
| 357 |
+
"logps/rejected": -423.46044921875,
|
| 358 |
+
"loss": 0.0991,
|
| 359 |
+
"rewards/accuracies": 0.8125,
|
| 360 |
+
"rewards/chosen": -1.2738369703292847,
|
| 361 |
+
"rewards/margins": 1.162742018699646,
|
| 362 |
+
"rewards/rejected": -2.4365792274475098,
|
| 363 |
"step": 220
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.26,
|
| 367 |
+
"learning_rate": 4.6209352273286095e-07,
|
| 368 |
+
"logits/chosen": -2.0840981006622314,
|
| 369 |
+
"logits/rejected": -2.009742259979248,
|
| 370 |
+
"logps/chosen": -487.56707763671875,
|
| 371 |
+
"logps/rejected": -457.1775817871094,
|
| 372 |
+
"loss": 0.1024,
|
| 373 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 374 |
+
"rewards/chosen": -1.396468162536621,
|
| 375 |
+
"rewards/margins": 1.1431598663330078,
|
| 376 |
+
"rewards/rejected": -2.539628267288208,
|
| 377 |
"step": 230
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 0.27,
|
| 381 |
+
"learning_rate": 4.56692272686805e-07,
|
| 382 |
+
"logits/chosen": -2.0599420070648193,
|
| 383 |
+
"logits/rejected": -1.9713146686553955,
|
| 384 |
+
"logps/chosen": -526.182373046875,
|
| 385 |
+
"logps/rejected": -502.25628662109375,
|
| 386 |
+
"loss": 0.0906,
|
| 387 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 388 |
+
"rewards/chosen": -1.8474916219711304,
|
| 389 |
+
"rewards/margins": 1.3336760997772217,
|
| 390 |
+
"rewards/rejected": -3.1811680793762207,
|
| 391 |
"step": 240
|
| 392 |
},
|
| 393 |
{
|
| 394 |
+
"epoch": 0.28,
|
| 395 |
+
"learning_rate": 4.5096748387656326e-07,
|
| 396 |
+
"logits/chosen": -2.058642625808716,
|
| 397 |
+
"logits/rejected": -1.9906930923461914,
|
| 398 |
+
"logps/chosen": -461.1634216308594,
|
| 399 |
+
"logps/rejected": -468.45086669921875,
|
| 400 |
+
"loss": 0.0905,
|
| 401 |
+
"rewards/accuracies": 0.75,
|
| 402 |
+
"rewards/chosen": -1.5445082187652588,
|
| 403 |
+
"rewards/margins": 1.150217890739441,
|
| 404 |
+
"rewards/rejected": -2.6947262287139893,
|
| 405 |
"step": 250
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.29,
|
| 409 |
+
"learning_rate": 4.4492811740683877e-07,
|
| 410 |
+
"logits/chosen": -2.1584465503692627,
|
| 411 |
+
"logits/rejected": -2.065124988555908,
|
| 412 |
+
"logps/chosen": -509.6869201660156,
|
| 413 |
+
"logps/rejected": -463.35711669921875,
|
| 414 |
+
"loss": 0.089,
|
| 415 |
+
"rewards/accuracies": 0.800000011920929,
|
| 416 |
+
"rewards/chosen": -1.6028919219970703,
|
| 417 |
+
"rewards/margins": 1.1219018697738647,
|
| 418 |
+
"rewards/rejected": -2.7247939109802246,
|
| 419 |
"step": 260
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 0.31,
|
| 423 |
+
"learning_rate": 4.3858362679584354e-07,
|
| 424 |
+
"logits/chosen": -2.0787439346313477,
|
| 425 |
+
"logits/rejected": -1.9619861841201782,
|
| 426 |
+
"logps/chosen": -476.6595764160156,
|
| 427 |
+
"logps/rejected": -454.09503173828125,
|
| 428 |
+
"loss": 0.0805,
|
| 429 |
+
"rewards/accuracies": 0.75,
|
| 430 |
+
"rewards/chosen": -1.6048253774642944,
|
| 431 |
+
"rewards/margins": 1.1343061923980713,
|
| 432 |
+
"rewards/rejected": -2.739131450653076,
|
| 433 |
"step": 270
|
| 434 |
},
|
| 435 |
{
|
| 436 |
+
"epoch": 0.32,
|
| 437 |
+
"learning_rate": 4.3194394317755245e-07,
|
| 438 |
+
"logits/chosen": -2.058042049407959,
|
| 439 |
+
"logits/rejected": -1.97593092918396,
|
| 440 |
+
"logps/chosen": -452.4337463378906,
|
| 441 |
+
"logps/rejected": -450.2781677246094,
|
| 442 |
+
"loss": 0.0989,
|
| 443 |
+
"rewards/accuracies": 0.731249988079071,
|
| 444 |
+
"rewards/chosen": -1.5429315567016602,
|
| 445 |
+
"rewards/margins": 1.1663429737091064,
|
| 446 |
+
"rewards/rejected": -2.7092742919921875,
|
| 447 |
"step": 280
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.33,
|
| 451 |
+
"learning_rate": 4.2501945975633914e-07,
|
| 452 |
+
"logits/chosen": -2.0520236492156982,
|
| 453 |
+
"logits/rejected": -1.9288972616195679,
|
| 454 |
+
"logps/chosen": -527.3427124023438,
|
| 455 |
+
"logps/rejected": -463.68621826171875,
|
| 456 |
+
"loss": 0.0823,
|
| 457 |
+
"rewards/accuracies": 0.768750011920929,
|
| 458 |
+
"rewards/chosen": -1.7575832605361938,
|
| 459 |
+
"rewards/margins": 1.1834232807159424,
|
| 460 |
+
"rewards/rejected": -2.9410064220428467,
|
| 461 |
"step": 290
|
| 462 |
},
|
| 463 |
{
|
| 464 |
+
"epoch": 0.34,
|
| 465 |
+
"learning_rate": 4.1782101553832405e-07,
|
| 466 |
+
"logits/chosen": -1.9703317880630493,
|
| 467 |
+
"logits/rejected": -1.8722751140594482,
|
| 468 |
+
"logps/chosen": -562.42333984375,
|
| 469 |
+
"logps/rejected": -565.2551879882812,
|
| 470 |
+
"loss": 0.0841,
|
| 471 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 472 |
+
"rewards/chosen": -2.0587549209594727,
|
| 473 |
+
"rewards/margins": 1.456610918045044,
|
| 474 |
+
"rewards/rejected": -3.5153656005859375,
|
| 475 |
"step": 300
|
| 476 |
},
|
| 477 |
{
|
| 478 |
+
"epoch": 0.34,
|
| 479 |
+
"eval_logits/chosen": -2.004652500152588,
|
| 480 |
+
"eval_logits/rejected": -1.9653990268707275,
|
| 481 |
+
"eval_logps/chosen": -485.7205505371094,
|
| 482 |
+
"eval_logps/rejected": -562.75634765625,
|
| 483 |
+
"eval_loss": 0.10304867476224899,
|
| 484 |
+
"eval_rewards/accuracies": 0.73828125,
|
| 485 |
+
"eval_rewards/chosen": -2.286808967590332,
|
| 486 |
+
"eval_rewards/margins": 0.7672209739685059,
|
| 487 |
+
"eval_rewards/rejected": -3.054029941558838,
|
| 488 |
+
"eval_runtime": 53.4196,
|
| 489 |
+
"eval_samples_per_second": 37.439,
|
| 490 |
+
"eval_steps_per_second": 0.599,
|
| 491 |
+
"step": 300
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.35,
|
| 495 |
+
"learning_rate": 4.103598783649029e-07,
|
| 496 |
+
"logits/chosen": -2.0496602058410645,
|
| 497 |
+
"logits/rejected": -1.9651731252670288,
|
| 498 |
+
"logps/chosen": -505.4261169433594,
|
| 499 |
+
"logps/rejected": -485.68310546875,
|
| 500 |
+
"loss": 0.0845,
|
| 501 |
+
"rewards/accuracies": 0.78125,
|
| 502 |
+
"rewards/chosen": -1.727452039718628,
|
| 503 |
+
"rewards/margins": 1.2490873336791992,
|
| 504 |
+
"rewards/rejected": -2.976539134979248,
|
| 505 |
"step": 310
|
| 506 |
},
|
| 507 |
{
|
| 508 |
+
"epoch": 0.36,
|
| 509 |
+
"learning_rate": 4.026477272750119e-07,
|
| 510 |
+
"logits/chosen": -2.020448684692383,
|
| 511 |
+
"logits/rejected": -1.9249324798583984,
|
| 512 |
+
"logps/chosen": -544.9754638671875,
|
| 513 |
+
"logps/rejected": -536.6004028320312,
|
| 514 |
+
"loss": 0.0727,
|
| 515 |
+
"rewards/accuracies": 0.824999988079071,
|
| 516 |
+
"rewards/chosen": -1.8779780864715576,
|
| 517 |
+
"rewards/margins": 1.4059292078018188,
|
| 518 |
+
"rewards/rejected": -3.283907413482666,
|
| 519 |
"step": 320
|
| 520 |
},
|
| 521 |
{
|
| 522 |
+
"epoch": 0.37,
|
| 523 |
+
"learning_rate": 3.9469663422373864e-07,
|
| 524 |
+
"logits/chosen": -2.028433322906494,
|
| 525 |
+
"logits/rejected": -1.9563522338867188,
|
| 526 |
+
"logps/chosen": -511.7374572753906,
|
| 527 |
+
"logps/rejected": -506.9154357910156,
|
| 528 |
+
"loss": 0.0726,
|
| 529 |
+
"rewards/accuracies": 0.762499988079071,
|
| 530 |
+
"rewards/chosen": -2.1345508098602295,
|
| 531 |
+
"rewards/margins": 1.1790883541107178,
|
| 532 |
+
"rewards/rejected": -3.3136394023895264,
|
| 533 |
"step": 330
|
| 534 |
},
|
| 535 |
{
|
| 536 |
+
"epoch": 0.38,
|
| 537 |
+
"learning_rate": 3.865190451858954e-07,
|
| 538 |
+
"logits/chosen": -1.9980262517929077,
|
| 539 |
+
"logits/rejected": -1.9208223819732666,
|
| 540 |
+
"logps/chosen": -555.2127075195312,
|
| 541 |
+
"logps/rejected": -533.1873779296875,
|
| 542 |
+
"loss": 0.0701,
|
| 543 |
+
"rewards/accuracies": 0.793749988079071,
|
| 544 |
+
"rewards/chosen": -1.9280259609222412,
|
| 545 |
+
"rewards/margins": 1.4416617155075073,
|
| 546 |
+
"rewards/rejected": -3.369687557220459,
|
| 547 |
"step": 340
|
| 548 |
},
|
| 549 |
{
|
| 550 |
+
"epoch": 0.4,
|
| 551 |
+
"learning_rate": 3.781277606741327e-07,
|
| 552 |
+
"logits/chosen": -2.0005669593811035,
|
| 553 |
+
"logits/rejected": -1.92441725730896,
|
| 554 |
+
"logps/chosen": -493.4820251464844,
|
| 555 |
+
"logps/rejected": -511.8729553222656,
|
| 556 |
+
"loss": 0.0836,
|
| 557 |
+
"rewards/accuracies": 0.78125,
|
| 558 |
+
"rewards/chosen": -2.063403606414795,
|
| 559 |
+
"rewards/margins": 1.1460031270980835,
|
| 560 |
+
"rewards/rejected": -3.209406614303589,
|
| 561 |
"step": 350
|
| 562 |
},
|
| 563 |
{
|
| 564 |
+
"epoch": 0.41,
|
| 565 |
+
"learning_rate": 3.6953591570208996e-07,
|
| 566 |
+
"logits/chosen": -2.074608325958252,
|
| 567 |
+
"logits/rejected": -1.983313798904419,
|
| 568 |
+
"logps/chosen": -513.7328491210938,
|
| 569 |
+
"logps/rejected": -485.75518798828125,
|
| 570 |
+
"loss": 0.0911,
|
| 571 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 572 |
+
"rewards/chosen": -1.652430772781372,
|
| 573 |
+
"rewards/margins": 1.3198424577713013,
|
| 574 |
+
"rewards/rejected": -2.9722726345062256,
|
| 575 |
"step": 360
|
| 576 |
},
|
| 577 |
{
|
| 578 |
+
"epoch": 0.42,
|
| 579 |
+
"learning_rate": 3.607569592239452e-07,
|
| 580 |
+
"logits/chosen": -1.9775346517562866,
|
| 581 |
+
"logits/rejected": -1.8758798837661743,
|
| 582 |
+
"logps/chosen": -490.39208984375,
|
| 583 |
+
"logps/rejected": -476.412109375,
|
| 584 |
+
"loss": 0.0983,
|
| 585 |
+
"rewards/accuracies": 0.824999988079071,
|
| 586 |
+
"rewards/chosen": -1.5720148086547852,
|
| 587 |
+
"rewards/margins": 1.3585755825042725,
|
| 588 |
+
"rewards/rejected": -2.9305903911590576,
|
| 589 |
"step": 370
|
| 590 |
},
|
| 591 |
{
|
| 592 |
+
"epoch": 0.43,
|
| 593 |
+
"learning_rate": 3.518046330825494e-07,
|
| 594 |
+
"logits/chosen": -1.9943759441375732,
|
| 595 |
+
"logits/rejected": -1.9340169429779053,
|
| 596 |
+
"logps/chosen": -556.3590087890625,
|
| 597 |
+
"logps/rejected": -578.3414306640625,
|
| 598 |
+
"loss": 0.0731,
|
| 599 |
+
"rewards/accuracies": 0.762499988079071,
|
| 600 |
+
"rewards/chosen": -2.216615676879883,
|
| 601 |
+
"rewards/margins": 1.2497565746307373,
|
| 602 |
+
"rewards/rejected": -3.46637225151062,
|
| 603 |
"step": 380
|
| 604 |
},
|
| 605 |
{
|
| 606 |
+
"epoch": 0.44,
|
| 607 |
+
"learning_rate": 3.4269295049909713e-07,
|
| 608 |
+
"logits/chosen": -1.9704225063323975,
|
| 609 |
+
"logits/rejected": -1.9079980850219727,
|
| 610 |
+
"logps/chosen": -492.43743896484375,
|
| 611 |
+
"logps/rejected": -508.32781982421875,
|
| 612 |
+
"loss": 0.0893,
|
| 613 |
+
"rewards/accuracies": 0.75,
|
| 614 |
+
"rewards/chosen": -1.9127576351165771,
|
| 615 |
+
"rewards/margins": 1.1991008520126343,
|
| 616 |
+
"rewards/rejected": -3.111858367919922,
|
| 617 |
"step": 390
|
| 618 |
},
|
| 619 |
{
|
| 620 |
+
"epoch": 0.45,
|
| 621 |
+
"learning_rate": 3.3343617413800453e-07,
|
| 622 |
+
"logits/chosen": -2.0571470260620117,
|
| 623 |
+
"logits/rejected": -1.9935848712921143,
|
| 624 |
+
"logps/chosen": -523.1668090820312,
|
| 625 |
+
"logps/rejected": -481.21844482421875,
|
| 626 |
+
"loss": 0.0916,
|
| 627 |
+
"rewards/accuracies": 0.793749988079071,
|
| 628 |
+
"rewards/chosen": -1.6335818767547607,
|
| 629 |
+
"rewards/margins": 1.325240969657898,
|
| 630 |
+
"rewards/rejected": -2.958822727203369,
|
| 631 |
"step": 400
|
| 632 |
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.45,
|
| 635 |
+
"eval_logits/chosen": -1.9811797142028809,
|
| 636 |
+
"eval_logits/rejected": -1.9393850564956665,
|
| 637 |
+
"eval_logps/chosen": -486.5964660644531,
|
| 638 |
+
"eval_logps/rejected": -566.7339477539062,
|
| 639 |
+
"eval_loss": 0.10801155120134354,
|
| 640 |
+
"eval_rewards/accuracies": 0.734375,
|
| 641 |
+
"eval_rewards/chosen": -2.2955682277679443,
|
| 642 |
+
"eval_rewards/margins": 0.7982385158538818,
|
| 643 |
+
"eval_rewards/rejected": -3.093806743621826,
|
| 644 |
+
"eval_runtime": 53.4276,
|
| 645 |
+
"eval_samples_per_second": 37.434,
|
| 646 |
+
"eval_steps_per_second": 0.599,
|
| 647 |
+
"step": 400
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.46,
|
| 651 |
+
"learning_rate": 3.2404879378132893e-07,
|
| 652 |
+
"logits/chosen": -1.9247922897338867,
|
| 653 |
+
"logits/rejected": -1.856414794921875,
|
| 654 |
+
"logps/chosen": -511.5186462402344,
|
| 655 |
+
"logps/rejected": -560.82177734375,
|
| 656 |
+
"loss": 0.0768,
|
| 657 |
+
"rewards/accuracies": 0.8125,
|
| 658 |
+
"rewards/chosen": -2.140298366546631,
|
| 659 |
+
"rewards/margins": 1.2967740297317505,
|
| 660 |
+
"rewards/rejected": -3.437072277069092,
|
| 661 |
+
"step": 410
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.48,
|
| 665 |
+
"learning_rate": 3.1454550364767894e-07,
|
| 666 |
+
"logits/chosen": -1.8825881481170654,
|
| 667 |
+
"logits/rejected": -1.7626222372055054,
|
| 668 |
+
"logps/chosen": -580.4396362304688,
|
| 669 |
+
"logps/rejected": -558.7975463867188,
|
| 670 |
+
"loss": 0.0629,
|
| 671 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 672 |
+
"rewards/chosen": -2.2527565956115723,
|
| 673 |
+
"rewards/margins": 1.5961545705795288,
|
| 674 |
+
"rewards/rejected": -3.848911762237549,
|
| 675 |
+
"step": 420
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.49,
|
| 679 |
+
"learning_rate": 3.049411793911154e-07,
|
| 680 |
+
"logits/chosen": -1.886850357055664,
|
| 681 |
+
"logits/rejected": -1.7896614074707031,
|
| 682 |
+
"logps/chosen": -581.1607055664062,
|
| 683 |
+
"logps/rejected": -578.7238159179688,
|
| 684 |
+
"loss": 0.074,
|
| 685 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 686 |
+
"rewards/chosen": -2.2639479637145996,
|
| 687 |
+
"rewards/margins": 1.3395296335220337,
|
| 688 |
+
"rewards/rejected": -3.6034774780273438,
|
| 689 |
+
"step": 430
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.5,
|
| 693 |
+
"learning_rate": 2.9525085481604914e-07,
|
| 694 |
+
"logits/chosen": -1.9346811771392822,
|
| 695 |
+
"logits/rejected": -1.8494796752929688,
|
| 696 |
+
"logps/chosen": -539.7020263671875,
|
| 697 |
+
"logps/rejected": -548.3917846679688,
|
| 698 |
+
"loss": 0.0813,
|
| 699 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 700 |
+
"rewards/chosen": -1.9064857959747314,
|
| 701 |
+
"rewards/margins": 1.5988764762878418,
|
| 702 |
+
"rewards/rejected": -3.5053622722625732,
|
| 703 |
+
"step": 440
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.51,
|
| 707 |
+
"learning_rate": 2.854896983445833e-07,
|
| 708 |
+
"logits/chosen": -1.9939569234848022,
|
| 709 |
+
"logits/rejected": -1.8174670934677124,
|
| 710 |
+
"logps/chosen": -547.39111328125,
|
| 711 |
+
"logps/rejected": -516.7273559570312,
|
| 712 |
+
"loss": 0.0755,
|
| 713 |
+
"rewards/accuracies": 0.78125,
|
| 714 |
+
"rewards/chosen": -1.8019745349884033,
|
| 715 |
+
"rewards/margins": 1.4809798002243042,
|
| 716 |
+
"rewards/rejected": -3.282953977584839,
|
| 717 |
+
"step": 450
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.52,
|
| 721 |
+
"learning_rate": 2.7567298927313654e-07,
|
| 722 |
+
"logits/chosen": -1.941044807434082,
|
| 723 |
+
"logits/rejected": -1.8167743682861328,
|
| 724 |
+
"logps/chosen": -517.2515869140625,
|
| 725 |
+
"logps/rejected": -488.5101013183594,
|
| 726 |
+
"loss": 0.0802,
|
| 727 |
+
"rewards/accuracies": 0.800000011920929,
|
| 728 |
+
"rewards/chosen": -1.866037130355835,
|
| 729 |
+
"rewards/margins": 1.3603498935699463,
|
| 730 |
+
"rewards/rejected": -3.2263870239257812,
|
| 731 |
+
"step": 460
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.53,
|
| 735 |
+
"learning_rate": 2.658160938555123e-07,
|
| 736 |
+
"logits/chosen": -1.8934457302093506,
|
| 737 |
+
"logits/rejected": -1.7888898849487305,
|
| 738 |
+
"logps/chosen": -532.8482055664062,
|
| 739 |
+
"logps/rejected": -564.9197998046875,
|
| 740 |
+
"loss": 0.0723,
|
| 741 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 742 |
+
"rewards/chosen": -2.034968852996826,
|
| 743 |
+
"rewards/margins": 1.6071611642837524,
|
| 744 |
+
"rewards/rejected": -3.642129898071289,
|
| 745 |
+
"step": 470
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.54,
|
| 749 |
+
"learning_rate": 2.559344412498532e-07,
|
| 750 |
+
"logits/chosen": -1.9257526397705078,
|
| 751 |
+
"logits/rejected": -1.8294801712036133,
|
| 752 |
+
"logps/chosen": -558.8524169921875,
|
| 753 |
+
"logps/rejected": -521.4874877929688,
|
| 754 |
+
"loss": 0.0725,
|
| 755 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 756 |
+
"rewards/chosen": -2.1087284088134766,
|
| 757 |
+
"rewards/margins": 1.3520355224609375,
|
| 758 |
+
"rewards/rejected": -3.460763454437256,
|
| 759 |
+
"step": 480
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.55,
|
| 763 |
+
"learning_rate": 2.460434993671294e-07,
|
| 764 |
+
"logits/chosen": -1.948136329650879,
|
| 765 |
+
"logits/rejected": -1.8431456089019775,
|
| 766 |
+
"logps/chosen": -519.21728515625,
|
| 767 |
+
"logps/rejected": -500.375,
|
| 768 |
+
"loss": 0.0792,
|
| 769 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 770 |
+
"rewards/chosen": -1.9937260150909424,
|
| 771 |
+
"rewards/margins": 1.335012674331665,
|
| 772 |
+
"rewards/rejected": -3.3287386894226074,
|
| 773 |
+
"step": 490
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.57,
|
| 777 |
+
"learning_rate": 2.361587506589672e-07,
|
| 778 |
+
"logits/chosen": -1.966451644897461,
|
| 779 |
+
"logits/rejected": -1.8704181909561157,
|
| 780 |
+
"logps/chosen": -534.3973388671875,
|
| 781 |
+
"logps/rejected": -500.2060546875,
|
| 782 |
+
"loss": 0.0864,
|
| 783 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 784 |
+
"rewards/chosen": -1.7493633031845093,
|
| 785 |
+
"rewards/margins": 1.3577125072479248,
|
| 786 |
+
"rewards/rejected": -3.1070759296417236,
|
| 787 |
+
"step": 500
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.57,
|
| 791 |
+
"eval_logits/chosen": -1.955956220626831,
|
| 792 |
+
"eval_logits/rejected": -1.9095466136932373,
|
| 793 |
+
"eval_logps/chosen": -488.0345153808594,
|
| 794 |
+
"eval_logps/rejected": -570.5159912109375,
|
| 795 |
+
"eval_loss": 0.09556370228528976,
|
| 796 |
+
"eval_rewards/accuracies": 0.74609375,
|
| 797 |
+
"eval_rewards/chosen": -2.309948682785034,
|
| 798 |
+
"eval_rewards/margins": 0.8216789960861206,
|
| 799 |
+
"eval_rewards/rejected": -3.1316275596618652,
|
| 800 |
+
"eval_runtime": 53.4131,
|
| 801 |
+
"eval_samples_per_second": 37.444,
|
| 802 |
+
"eval_steps_per_second": 0.599,
|
| 803 |
+
"step": 500
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.58,
|
| 807 |
+
"learning_rate": 2.2629566788271613e-07,
|
| 808 |
+
"logits/chosen": -1.9134633541107178,
|
| 809 |
+
"logits/rejected": -1.8345638513565063,
|
| 810 |
+
"logps/chosen": -539.9205932617188,
|
| 811 |
+
"logps/rejected": -500.9922790527344,
|
| 812 |
+
"loss": 0.0694,
|
| 813 |
+
"rewards/accuracies": 0.768750011920929,
|
| 814 |
+
"rewards/chosen": -2.060006618499756,
|
| 815 |
+
"rewards/margins": 1.1839641332626343,
|
| 816 |
+
"rewards/rejected": -3.2439708709716797,
|
| 817 |
+
"step": 510
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.59,
|
| 821 |
+
"learning_rate": 2.1646968988169135e-07,
|
| 822 |
+
"logits/chosen": -1.9419372081756592,
|
| 823 |
+
"logits/rejected": -1.8542063236236572,
|
| 824 |
+
"logps/chosen": -510.229736328125,
|
| 825 |
+
"logps/rejected": -508.4148864746094,
|
| 826 |
+
"loss": 0.0739,
|
| 827 |
+
"rewards/accuracies": 0.831250011920929,
|
| 828 |
+
"rewards/chosen": -1.9082494974136353,
|
| 829 |
+
"rewards/margins": 1.5180120468139648,
|
| 830 |
+
"rewards/rejected": -3.4262614250183105,
|
| 831 |
+
"step": 520
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.6,
|
| 835 |
+
"learning_rate": 2.0669619741850232e-07,
|
| 836 |
+
"logits/chosen": -1.8845760822296143,
|
| 837 |
+
"logits/rejected": -1.7653076648712158,
|
| 838 |
+
"logps/chosen": -569.1829833984375,
|
| 839 |
+
"logps/rejected": -555.857666015625,
|
| 840 |
+
"loss": 0.0678,
|
| 841 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 842 |
+
"rewards/chosen": -2.045509099960327,
|
| 843 |
+
"rewards/margins": 1.7232000827789307,
|
| 844 |
+
"rewards/rejected": -3.768709182739258,
|
| 845 |
+
"step": 530
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.61,
|
| 849 |
+
"learning_rate": 1.9699048909929518e-07,
|
| 850 |
+
"logits/chosen": -1.9015051126480103,
|
| 851 |
+
"logits/rejected": -1.8490610122680664,
|
| 852 |
+
"logps/chosen": -547.8323974609375,
|
| 853 |
+
"logps/rejected": -538.3102416992188,
|
| 854 |
+
"loss": 0.0707,
|
| 855 |
+
"rewards/accuracies": 0.737500011920929,
|
| 856 |
+
"rewards/chosen": -2.1869750022888184,
|
| 857 |
+
"rewards/margins": 1.2785483598709106,
|
| 858 |
+
"rewards/rejected": -3.4655234813690186,
|
| 859 |
+
"step": 540
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.62,
|
| 863 |
+
"learning_rate": 1.8736775742659732e-07,
|
| 864 |
+
"logits/chosen": -1.981376051902771,
|
| 865 |
+
"logits/rejected": -1.8454633951187134,
|
| 866 |
+
"logps/chosen": -589.72509765625,
|
| 867 |
+
"logps/rejected": -503.4346618652344,
|
| 868 |
+
"loss": 0.073,
|
| 869 |
+
"rewards/accuracies": 0.75,
|
| 870 |
+
"rewards/chosen": -1.9054787158966064,
|
| 871 |
+
"rewards/margins": 1.4781739711761475,
|
| 872 |
+
"rewards/rejected": -3.383652925491333,
|
| 873 |
+
"step": 550
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.63,
|
| 877 |
+
"learning_rate": 1.7784306501824616e-07,
|
| 878 |
+
"logits/chosen": -1.968064308166504,
|
| 879 |
+
"logits/rejected": -1.878286361694336,
|
| 880 |
+
"logps/chosen": -508.1175842285156,
|
| 881 |
+
"logps/rejected": -488.2872619628906,
|
| 882 |
+
"loss": 0.0735,
|
| 883 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 884 |
+
"rewards/chosen": -2.008474826812744,
|
| 885 |
+
"rewards/margins": 1.3027125597000122,
|
| 886 |
+
"rewards/rejected": -3.311187267303467,
|
| 887 |
+
"step": 560
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.65,
|
| 891 |
+
"learning_rate": 1.6843132102963025e-07,
|
| 892 |
+
"logits/chosen": -1.9995759725570679,
|
| 893 |
+
"logits/rejected": -1.8707059621810913,
|
| 894 |
+
"logps/chosen": -571.5035400390625,
|
| 895 |
+
"logps/rejected": -551.8418579101562,
|
| 896 |
+
"loss": 0.0724,
|
| 897 |
+
"rewards/accuracies": 0.78125,
|
| 898 |
+
"rewards/chosen": -2.143575429916382,
|
| 899 |
+
"rewards/margins": 1.4678400754928589,
|
| 900 |
+
"rewards/rejected": -3.611415386199951,
|
| 901 |
+
"step": 570
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.66,
|
| 905 |
+
"learning_rate": 1.591472578161458e-07,
|
| 906 |
+
"logits/chosen": -1.9438421726226807,
|
| 907 |
+
"logits/rejected": -1.8859055042266846,
|
| 908 |
+
"logps/chosen": -556.9771728515625,
|
| 909 |
+
"logps/rejected": -608.1251220703125,
|
| 910 |
+
"loss": 0.0776,
|
| 911 |
+
"rewards/accuracies": 0.800000011920929,
|
| 912 |
+
"rewards/chosen": -2.206200122833252,
|
| 913 |
+
"rewards/margins": 1.4931509494781494,
|
| 914 |
+
"rewards/rejected": -3.6993508338928223,
|
| 915 |
+
"step": 580
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.67,
|
| 919 |
+
"learning_rate": 1.5000540787240274e-07,
|
| 920 |
+
"logits/chosen": -1.992211937904358,
|
| 921 |
+
"logits/rejected": -1.881553053855896,
|
| 922 |
+
"logps/chosen": -552.2855224609375,
|
| 923 |
+
"logps/rejected": -532.5646362304688,
|
| 924 |
+
"loss": 0.0643,
|
| 925 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 926 |
+
"rewards/chosen": -2.1940808296203613,
|
| 927 |
+
"rewards/margins": 1.3729619979858398,
|
| 928 |
+
"rewards/rejected": -3.567042589187622,
|
| 929 |
+
"step": 590
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.68,
|
| 933 |
+
"learning_rate": 1.410200810842749e-07,
|
| 934 |
+
"logits/chosen": -1.9430984258651733,
|
| 935 |
+
"logits/rejected": -1.8172378540039062,
|
| 936 |
+
"logps/chosen": -613.967529296875,
|
| 937 |
+
"logps/rejected": -598.5076904296875,
|
| 938 |
+
"loss": 0.065,
|
| 939 |
+
"rewards/accuracies": 0.800000011920929,
|
| 940 |
+
"rewards/chosen": -2.2278945446014404,
|
| 941 |
+
"rewards/margins": 1.6444116830825806,
|
| 942 |
+
"rewards/rejected": -3.8723063468933105,
|
| 943 |
+
"step": 600
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.68,
|
| 947 |
+
"eval_logits/chosen": -1.9405471086502075,
|
| 948 |
+
"eval_logits/rejected": -1.886796236038208,
|
| 949 |
+
"eval_logps/chosen": -542.6825561523438,
|
| 950 |
+
"eval_logps/rejected": -633.113525390625,
|
| 951 |
+
"eval_loss": 0.08488854765892029,
|
| 952 |
+
"eval_rewards/accuracies": 0.7265625,
|
| 953 |
+
"eval_rewards/chosen": -2.856428861618042,
|
| 954 |
+
"eval_rewards/margins": 0.9011733531951904,
|
| 955 |
+
"eval_rewards/rejected": -3.7576024532318115,
|
| 956 |
+
"eval_runtime": 53.397,
|
| 957 |
+
"eval_samples_per_second": 37.455,
|
| 958 |
+
"eval_steps_per_second": 0.599,
|
| 959 |
+
"step": 600
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.69,
|
| 963 |
+
"learning_rate": 1.322053423294041e-07,
|
| 964 |
+
"logits/chosen": -1.989729881286621,
|
| 965 |
+
"logits/rejected": -1.911266565322876,
|
| 966 |
+
"logps/chosen": -580.957275390625,
|
| 967 |
+
"logps/rejected": -575.7573852539062,
|
| 968 |
+
"loss": 0.0664,
|
| 969 |
+
"rewards/accuracies": 0.762499988079071,
|
| 970 |
+
"rewards/chosen": -2.4196839332580566,
|
| 971 |
+
"rewards/margins": 1.2277535200119019,
|
| 972 |
+
"rewards/rejected": -3.647437334060669,
|
| 973 |
+
"step": 610
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.7,
|
| 977 |
+
"learning_rate": 1.2357498946121905e-07,
|
| 978 |
+
"logits/chosen": -1.930645227432251,
|
| 979 |
+
"logits/rejected": -1.8115425109863281,
|
| 980 |
+
"logps/chosen": -580.1464233398438,
|
| 981 |
+
"logps/rejected": -578.2677612304688,
|
| 982 |
+
"loss": 0.0632,
|
| 983 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 984 |
+
"rewards/chosen": -2.2879867553710938,
|
| 985 |
+
"rewards/margins": 1.716691017150879,
|
| 986 |
+
"rewards/rejected": -4.004677772521973,
|
| 987 |
+
"step": 620
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.71,
|
| 991 |
+
"learning_rate": 1.1514253171093161e-07,
|
| 992 |
+
"logits/chosen": -1.9917194843292236,
|
| 993 |
+
"logits/rejected": -1.8586351871490479,
|
| 994 |
+
"logps/chosen": -593.2454833984375,
|
| 995 |
+
"logps/rejected": -554.43115234375,
|
| 996 |
+
"loss": 0.0681,
|
| 997 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 998 |
+
"rewards/chosen": -2.304713726043701,
|
| 999 |
+
"rewards/margins": 1.4881376028060913,
|
| 1000 |
+
"rewards/rejected": -3.792851209640503,
|
| 1001 |
+
"step": 630
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 0.72,
|
| 1005 |
+
"learning_rate": 1.0692116854131883e-07,
|
| 1006 |
+
"logits/chosen": -1.963966727256775,
|
| 1007 |
+
"logits/rejected": -1.876705527305603,
|
| 1008 |
+
"logps/chosen": -572.40625,
|
| 1009 |
+
"logps/rejected": -571.9288940429688,
|
| 1010 |
+
"loss": 0.07,
|
| 1011 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1012 |
+
"rewards/chosen": -2.363751173019409,
|
| 1013 |
+
"rewards/margins": 1.4736993312835693,
|
| 1014 |
+
"rewards/rejected": -3.8374505043029785,
|
| 1015 |
+
"step": 640
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.74,
|
| 1019 |
+
"learning_rate": 9.89237689853889e-08,
|
| 1020 |
+
"logits/chosen": -1.9835160970687866,
|
| 1021 |
+
"logits/rejected": -1.9168879985809326,
|
| 1022 |
+
"logps/chosen": -562.7428588867188,
|
| 1023 |
+
"logps/rejected": -593.8363037109375,
|
| 1024 |
+
"loss": 0.0689,
|
| 1025 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1026 |
+
"rewards/chosen": -2.1839652061462402,
|
| 1027 |
+
"rewards/margins": 1.3724231719970703,
|
| 1028 |
+
"rewards/rejected": -3.5563888549804688,
|
| 1029 |
+
"step": 650
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.75,
|
| 1033 |
+
"learning_rate": 9.11628515022765e-08,
|
| 1034 |
+
"logits/chosen": -2.0633797645568848,
|
| 1035 |
+
"logits/rejected": -1.9203882217407227,
|
| 1036 |
+
"logps/chosen": -586.35009765625,
|
| 1037 |
+
"logps/rejected": -529.8253173828125,
|
| 1038 |
+
"loss": 0.0734,
|
| 1039 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1040 |
+
"rewards/chosen": -2.0526123046875,
|
| 1041 |
+
"rewards/margins": 1.624458909034729,
|
| 1042 |
+
"rewards/rejected": -3.6770706176757812,
|
| 1043 |
+
"step": 660
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.76,
|
| 1047 |
+
"learning_rate": 8.365056438189486e-08,
|
| 1048 |
+
"logits/chosen": -2.0234529972076416,
|
| 1049 |
+
"logits/rejected": -1.9365705251693726,
|
| 1050 |
+
"logps/chosen": -527.53662109375,
|
| 1051 |
+
"logps/rejected": -542.6231689453125,
|
| 1052 |
+
"loss": 0.0723,
|
| 1053 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1054 |
+
"rewards/chosen": -2.2389750480651855,
|
| 1055 |
+
"rewards/margins": 1.2195186614990234,
|
| 1056 |
+
"rewards/rejected": -3.45849347114563,
|
| 1057 |
+
"step": 670
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.77,
|
| 1061 |
+
"learning_rate": 7.639866672902101e-08,
|
| 1062 |
+
"logits/chosen": -1.9303470849990845,
|
| 1063 |
+
"logits/rejected": -1.8580068349838257,
|
| 1064 |
+
"logps/chosen": -554.188720703125,
|
| 1065 |
+
"logps/rejected": -569.6090698242188,
|
| 1066 |
+
"loss": 0.0703,
|
| 1067 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1068 |
+
"rewards/chosen": -2.2495415210723877,
|
| 1069 |
+
"rewards/margins": 1.5471585988998413,
|
| 1070 |
+
"rewards/rejected": -3.7967002391815186,
|
| 1071 |
+
"step": 680
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.78,
|
| 1075 |
+
"learning_rate": 6.941851005657851e-08,
|
| 1076 |
+
"logits/chosen": -2.0182881355285645,
|
| 1077 |
+
"logits/rejected": -1.9131933450698853,
|
| 1078 |
+
"logps/chosen": -571.5795288085938,
|
| 1079 |
+
"logps/rejected": -563.0739135742188,
|
| 1080 |
+
"loss": 0.0668,
|
| 1081 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1082 |
+
"rewards/chosen": -2.189399242401123,
|
| 1083 |
+
"rewards/margins": 1.6238410472869873,
|
| 1084 |
+
"rewards/rejected": -3.8132405281066895,
|
| 1085 |
+
"step": 690
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.79,
|
| 1089 |
+
"learning_rate": 6.272102051693051e-08,
|
| 1090 |
+
"logits/chosen": -1.9295063018798828,
|
| 1091 |
+
"logits/rejected": -1.8629966974258423,
|
| 1092 |
+
"logps/chosen": -551.6529541015625,
|
| 1093 |
+
"logps/rejected": -583.9489135742188,
|
| 1094 |
+
"loss": 0.0663,
|
| 1095 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 1096 |
+
"rewards/chosen": -2.4394476413726807,
|
| 1097 |
+
"rewards/margins": 1.471626877784729,
|
| 1098 |
+
"rewards/rejected": -3.91107439994812,
|
| 1099 |
+
"step": 700
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.79,
|
| 1103 |
+
"eval_logits/chosen": -1.962702751159668,
|
| 1104 |
+
"eval_logits/rejected": -1.9099680185317993,
|
| 1105 |
+
"eval_logps/chosen": -538.3110961914062,
|
| 1106 |
+
"eval_logps/rejected": -630.755615234375,
|
| 1107 |
+
"eval_loss": 0.08400725573301315,
|
| 1108 |
+
"eval_rewards/accuracies": 0.73828125,
|
| 1109 |
+
"eval_rewards/chosen": -2.812714099884033,
|
| 1110 |
+
"eval_rewards/margins": 0.921308696269989,
|
| 1111 |
+
"eval_rewards/rejected": -3.734023094177246,
|
| 1112 |
+
"eval_runtime": 53.3939,
|
| 1113 |
+
"eval_samples_per_second": 37.457,
|
| 1114 |
+
"eval_steps_per_second": 0.599,
|
| 1115 |
+
"step": 700
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.8,
|
| 1119 |
+
"learning_rate": 5.6316681798995844e-08,
|
| 1120 |
+
"logits/chosen": -1.998659372329712,
|
| 1121 |
+
"logits/rejected": -1.9016752243041992,
|
| 1122 |
+
"logps/chosen": -596.6257934570312,
|
| 1123 |
+
"logps/rejected": -589.1134033203125,
|
| 1124 |
+
"loss": 0.0751,
|
| 1125 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1126 |
+
"rewards/chosen": -2.465219020843506,
|
| 1127 |
+
"rewards/margins": 1.4250514507293701,
|
| 1128 |
+
"rewards/rejected": -3.890270948410034,
|
| 1129 |
+
"step": 710
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.81,
|
| 1133 |
+
"learning_rate": 5.0215518717961256e-08,
|
| 1134 |
+
"logits/chosen": -1.9532493352890015,
|
| 1135 |
+
"logits/rejected": -1.861000418663025,
|
| 1136 |
+
"logps/chosen": -530.2694091796875,
|
| 1137 |
+
"logps/rejected": -542.3233642578125,
|
| 1138 |
+
"loss": 0.0666,
|
| 1139 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1140 |
+
"rewards/chosen": -2.3267478942871094,
|
| 1141 |
+
"rewards/margins": 1.403226613998413,
|
| 1142 |
+
"rewards/rejected": -3.7299742698669434,
|
| 1143 |
+
"step": 720
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.83,
|
| 1147 |
+
"learning_rate": 4.4427081523275925e-08,
|
| 1148 |
+
"logits/chosen": -1.9771426916122437,
|
| 1149 |
+
"logits/rejected": -1.8616434335708618,
|
| 1150 |
+
"logps/chosen": -516.3856201171875,
|
| 1151 |
+
"logps/rejected": -545.8884887695312,
|
| 1152 |
+
"loss": 0.0705,
|
| 1153 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1154 |
+
"rewards/chosen": -2.18568754196167,
|
| 1155 |
+
"rewards/margins": 1.4953739643096924,
|
| 1156 |
+
"rewards/rejected": -3.6810615062713623,
|
| 1157 |
+
"step": 730
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.84,
|
| 1161 |
+
"learning_rate": 3.896043094949061e-08,
|
| 1162 |
+
"logits/chosen": -1.9744873046875,
|
| 1163 |
+
"logits/rejected": -1.8698228597640991,
|
| 1164 |
+
"logps/chosen": -564.7574462890625,
|
| 1165 |
+
"logps/rejected": -554.6087036132812,
|
| 1166 |
+
"loss": 0.0707,
|
| 1167 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1168 |
+
"rewards/chosen": -2.351306438446045,
|
| 1169 |
+
"rewards/margins": 1.3112602233886719,
|
| 1170 |
+
"rewards/rejected": -3.662567138671875,
|
| 1171 |
+
"step": 740
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.85,
|
| 1175 |
+
"learning_rate": 3.3824124033343557e-08,
|
| 1176 |
+
"logits/chosen": -1.971381425857544,
|
| 1177 |
+
"logits/rejected": -1.8558752536773682,
|
| 1178 |
+
"logps/chosen": -550.445556640625,
|
| 1179 |
+
"logps/rejected": -536.169921875,
|
| 1180 |
+
"loss": 0.0663,
|
| 1181 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1182 |
+
"rewards/chosen": -2.283639430999756,
|
| 1183 |
+
"rewards/margins": 1.291291356086731,
|
| 1184 |
+
"rewards/rejected": -3.5749306678771973,
|
| 1185 |
+
"step": 750
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.86,
|
| 1189 |
+
"learning_rate": 2.9026200719291904e-08,
|
| 1190 |
+
"logits/chosen": -1.8978159427642822,
|
| 1191 |
+
"logits/rejected": -1.8259022235870361,
|
| 1192 |
+
"logps/chosen": -486.37152099609375,
|
| 1193 |
+
"logps/rejected": -516.701416015625,
|
| 1194 |
+
"loss": 0.0672,
|
| 1195 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1196 |
+
"rewards/chosen": -2.1406826972961426,
|
| 1197 |
+
"rewards/margins": 1.2746713161468506,
|
| 1198 |
+
"rewards/rejected": -3.4153542518615723,
|
| 1199 |
+
"step": 760
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.87,
|
| 1203 |
+
"learning_rate": 2.4574171274456433e-08,
|
| 1204 |
+
"logits/chosen": -1.9380009174346924,
|
| 1205 |
+
"logits/rejected": -1.7994283437728882,
|
| 1206 |
+
"logps/chosen": -570.5930786132812,
|
| 1207 |
+
"logps/rejected": -562.5328369140625,
|
| 1208 |
+
"loss": 0.0668,
|
| 1209 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1210 |
+
"rewards/chosen": -2.2663681507110596,
|
| 1211 |
+
"rewards/margins": 1.4617531299591064,
|
| 1212 |
+
"rewards/rejected": -3.728121519088745,
|
| 1213 |
+
"step": 770
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.88,
|
| 1217 |
+
"learning_rate": 2.047500453267881e-08,
|
| 1218 |
+
"logits/chosen": -1.9811604022979736,
|
| 1219 |
+
"logits/rejected": -1.8174915313720703,
|
| 1220 |
+
"logps/chosen": -585.8560180664062,
|
| 1221 |
+
"logps/rejected": -575.7230834960938,
|
| 1222 |
+
"loss": 0.0723,
|
| 1223 |
+
"rewards/accuracies": 0.78125,
|
| 1224 |
+
"rewards/chosen": -2.2959883213043213,
|
| 1225 |
+
"rewards/margins": 1.562656283378601,
|
| 1226 |
+
"rewards/rejected": -3.858644485473633,
|
| 1227 |
+
"step": 780
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.89,
|
| 1231 |
+
"learning_rate": 1.673511698609292e-08,
|
| 1232 |
+
"logits/chosen": -1.9835258722305298,
|
| 1233 |
+
"logits/rejected": -1.8875339031219482,
|
| 1234 |
+
"logps/chosen": -577.6177978515625,
|
| 1235 |
+
"logps/rejected": -575.160400390625,
|
| 1236 |
+
"loss": 0.074,
|
| 1237 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1238 |
+
"rewards/chosen": -2.322665214538574,
|
| 1239 |
+
"rewards/margins": 1.4801914691925049,
|
| 1240 |
+
"rewards/rejected": -3.8028564453125,
|
| 1241 |
+
"step": 790
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.91,
|
| 1245 |
+
"learning_rate": 1.3360362741285769e-08,
|
| 1246 |
+
"logits/chosen": -1.877092719078064,
|
| 1247 |
+
"logits/rejected": -1.8041622638702393,
|
| 1248 |
+
"logps/chosen": -503.6748962402344,
|
| 1249 |
+
"logps/rejected": -512.8577270507812,
|
| 1250 |
+
"loss": 0.0663,
|
| 1251 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1252 |
+
"rewards/chosen": -2.3050992488861084,
|
| 1253 |
+
"rewards/margins": 1.193604826927185,
|
| 1254 |
+
"rewards/rejected": -3.498703718185425,
|
| 1255 |
+
"step": 800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.91,
|
| 1259 |
+
"eval_logits/chosen": -1.9548935890197754,
|
| 1260 |
+
"eval_logits/rejected": -1.901588797569275,
|
| 1261 |
+
"eval_logps/chosen": -527.3272705078125,
|
| 1262 |
+
"eval_logps/rejected": -619.4178466796875,
|
| 1263 |
+
"eval_loss": 0.0869935154914856,
|
| 1264 |
+
"eval_rewards/accuracies": 0.74609375,
|
| 1265 |
+
"eval_rewards/chosen": -2.702876567840576,
|
| 1266 |
+
"eval_rewards/margins": 0.9177693128585815,
|
| 1267 |
+
"eval_rewards/rejected": -3.620645761489868,
|
| 1268 |
+
"eval_runtime": 53.3906,
|
| 1269 |
+
"eval_samples_per_second": 37.46,
|
| 1270 |
+
"eval_steps_per_second": 0.599,
|
| 1271 |
+
"step": 800
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.92,
|
| 1275 |
+
"learning_rate": 1.0356024355769433e-08,
|
| 1276 |
+
"logits/chosen": -1.9531776905059814,
|
| 1277 |
+
"logits/rejected": -1.8769395351409912,
|
| 1278 |
+
"logps/chosen": -578.2391357421875,
|
| 1279 |
+
"logps/rejected": -570.4874267578125,
|
| 1280 |
+
"loss": 0.0724,
|
| 1281 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1282 |
+
"rewards/chosen": -2.243154287338257,
|
| 1283 |
+
"rewards/margins": 1.2938239574432373,
|
| 1284 |
+
"rewards/rejected": -3.5369784832000732,
|
| 1285 |
+
"step": 810
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.93,
|
| 1289 |
+
"learning_rate": 7.726804569108597e-09,
|
| 1290 |
+
"logits/chosen": -1.908215880393982,
|
| 1291 |
+
"logits/rejected": -1.777242660522461,
|
| 1292 |
+
"logps/chosen": -549.6116333007812,
|
| 1293 |
+
"logps/rejected": -539.8973999023438,
|
| 1294 |
+
"loss": 0.0635,
|
| 1295 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1296 |
+
"rewards/chosen": -2.2482357025146484,
|
| 1297 |
+
"rewards/margins": 1.3119279146194458,
|
| 1298 |
+
"rewards/rejected": -3.5601630210876465,
|
| 1299 |
+
"step": 820
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.94,
|
| 1303 |
+
"learning_rate": 5.476818941645561e-09,
|
| 1304 |
+
"logits/chosen": -1.9380836486816406,
|
| 1305 |
+
"logits/rejected": -1.8248558044433594,
|
| 1306 |
+
"logps/chosen": -566.2852172851562,
|
| 1307 |
+
"logps/rejected": -571.448974609375,
|
| 1308 |
+
"loss": 0.0626,
|
| 1309 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1310 |
+
"rewards/chosen": -2.232658863067627,
|
| 1311 |
+
"rewards/margins": 1.5336438417434692,
|
| 1312 |
+
"rewards/rejected": -3.7663028240203857,
|
| 1313 |
+
"step": 830
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.95,
|
| 1317 |
+
"learning_rate": 3.609589412347347e-09,
|
| 1318 |
+
"logits/chosen": -1.99057137966156,
|
| 1319 |
+
"logits/rejected": -1.887397050857544,
|
| 1320 |
+
"logps/chosen": -581.9739990234375,
|
| 1321 |
+
"logps/rejected": -585.4699096679688,
|
| 1322 |
+
"loss": 0.0769,
|
| 1323 |
+
"rewards/accuracies": 0.8125,
|
| 1324 |
+
"rewards/chosen": -2.106590747833252,
|
| 1325 |
+
"rewards/margins": 1.779883623123169,
|
| 1326 |
+
"rewards/rejected": -3.886474132537842,
|
| 1327 |
+
"step": 840
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.96,
|
| 1331 |
+
"learning_rate": 2.1280387858572667e-09,
|
| 1332 |
+
"logits/chosen": -1.925265908241272,
|
| 1333 |
+
"logits/rejected": -1.8459047079086304,
|
| 1334 |
+
"logps/chosen": -552.0723876953125,
|
| 1335 |
+
"logps/rejected": -599.8142700195312,
|
| 1336 |
+
"loss": 0.0739,
|
| 1337 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 1338 |
+
"rewards/chosen": -2.3117988109588623,
|
| 1339 |
+
"rewards/margins": 1.3644450902938843,
|
| 1340 |
+
"rewards/rejected": -3.676244020462036,
|
| 1341 |
+
"step": 850
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.97,
|
| 1345 |
+
"learning_rate": 1.03448615738172e-09,
|
| 1346 |
+
"logits/chosen": -1.9360536336898804,
|
| 1347 |
+
"logits/rejected": -1.8282134532928467,
|
| 1348 |
+
"logps/chosen": -535.2977294921875,
|
| 1349 |
+
"logps/rejected": -555.2935791015625,
|
| 1350 |
+
"loss": 0.0656,
|
| 1351 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 1352 |
+
"rewards/chosen": -2.1976685523986816,
|
| 1353 |
+
"rewards/margins": 1.5509238243103027,
|
| 1354 |
+
"rewards/rejected": -3.748591899871826,
|
| 1355 |
+
"step": 860
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.98,
|
| 1359 |
+
"learning_rate": 3.3064328257259575e-10,
|
| 1360 |
+
"logits/chosen": -1.9918015003204346,
|
| 1361 |
+
"logits/rejected": -1.8840529918670654,
|
| 1362 |
+
"logps/chosen": -587.255615234375,
|
| 1363 |
+
"logps/rejected": -566.5502319335938,
|
| 1364 |
+
"loss": 0.0679,
|
| 1365 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1366 |
+
"rewards/chosen": -2.329481840133667,
|
| 1367 |
+
"rewards/margins": 1.393025279045105,
|
| 1368 |
+
"rewards/rejected": -3.7225069999694824,
|
| 1369 |
+
"step": 870
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.0,
|
| 1373 |
+
"learning_rate": 1.7611898088715216e-11,
|
| 1374 |
+
"logits/chosen": -2.004169225692749,
|
| 1375 |
+
"logits/rejected": -1.8871746063232422,
|
| 1376 |
+
"logps/chosen": -655.3201293945312,
|
| 1377 |
+
"logps/rejected": -616.4334106445312,
|
| 1378 |
+
"loss": 0.0723,
|
| 1379 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1380 |
+
"rewards/chosen": -2.3969874382019043,
|
| 1381 |
+
"rewards/margins": 1.6344892978668213,
|
| 1382 |
+
"rewards/rejected": -4.031477451324463,
|
| 1383 |
+
"step": 880
|
| 1384 |
+
},
|
| 1385 |
{
|
| 1386 |
"epoch": 1.0,
|
| 1387 |
+
"step": 883,
|
| 1388 |
"total_flos": 0.0,
|
| 1389 |
+
"train_loss": 0.11349766382510908,
|
| 1390 |
+
"train_runtime": 8005.048,
|
| 1391 |
+
"train_samples_per_second": 14.12,
|
| 1392 |
+
"train_steps_per_second": 0.11
|
| 1393 |
}
|
| 1394 |
],
|
| 1395 |
"logging_steps": 10,
|
| 1396 |
+
"max_steps": 883,
|
| 1397 |
"num_train_epochs": 1,
|
| 1398 |
+
"save_steps": 100,
|
| 1399 |
"total_flos": 0.0,
|
| 1400 |
"trial_name": null,
|
| 1401 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5944
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:34facb59b2833ff9a65d1ea6ca0671f7143189081be77d079ad67a7343d5aa7d
|
| 3 |
size 5944
|