Model save
Browse files- README.md +24 -3
- all_results.json +5 -5
- model-00001-of-00004.safetensors +1 -1
- model-00002-of-00004.safetensors +1 -1
- model-00003-of-00004.safetensors +1 -1
- model-00004-of-00004.safetensors +1 -1
- train_results.json +5 -5
- trainer_state.json +1265 -785
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -14,6 +14,16 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
# zephyr-7b-dpo-full
|
| 15 |
|
| 16 |
This model was trained from scratch on the None dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
## Model description
|
| 19 |
|
|
@@ -33,12 +43,12 @@ More information needed
|
|
| 33 |
|
| 34 |
The following hyperparameters were used during training:
|
| 35 |
- learning_rate: 1e-06
|
| 36 |
-
- train_batch_size:
|
| 37 |
- eval_batch_size: 8
|
| 38 |
-
- seed:
|
| 39 |
- distributed_type: multi-GPU
|
| 40 |
- num_devices: 8
|
| 41 |
-
- gradient_accumulation_steps:
|
| 42 |
- total_train_batch_size: 128
|
| 43 |
- total_eval_batch_size: 64
|
| 44 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
|
@@ -48,6 +58,17 @@ The following hyperparameters were used during training:
|
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 51 |
|
| 52 |
|
| 53 |
### Framework versions
|
|
|
|
| 14 |
# zephyr-7b-dpo-full
|
| 15 |
|
| 16 |
This model was trained from scratch on the None dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.0216
|
| 19 |
+
- Rewards/chosen: -2.0018
|
| 20 |
+
- Rewards/rejected: -3.2655
|
| 21 |
+
- Rewards/accuracies: 0.7344
|
| 22 |
+
- Rewards/margins: 1.2637
|
| 23 |
+
- Logps/rejected: -638.1763
|
| 24 |
+
- Logps/chosen: -503.7670
|
| 25 |
+
- Logits/rejected: 0.7139
|
| 26 |
+
- Logits/chosen: 0.6104
|
| 27 |
|
| 28 |
## Model description
|
| 29 |
|
|
|
|
| 43 |
|
| 44 |
The following hyperparameters were used during training:
|
| 45 |
- learning_rate: 1e-06
|
| 46 |
+
- train_batch_size: 4
|
| 47 |
- eval_batch_size: 8
|
| 48 |
+
- seed: 2
|
| 49 |
- distributed_type: multi-GPU
|
| 50 |
- num_devices: 8
|
| 51 |
+
- gradient_accumulation_steps: 4
|
| 52 |
- total_train_batch_size: 128
|
| 53 |
- total_eval_batch_size: 64
|
| 54 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
|
|
|
| 58 |
|
| 59 |
### Training results
|
| 60 |
|
| 61 |
+
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|
| 62 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
|
| 63 |
+
| 0.1135 | 0.21 | 100 | 0.1060 | -0.4112 | -0.7633 | 0.7070 | 0.3521 | -387.9543 | -344.7048 | 0.5669 | 0.5124 |
|
| 64 |
+
| 0.0648 | 0.42 | 200 | 0.0664 | -0.6849 | -1.3101 | 0.7188 | 0.6252 | -442.6332 | -372.0718 | 0.3867 | 0.3105 |
|
| 65 |
+
| 0.0565 | 0.63 | 300 | 0.0585 | -0.7598 | -1.4962 | 0.7305 | 0.7364 | -461.2486 | -379.5646 | 0.3625 | 0.2834 |
|
| 66 |
+
| 0.0643 | 0.84 | 400 | 0.0623 | -0.6660 | -1.4002 | 0.7305 | 0.7341 | -451.6428 | -370.1878 | 0.2607 | 0.1862 |
|
| 67 |
+
| 0.0214 | 1.05 | 500 | 0.0310 | -1.5688 | -2.7542 | 0.7422 | 1.1854 | -587.0469 | -460.4668 | 0.5000 | 0.4214 |
|
| 68 |
+
| 0.0145 | 1.26 | 600 | 0.0245 | -1.7918 | -3.0398 | 0.7344 | 1.2481 | -615.6083 | -482.7609 | 0.6354 | 0.5416 |
|
| 69 |
+
| 0.0098 | 1.47 | 700 | 0.0210 | -2.0363 | -3.3488 | 0.7305 | 1.3125 | -646.5033 | -507.2101 | 0.6871 | 0.5851 |
|
| 70 |
+
| 0.0101 | 1.67 | 800 | 0.0218 | -1.9924 | -3.2641 | 0.7344 | 1.2718 | -638.0398 | -502.8234 | 0.6981 | 0.5937 |
|
| 71 |
+
| 0.0103 | 1.88 | 900 | 0.0216 | -2.0018 | -3.2655 | 0.7344 | 1.2637 | -638.1763 | -503.7670 | 0.7139 | 0.6104 |
|
| 72 |
|
| 73 |
|
| 74 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 8
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"train_loss": 0.048426837674039463,
|
| 4 |
+
"train_runtime": 8860.0868,
|
| 5 |
+
"train_samples": 61134,
|
| 6 |
+
"train_samples_per_second": 13.8,
|
| 7 |
+
"train_steps_per_second": 0.108
|
| 8 |
}
|
model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4976698672
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bbf762dfc95b30382f7d12aa322403420f2fedeb1f40de707e07a41b4b65a74
|
| 3 |
size 4976698672
|
model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4999802720
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b360ad3bfe2c13ef13d8ec3499feb010e9e772efbcd73344a4fa33c853c08df
|
| 3 |
size 4999802720
|
model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4915916176
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:611e09fdf80d346d40d4790f93b01096529390ec9de8882089baa707b39f1a31
|
| 3 |
size 4915916176
|
model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1168138808
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c34cb68b36e33fe3b029cc36b78c1729aa55f897d89fb8a7def6c059f706b100
|
| 3 |
size 1168138808
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second": 8
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 2.0,
|
| 3 |
+
"train_loss": 0.048426837674039463,
|
| 4 |
+
"train_runtime": 8860.0868,
|
| 5 |
+
"train_samples": 61134,
|
| 6 |
+
"train_samples_per_second": 13.8,
|
| 7 |
+
"train_steps_per_second": 0.108
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,1021 +1,1501 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 1.
|
| 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.
|
| 13 |
-
"learning_rate": 1.
|
| 14 |
-
"logits/chosen": 0.
|
| 15 |
-
"logits/rejected": 0.
|
| 16 |
-
"logps/chosen": -
|
| 17 |
-
"logps/rejected": -
|
| 18 |
-
"loss": 0.
|
| 19 |
-
"rewards/accuracies": 0.
|
| 20 |
-
"rewards/chosen":
|
| 21 |
-
"rewards/margins": 0.
|
| 22 |
-
"rewards/rejected": -0.
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"learning_rate": 2.
|
| 28 |
-
"logits/chosen":
|
| 29 |
-
"logits/rejected": 0.
|
| 30 |
-
"logps/chosen": -
|
| 31 |
-
"logps/rejected": -
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"rewards/accuracies": 0.
|
| 34 |
-
"rewards/chosen": 0.
|
| 35 |
-
"rewards/margins": 0.
|
| 36 |
-
"rewards/rejected": -0.
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"learning_rate":
|
| 42 |
-
"logits/chosen": 0.
|
| 43 |
-
"logits/rejected": 0.
|
| 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": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"logits/chosen": 0.
|
| 57 |
-
"logits/rejected": 0.
|
| 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": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch": 0.
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"logits/chosen": 0.
|
| 71 |
-
"logits/rejected": 0.
|
| 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": 50
|
| 80 |
},
|
| 81 |
{
|
| 82 |
-
"epoch": 0.
|
| 83 |
-
"learning_rate":
|
| 84 |
-
"logits/chosen": 0.
|
| 85 |
-
"logits/rejected": 0.
|
| 86 |
-
"logps/chosen": -
|
| 87 |
-
"logps/rejected": -
|
| 88 |
-
"loss": 0.
|
| 89 |
-
"rewards/accuracies": 0.
|
| 90 |
-
"rewards/chosen": -0.
|
| 91 |
-
"rewards/margins": 0.
|
| 92 |
-
"rewards/rejected": -0.
|
| 93 |
"step": 60
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"learning_rate":
|
| 98 |
-
"logits/chosen":
|
| 99 |
-
"logits/rejected": 0.
|
| 100 |
-
"logps/chosen": -
|
| 101 |
-
"logps/rejected": -
|
| 102 |
-
"loss": 0.
|
| 103 |
-
"rewards/accuracies": 0.
|
| 104 |
-
"rewards/chosen": -0.
|
| 105 |
-
"rewards/margins": 0.
|
| 106 |
-
"rewards/rejected": -0.
|
| 107 |
"step": 70
|
| 108 |
},
|
| 109 |
{
|
| 110 |
-
"epoch": 0.
|
| 111 |
-
"learning_rate":
|
| 112 |
-
"logits/chosen":
|
| 113 |
-
"logits/rejected":
|
| 114 |
-
"logps/chosen": -
|
| 115 |
-
"logps/rejected": -
|
| 116 |
-
"loss": 0.
|
| 117 |
-
"rewards/accuracies": 0.
|
| 118 |
-
"rewards/chosen": 0.
|
| 119 |
-
"rewards/margins": 0.
|
| 120 |
-
"rewards/rejected": -0.
|
| 121 |
"step": 80
|
| 122 |
},
|
| 123 |
{
|
| 124 |
-
"epoch": 0.
|
| 125 |
-
"learning_rate": 9.
|
| 126 |
-
"logits/chosen":
|
| 127 |
-
"logits/rejected":
|
| 128 |
-
"logps/chosen": -
|
| 129 |
-
"logps/rejected": -
|
| 130 |
-
"loss": 0.
|
| 131 |
"rewards/accuracies": 0.606249988079071,
|
| 132 |
-
"rewards/chosen": 0.
|
| 133 |
-
"rewards/margins": 0.
|
| 134 |
-
"rewards/rejected": -0.
|
| 135 |
"step": 90
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"learning_rate": 9.
|
| 140 |
-
"logits/chosen":
|
| 141 |
-
"logits/rejected":
|
| 142 |
-
"logps/chosen": -
|
| 143 |
-
"logps/rejected": -
|
| 144 |
-
"loss": 0.
|
| 145 |
-
"rewards/accuracies": 0.
|
| 146 |
-
"rewards/chosen": -0.
|
| 147 |
-
"rewards/margins": 0.
|
| 148 |
-
"rewards/rejected": -0.
|
| 149 |
"step": 100
|
| 150 |
},
|
| 151 |
{
|
| 152 |
-
"epoch": 0.
|
| 153 |
-
"
|
| 154 |
-
"
|
| 155 |
-
"
|
| 156 |
-
"
|
| 157 |
-
"
|
| 158 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 159 |
"rewards/accuracies": 0.643750011920929,
|
| 160 |
-
"rewards/chosen": 0.
|
| 161 |
-
"rewards/margins": 0.
|
| 162 |
-
"rewards/rejected": -0.
|
| 163 |
"step": 110
|
| 164 |
},
|
| 165 |
{
|
| 166 |
-
"epoch": 0.
|
| 167 |
-
"learning_rate": 9.
|
| 168 |
-
"logits/chosen":
|
| 169 |
-
"logits/rejected":
|
| 170 |
-
"logps/chosen": -
|
| 171 |
-
"logps/rejected": -
|
| 172 |
-
"loss": 0.
|
| 173 |
"rewards/accuracies": 0.65625,
|
| 174 |
-
"rewards/chosen": 0.
|
| 175 |
-
"rewards/margins": 0.
|
| 176 |
-
"rewards/rejected": -0.
|
| 177 |
"step": 120
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
-
"learning_rate": 9.
|
| 182 |
-
"logits/chosen":
|
| 183 |
-
"logits/rejected":
|
| 184 |
-
"logps/chosen": -
|
| 185 |
-
"logps/rejected": -
|
| 186 |
-
"loss": 0.
|
| 187 |
-
"rewards/accuracies": 0.
|
| 188 |
-
"rewards/chosen": -0.
|
| 189 |
-
"rewards/margins": 0.
|
| 190 |
-
"rewards/rejected": -0.
|
| 191 |
"step": 130
|
| 192 |
},
|
| 193 |
{
|
| 194 |
-
"epoch": 0.
|
| 195 |
-
"learning_rate": 9.
|
| 196 |
-
"logits/chosen":
|
| 197 |
-
"logits/rejected":
|
| 198 |
-
"logps/chosen": -
|
| 199 |
-
"logps/rejected": -
|
| 200 |
-
"loss": 0.
|
| 201 |
-
"rewards/accuracies": 0.
|
| 202 |
-
"rewards/chosen": -0.
|
| 203 |
-
"rewards/margins": 0.
|
| 204 |
-
"rewards/rejected": -0.
|
| 205 |
"step": 140
|
| 206 |
},
|
| 207 |
{
|
| 208 |
-
"epoch": 0.
|
| 209 |
-
"learning_rate": 9.
|
| 210 |
-
"logits/chosen":
|
| 211 |
-
"logits/rejected":
|
| 212 |
-
"logps/chosen": -
|
| 213 |
-
"logps/rejected": -
|
| 214 |
-
"loss": 0.
|
| 215 |
-
"rewards/accuracies": 0.
|
| 216 |
-
"rewards/chosen": -0.
|
| 217 |
-
"rewards/margins": 0.
|
| 218 |
-
"rewards/rejected": -0.
|
| 219 |
"step": 150
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
-
"learning_rate": 9.
|
| 224 |
-
"logits/chosen":
|
| 225 |
-
"logits/rejected":
|
| 226 |
-
"logps/chosen": -
|
| 227 |
-
"logps/rejected": -
|
| 228 |
-
"loss": 0.
|
| 229 |
-
"rewards/accuracies": 0.
|
| 230 |
-
"rewards/chosen": -0.
|
| 231 |
-
"rewards/margins": 0.
|
| 232 |
-
"rewards/rejected": -
|
| 233 |
"step": 160
|
| 234 |
},
|
| 235 |
{
|
| 236 |
-
"epoch": 0.
|
| 237 |
-
"learning_rate": 9.
|
| 238 |
-
"logits/chosen":
|
| 239 |
-
"logits/rejected":
|
| 240 |
-
"logps/chosen": -
|
| 241 |
-
"logps/rejected": -
|
| 242 |
-
"loss": 0.
|
| 243 |
-
"rewards/accuracies": 0.
|
| 244 |
-
"rewards/chosen": -0.
|
| 245 |
-
"rewards/margins": 0.
|
| 246 |
-
"rewards/rejected": -
|
| 247 |
"step": 170
|
| 248 |
},
|
| 249 |
{
|
| 250 |
-
"epoch": 0.
|
| 251 |
-
"learning_rate": 9.
|
| 252 |
-
"logits/chosen":
|
| 253 |
-
"logits/rejected":
|
| 254 |
-
"logps/chosen": -
|
| 255 |
-
"logps/rejected": -
|
| 256 |
-
"loss": 0.
|
| 257 |
-
"rewards/accuracies": 0.
|
| 258 |
-
"rewards/chosen": 0.
|
| 259 |
-
"rewards/margins": 0.
|
| 260 |
-
"rewards/rejected": -
|
| 261 |
"step": 180
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
-
"learning_rate": 9.
|
| 266 |
-
"logits/chosen":
|
| 267 |
-
"logits/rejected":
|
| 268 |
-
"logps/chosen": -
|
| 269 |
-
"logps/rejected": -
|
| 270 |
-
"loss": 0.
|
| 271 |
"rewards/accuracies": 0.6499999761581421,
|
| 272 |
-
"rewards/chosen": -0.
|
| 273 |
-
"rewards/margins": 0.
|
| 274 |
-
"rewards/rejected": -
|
| 275 |
"step": 190
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 0.
|
| 279 |
-
"learning_rate": 9.
|
| 280 |
-
"logits/chosen":
|
| 281 |
-
"logits/rejected":
|
| 282 |
-
"logps/chosen": -
|
| 283 |
-
"logps/rejected": -
|
| 284 |
-
"loss": 0.
|
| 285 |
-
"rewards/accuracies": 0.
|
| 286 |
-
"rewards/chosen": -0.
|
| 287 |
-
"rewards/margins": 0.
|
| 288 |
-
"rewards/rejected": -
|
| 289 |
"step": 200
|
| 290 |
},
|
| 291 |
{
|
| 292 |
-
"epoch": 0.
|
| 293 |
-
"
|
| 294 |
-
"
|
| 295 |
-
"
|
| 296 |
-
"
|
| 297 |
-
"
|
| 298 |
-
"
|
| 299 |
-
"
|
| 300 |
-
"
|
| 301 |
-
"
|
| 302 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 303 |
"step": 210
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 0.
|
| 307 |
-
"learning_rate":
|
| 308 |
-
"logits/chosen":
|
| 309 |
-
"logits/rejected":
|
| 310 |
-
"logps/chosen": -
|
| 311 |
-
"logps/rejected": -
|
| 312 |
-
"loss": 0.
|
| 313 |
-
"rewards/accuracies": 0.
|
| 314 |
-
"rewards/chosen": -0.
|
| 315 |
-
"rewards/margins": 0.
|
| 316 |
-
"rewards/rejected": -
|
| 317 |
"step": 220
|
| 318 |
},
|
| 319 |
{
|
| 320 |
-
"epoch": 0.
|
| 321 |
-
"learning_rate":
|
| 322 |
-
"logits/chosen":
|
| 323 |
-
"logits/rejected":
|
| 324 |
-
"logps/chosen": -
|
| 325 |
-
"logps/rejected": -
|
| 326 |
-
"loss": 0.
|
| 327 |
-
"rewards/accuracies": 0.
|
| 328 |
-
"rewards/chosen": -
|
| 329 |
-
"rewards/margins": 0.
|
| 330 |
-
"rewards/rejected": -
|
| 331 |
"step": 230
|
| 332 |
},
|
| 333 |
{
|
| 334 |
-
"epoch": 0.
|
| 335 |
-
"learning_rate":
|
| 336 |
-
"logits/chosen":
|
| 337 |
-
"logits/rejected":
|
| 338 |
-
"logps/chosen": -
|
| 339 |
-
"logps/rejected": -
|
| 340 |
-
"loss": 0.
|
| 341 |
-
"rewards/accuracies": 0.
|
| 342 |
-
"rewards/chosen": -0.
|
| 343 |
-
"rewards/margins": 0.
|
| 344 |
-
"rewards/rejected": -
|
| 345 |
"step": 240
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 0.
|
| 349 |
-
"learning_rate":
|
| 350 |
-
"logits/chosen":
|
| 351 |
-
"logits/rejected":
|
| 352 |
-
"logps/chosen": -
|
| 353 |
-
"logps/rejected": -
|
| 354 |
-
"loss": 0.
|
| 355 |
-
"rewards/accuracies": 0.
|
| 356 |
-
"rewards/chosen": -0.
|
| 357 |
-
"rewards/margins": 0.
|
| 358 |
-
"rewards/rejected": -
|
| 359 |
"step": 250
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch": 0.
|
| 363 |
-
"learning_rate":
|
| 364 |
-
"logits/chosen":
|
| 365 |
-
"logits/rejected":
|
| 366 |
-
"logps/chosen": -
|
| 367 |
-
"logps/rejected": -
|
| 368 |
-
"loss": 0.
|
| 369 |
-
"rewards/accuracies": 0.
|
| 370 |
-
"rewards/chosen": -0.
|
| 371 |
-
"rewards/margins": 0.
|
| 372 |
-
"rewards/rejected": -
|
| 373 |
"step": 260
|
| 374 |
},
|
| 375 |
{
|
| 376 |
-
"epoch": 0.
|
| 377 |
-
"learning_rate":
|
| 378 |
-
"logits/chosen":
|
| 379 |
-
"logits/rejected":
|
| 380 |
-
"logps/chosen": -
|
| 381 |
-
"logps/rejected": -
|
| 382 |
-
"loss": 0.
|
| 383 |
-
"rewards/accuracies": 0.
|
| 384 |
-
"rewards/chosen": -0.
|
| 385 |
-
"rewards/margins": 0.
|
| 386 |
-
"rewards/rejected": -
|
| 387 |
"step": 270
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch": 0.
|
| 391 |
-
"learning_rate":
|
| 392 |
-
"logits/chosen":
|
| 393 |
-
"logits/rejected":
|
| 394 |
-
"logps/chosen": -
|
| 395 |
-
"logps/rejected": -
|
| 396 |
-
"loss": 0.
|
| 397 |
-
"rewards/accuracies": 0.
|
| 398 |
-
"rewards/chosen": -0.
|
| 399 |
-
"rewards/margins": 0.
|
| 400 |
-
"rewards/rejected": -
|
| 401 |
"step": 280
|
| 402 |
},
|
| 403 |
{
|
| 404 |
-
"epoch": 0.
|
| 405 |
-
"learning_rate":
|
| 406 |
-
"logits/chosen":
|
| 407 |
-
"logits/rejected":
|
| 408 |
-
"logps/chosen": -
|
| 409 |
-
"logps/rejected": -
|
| 410 |
-
"loss": 0.
|
| 411 |
-
"rewards/accuracies": 0.
|
| 412 |
-
"rewards/chosen": -0.
|
| 413 |
-
"rewards/margins": 0.
|
| 414 |
-
"rewards/rejected": -
|
| 415 |
"step": 290
|
| 416 |
},
|
| 417 |
{
|
| 418 |
-
"epoch": 0.
|
| 419 |
-
"learning_rate":
|
| 420 |
-
"logits/chosen":
|
| 421 |
-
"logits/rejected":
|
| 422 |
-
"logps/chosen": -
|
| 423 |
-
"logps/rejected": -
|
| 424 |
-
"loss": 0.
|
| 425 |
-
"rewards/accuracies": 0.
|
| 426 |
-
"rewards/chosen": -0.
|
| 427 |
-
"rewards/margins": 0.
|
| 428 |
-
"rewards/rejected": -
|
| 429 |
"step": 300
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch": 0.
|
| 433 |
-
"
|
| 434 |
-
"
|
| 435 |
-
"
|
| 436 |
-
"
|
| 437 |
-
"
|
| 438 |
-
"
|
| 439 |
-
"
|
| 440 |
-
"
|
| 441 |
-
"
|
| 442 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 443 |
"step": 310
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch": 0.
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"logits/chosen":
|
| 449 |
-
"logits/rejected":
|
| 450 |
-
"logps/chosen": -
|
| 451 |
-
"logps/rejected": -
|
| 452 |
-
"loss": 0.
|
| 453 |
"rewards/accuracies": 0.625,
|
| 454 |
-
"rewards/chosen": -0.
|
| 455 |
-
"rewards/margins": 0.
|
| 456 |
-
"rewards/rejected": -
|
| 457 |
"step": 320
|
| 458 |
},
|
| 459 |
{
|
| 460 |
-
"epoch": 0.
|
| 461 |
-
"learning_rate":
|
| 462 |
-
"logits/chosen":
|
| 463 |
-
"logits/rejected":
|
| 464 |
-
"logps/chosen": -
|
| 465 |
-
"logps/rejected": -
|
| 466 |
-
"loss": 0.
|
| 467 |
-
"rewards/accuracies": 0.
|
| 468 |
-
"rewards/chosen": -0.
|
| 469 |
-
"rewards/margins": 0.
|
| 470 |
-
"rewards/rejected": -
|
| 471 |
"step": 330
|
| 472 |
},
|
| 473 |
{
|
| 474 |
-
"epoch": 0.
|
| 475 |
-
"learning_rate":
|
| 476 |
-
"logits/chosen":
|
| 477 |
-
"logits/rejected":
|
| 478 |
-
"logps/chosen": -
|
| 479 |
-
"logps/rejected": -
|
| 480 |
-
"loss": 0.
|
| 481 |
-
"rewards/accuracies": 0.
|
| 482 |
-
"rewards/chosen": -
|
| 483 |
-
"rewards/margins": 0.
|
| 484 |
-
"rewards/rejected": -
|
| 485 |
"step": 340
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch": 0.
|
| 489 |
-
"learning_rate":
|
| 490 |
-
"logits/chosen":
|
| 491 |
-
"logits/rejected":
|
| 492 |
-
"logps/chosen": -
|
| 493 |
-
"logps/rejected": -
|
| 494 |
-
"loss": 0.
|
| 495 |
-
"rewards/accuracies": 0.
|
| 496 |
-
"rewards/chosen": -
|
| 497 |
-
"rewards/margins": 0.
|
| 498 |
-
"rewards/rejected": -
|
| 499 |
"step": 350
|
| 500 |
},
|
| 501 |
{
|
| 502 |
-
"epoch":
|
| 503 |
-
"learning_rate":
|
| 504 |
-
"logits/chosen":
|
| 505 |
-
"logits/rejected":
|
| 506 |
-
"logps/chosen": -
|
| 507 |
-
"logps/rejected": -
|
| 508 |
-
"loss": 0.
|
| 509 |
-
"rewards/accuracies": 0.
|
| 510 |
-
"rewards/chosen": -0.
|
| 511 |
-
"rewards/margins": 0.
|
| 512 |
-
"rewards/rejected": -
|
| 513 |
"step": 360
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch":
|
| 517 |
-
"learning_rate":
|
| 518 |
-
"logits/chosen":
|
| 519 |
-
"logits/rejected":
|
| 520 |
-
"logps/chosen": -
|
| 521 |
-
"logps/rejected": -
|
| 522 |
-
"loss": 0.
|
| 523 |
-
"rewards/accuracies": 0.
|
| 524 |
-
"rewards/chosen": -0.
|
| 525 |
-
"rewards/margins": 0.
|
| 526 |
-
"rewards/rejected": -
|
| 527 |
"step": 370
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch":
|
| 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": -
|
| 539 |
-
"rewards/margins": 0.
|
| 540 |
-
"rewards/rejected": -1.
|
| 541 |
"step": 380
|
| 542 |
},
|
| 543 |
{
|
| 544 |
-
"epoch":
|
| 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": -0.
|
| 553 |
-
"rewards/margins": 0.
|
| 554 |
-
"rewards/rejected": -1.
|
| 555 |
"step": 390
|
| 556 |
},
|
| 557 |
{
|
| 558 |
-
"epoch":
|
| 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": -0.
|
| 567 |
-
"rewards/margins": 0.
|
| 568 |
-
"rewards/rejected": -1.
|
| 569 |
"step": 400
|
| 570 |
},
|
| 571 |
{
|
| 572 |
-
"epoch":
|
| 573 |
-
"
|
| 574 |
-
"
|
| 575 |
-
"
|
| 576 |
-
"
|
| 577 |
-
"
|
| 578 |
-
"
|
| 579 |
-
"
|
| 580 |
-
"
|
| 581 |
-
"
|
| 582 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 583 |
"step": 410
|
| 584 |
},
|
| 585 |
{
|
| 586 |
-
"epoch":
|
| 587 |
-
"learning_rate":
|
| 588 |
-
"logits/chosen":
|
| 589 |
-
"logits/rejected":
|
| 590 |
-
"logps/chosen": -
|
| 591 |
-
"logps/rejected": -
|
| 592 |
-
"loss": 0.
|
| 593 |
-
"rewards/accuracies": 0.
|
| 594 |
-
"rewards/chosen": -0.
|
| 595 |
-
"rewards/margins": 0.
|
| 596 |
-
"rewards/rejected": -1.
|
| 597 |
"step": 420
|
| 598 |
},
|
| 599 |
{
|
| 600 |
-
"epoch":
|
| 601 |
-
"learning_rate":
|
| 602 |
-
"logits/chosen":
|
| 603 |
-
"logits/rejected":
|
| 604 |
-
"logps/chosen": -
|
| 605 |
-
"logps/rejected": -
|
| 606 |
-
"loss": 0.
|
| 607 |
-
"rewards/accuracies": 0.
|
| 608 |
-
"rewards/chosen": -0.
|
| 609 |
-
"rewards/margins": 0.
|
| 610 |
-
"rewards/rejected": -1.
|
| 611 |
"step": 430
|
| 612 |
},
|
| 613 |
{
|
| 614 |
-
"epoch":
|
| 615 |
-
"learning_rate":
|
| 616 |
-
"logits/chosen":
|
| 617 |
-
"logits/rejected":
|
| 618 |
-
"logps/chosen": -
|
| 619 |
-
"logps/rejected": -
|
| 620 |
-
"loss": 0.
|
| 621 |
-
"rewards/accuracies": 0.
|
| 622 |
-
"rewards/chosen": -0.
|
| 623 |
-
"rewards/margins": 0.
|
| 624 |
-
"rewards/rejected": -1.
|
| 625 |
"step": 440
|
| 626 |
},
|
| 627 |
{
|
| 628 |
-
"epoch":
|
| 629 |
-
"learning_rate":
|
| 630 |
-
"logits/chosen":
|
| 631 |
-
"logits/rejected":
|
| 632 |
-
"logps/chosen": -
|
| 633 |
-
"logps/rejected": -
|
| 634 |
-
"loss": 0.
|
| 635 |
-
"rewards/accuracies": 0.
|
| 636 |
-
"rewards/chosen": -0.
|
| 637 |
-
"rewards/margins": 0.
|
| 638 |
-
"rewards/rejected": -1.
|
| 639 |
"step": 450
|
| 640 |
},
|
| 641 |
{
|
| 642 |
-
"epoch":
|
| 643 |
-
"learning_rate":
|
| 644 |
-
"logits/chosen":
|
| 645 |
-
"logits/rejected":
|
| 646 |
-
"logps/chosen": -
|
| 647 |
-
"logps/rejected": -
|
| 648 |
-
"loss": 0.
|
| 649 |
-
"rewards/accuracies": 0.
|
| 650 |
-
"rewards/chosen": -1.
|
| 651 |
-
"rewards/margins": 0.
|
| 652 |
-
"rewards/rejected": -1.
|
| 653 |
"step": 460
|
| 654 |
},
|
| 655 |
{
|
| 656 |
-
"epoch":
|
| 657 |
-
"learning_rate":
|
| 658 |
-
"logits/chosen":
|
| 659 |
-
"logits/rejected":
|
| 660 |
-
"logps/chosen": -
|
| 661 |
-
"logps/rejected": -
|
| 662 |
-
"loss": 0.
|
| 663 |
-
"rewards/accuracies": 0.
|
| 664 |
-
"rewards/chosen": -1.
|
| 665 |
-
"rewards/margins": 0.
|
| 666 |
-
"rewards/rejected": -1.
|
| 667 |
"step": 470
|
| 668 |
},
|
| 669 |
{
|
| 670 |
-
"epoch": 1.
|
| 671 |
-
"learning_rate":
|
| 672 |
-
"logits/chosen": 0.
|
| 673 |
-
"logits/rejected": 0.
|
| 674 |
-
"logps/chosen": -
|
| 675 |
-
"logps/rejected": -
|
| 676 |
-
"loss": 0.
|
| 677 |
-
"rewards/accuracies": 0.
|
| 678 |
-
"rewards/chosen": -
|
| 679 |
-
"rewards/margins": 0.
|
| 680 |
-
"rewards/rejected": -1.
|
| 681 |
"step": 480
|
| 682 |
},
|
| 683 |
{
|
| 684 |
-
"epoch": 1.
|
| 685 |
-
"learning_rate":
|
| 686 |
-
"logits/chosen":
|
| 687 |
-
"logits/rejected": 0.
|
| 688 |
-
"logps/chosen": -
|
| 689 |
-
"logps/rejected": -
|
| 690 |
-
"loss": 0.
|
| 691 |
-
"rewards/accuracies": 0.
|
| 692 |
-
"rewards/chosen": -
|
| 693 |
-
"rewards/margins":
|
| 694 |
-
"rewards/rejected": -
|
| 695 |
"step": 490
|
| 696 |
},
|
| 697 |
{
|
| 698 |
-
"epoch": 1.
|
| 699 |
-
"learning_rate":
|
| 700 |
-
"logits/chosen": 0.
|
| 701 |
-
"logits/rejected": 0.
|
| 702 |
-
"logps/chosen": -
|
| 703 |
-
"logps/rejected": -
|
| 704 |
-
"loss": 0.
|
| 705 |
-
"rewards/accuracies": 0.
|
| 706 |
-
"rewards/chosen": -1.
|
| 707 |
-
"rewards/margins":
|
| 708 |
-
"rewards/rejected": -2.
|
| 709 |
"step": 500
|
| 710 |
},
|
| 711 |
{
|
| 712 |
-
"epoch": 1.
|
| 713 |
-
"
|
| 714 |
-
"
|
| 715 |
-
"
|
| 716 |
-
"
|
| 717 |
-
"
|
| 718 |
-
"
|
| 719 |
-
"
|
| 720 |
-
"
|
| 721 |
-
"
|
| 722 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 723 |
"step": 510
|
| 724 |
},
|
| 725 |
{
|
| 726 |
-
"epoch": 1.
|
| 727 |
-
"learning_rate":
|
| 728 |
-
"logits/chosen":
|
| 729 |
-
"logits/rejected": 0.
|
| 730 |
-
"logps/chosen": -
|
| 731 |
-
"logps/rejected": -
|
| 732 |
-
"loss": 0.
|
| 733 |
-
"rewards/accuracies": 0.
|
| 734 |
-
"rewards/chosen": -1.
|
| 735 |
-
"rewards/margins":
|
| 736 |
-
"rewards/rejected": -
|
| 737 |
"step": 520
|
| 738 |
},
|
| 739 |
{
|
| 740 |
-
"epoch": 1.
|
| 741 |
-
"learning_rate":
|
| 742 |
-
"logits/chosen":
|
| 743 |
-
"logits/rejected":
|
| 744 |
-
"logps/chosen": -
|
| 745 |
-
"logps/rejected": -
|
| 746 |
-
"loss": 0.
|
| 747 |
-
"rewards/accuracies": 0.
|
| 748 |
-
"rewards/chosen": -1.
|
| 749 |
-
"rewards/margins": 1.
|
| 750 |
-
"rewards/rejected": -
|
| 751 |
"step": 530
|
| 752 |
},
|
| 753 |
{
|
| 754 |
-
"epoch": 1.
|
| 755 |
-
"learning_rate":
|
| 756 |
-
"logits/chosen":
|
| 757 |
-
"logits/rejected": 0.
|
| 758 |
-
"logps/chosen": -
|
| 759 |
-
"logps/rejected": -
|
| 760 |
-
"loss": 0.
|
| 761 |
-
"rewards/accuracies": 0.
|
| 762 |
-
"rewards/chosen": -1.
|
| 763 |
-
"rewards/margins": 1.
|
| 764 |
-
"rewards/rejected": -2.
|
| 765 |
"step": 540
|
| 766 |
},
|
| 767 |
{
|
| 768 |
-
"epoch": 1.
|
| 769 |
-
"learning_rate":
|
| 770 |
-
"logits/chosen": 0.
|
| 771 |
-
"logits/rejected": 0.
|
| 772 |
-
"logps/chosen": -
|
| 773 |
-
"logps/rejected": -
|
| 774 |
-
"loss": 0.
|
| 775 |
-
"rewards/accuracies": 0.
|
| 776 |
-
"rewards/chosen": -1.
|
| 777 |
-
"rewards/margins":
|
| 778 |
-
"rewards/rejected": -2.
|
| 779 |
"step": 550
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch": 1.
|
| 783 |
-
"learning_rate":
|
| 784 |
-
"logits/chosen": 0.
|
| 785 |
-
"logits/rejected": 0.
|
| 786 |
-
"logps/chosen": -
|
| 787 |
-
"logps/rejected": -
|
| 788 |
-
"loss": 0.
|
| 789 |
"rewards/accuracies": 0.800000011920929,
|
| 790 |
-
"rewards/chosen": -1.
|
| 791 |
-
"rewards/margins":
|
| 792 |
-
"rewards/rejected": -2.
|
| 793 |
"step": 560
|
| 794 |
},
|
| 795 |
{
|
| 796 |
-
"epoch": 1.
|
| 797 |
-
"learning_rate":
|
| 798 |
-
"logits/chosen": 0.
|
| 799 |
-
"logits/rejected": 0.
|
| 800 |
-
"logps/chosen": -
|
| 801 |
-
"logps/rejected": -
|
| 802 |
-
"loss": 0.
|
| 803 |
-
"rewards/accuracies": 0.
|
| 804 |
-
"rewards/chosen": -1.
|
| 805 |
-
"rewards/margins": 1.
|
| 806 |
-
"rewards/rejected": -
|
| 807 |
"step": 570
|
| 808 |
},
|
| 809 |
{
|
| 810 |
-
"epoch": 1.
|
| 811 |
-
"learning_rate":
|
| 812 |
-
"logits/chosen": 0.
|
| 813 |
-
"logits/rejected": 0.
|
| 814 |
-
"logps/chosen": -
|
| 815 |
-
"logps/rejected": -
|
| 816 |
-
"loss": 0.
|
| 817 |
-
"rewards/accuracies": 0.
|
| 818 |
-
"rewards/chosen": -
|
| 819 |
-
"rewards/margins":
|
| 820 |
-
"rewards/rejected": -
|
| 821 |
"step": 580
|
| 822 |
},
|
| 823 |
{
|
| 824 |
-
"epoch": 1.
|
| 825 |
-
"learning_rate":
|
| 826 |
-
"logits/chosen": 0.
|
| 827 |
-
"logits/rejected": 0.
|
| 828 |
-
"logps/chosen": -
|
| 829 |
-
"logps/rejected": -
|
| 830 |
-
"loss": 0.
|
| 831 |
-
"rewards/accuracies": 0.
|
| 832 |
-
"rewards/chosen": -1.
|
| 833 |
-
"rewards/margins": 1.
|
| 834 |
-
"rewards/rejected": -
|
| 835 |
"step": 590
|
| 836 |
},
|
| 837 |
{
|
| 838 |
-
"epoch": 1.
|
| 839 |
-
"learning_rate":
|
| 840 |
-
"logits/chosen": 0.
|
| 841 |
-
"logits/rejected": 0.
|
| 842 |
-
"logps/chosen": -
|
| 843 |
-
"logps/rejected": -
|
| 844 |
-
"loss": 0.
|
| 845 |
-
"rewards/accuracies": 0.
|
| 846 |
-
"rewards/chosen": -1.
|
| 847 |
-
"rewards/margins": 1.
|
| 848 |
-
"rewards/rejected": -
|
| 849 |
"step": 600
|
| 850 |
},
|
| 851 |
{
|
| 852 |
-
"epoch": 1.
|
| 853 |
-
"
|
| 854 |
-
"
|
| 855 |
-
"
|
| 856 |
-
"
|
| 857 |
-
"
|
| 858 |
-
"
|
| 859 |
-
"
|
| 860 |
-
"
|
| 861 |
-
"
|
| 862 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 863 |
"step": 610
|
| 864 |
},
|
| 865 |
{
|
| 866 |
-
"epoch": 1.
|
| 867 |
-
"learning_rate":
|
| 868 |
-
"logits/chosen": 0.
|
| 869 |
-
"logits/rejected": 0.
|
| 870 |
-
"logps/chosen": -
|
| 871 |
-
"logps/rejected": -
|
| 872 |
-
"loss": 0.
|
| 873 |
-
"rewards/accuracies": 0.
|
| 874 |
-
"rewards/chosen": -1.
|
| 875 |
-
"rewards/margins": 1.
|
| 876 |
-
"rewards/rejected": -
|
| 877 |
"step": 620
|
| 878 |
},
|
| 879 |
{
|
| 880 |
-
"epoch": 1.
|
| 881 |
-
"learning_rate": 3.
|
| 882 |
-
"logits/chosen": 0.
|
| 883 |
-
"logits/rejected": 0.
|
| 884 |
-
"logps/chosen": -
|
| 885 |
-
"logps/rejected": -
|
| 886 |
-
"loss": 0.
|
| 887 |
-
"rewards/accuracies": 0.
|
| 888 |
-
"rewards/chosen": -1.
|
| 889 |
-
"rewards/margins":
|
| 890 |
-
"rewards/rejected": -
|
| 891 |
"step": 630
|
| 892 |
},
|
| 893 |
{
|
| 894 |
-
"epoch": 1.
|
| 895 |
-
"learning_rate": 2.
|
| 896 |
-
"logits/chosen": 0.
|
| 897 |
-
"logits/rejected": 0.
|
| 898 |
-
"logps/chosen": -
|
| 899 |
-
"logps/rejected": -
|
| 900 |
-
"loss": 0.
|
| 901 |
-
"rewards/accuracies": 0.
|
| 902 |
-
"rewards/chosen": -1.
|
| 903 |
-
"rewards/margins":
|
| 904 |
-
"rewards/rejected": -
|
| 905 |
"step": 640
|
| 906 |
},
|
| 907 |
{
|
| 908 |
-
"epoch": 1.
|
| 909 |
-
"learning_rate": 2.
|
| 910 |
-
"logits/chosen":
|
| 911 |
-
"logits/rejected":
|
| 912 |
-
"logps/chosen": -
|
| 913 |
-
"logps/rejected": -
|
| 914 |
-
"loss": 0.
|
| 915 |
-
"rewards/accuracies": 0.
|
| 916 |
-
"rewards/chosen": -1.
|
| 917 |
-
"rewards/margins":
|
| 918 |
-
"rewards/rejected": -
|
| 919 |
"step": 650
|
| 920 |
},
|
| 921 |
{
|
| 922 |
-
"epoch": 1.
|
| 923 |
-
"learning_rate":
|
| 924 |
-
"logits/chosen": 0.
|
| 925 |
-
"logits/rejected": 0.
|
| 926 |
-
"logps/chosen": -
|
| 927 |
-
"logps/rejected": -
|
| 928 |
-
"loss": 0.
|
| 929 |
-
"rewards/accuracies": 0.
|
| 930 |
-
"rewards/chosen": -1.
|
| 931 |
-
"rewards/margins": 1.
|
| 932 |
-
"rewards/rejected": -
|
| 933 |
"step": 660
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch": 1.
|
| 937 |
-
"learning_rate":
|
| 938 |
-
"logits/chosen": 0.
|
| 939 |
-
"logits/rejected": 0.
|
| 940 |
-
"logps/chosen": -
|
| 941 |
-
"logps/rejected": -
|
| 942 |
-
"loss": 0.
|
| 943 |
-
"rewards/accuracies": 0.
|
| 944 |
-
"rewards/chosen": -
|
| 945 |
-
"rewards/margins":
|
| 946 |
-
"rewards/rejected": -
|
| 947 |
"step": 670
|
| 948 |
},
|
| 949 |
{
|
| 950 |
-
"epoch": 1.
|
| 951 |
-
"learning_rate":
|
| 952 |
-
"logits/chosen": 0.
|
| 953 |
-
"logits/rejected": 0.
|
| 954 |
-
"logps/chosen": -
|
| 955 |
-
"logps/rejected": -
|
| 956 |
-
"loss": 0.
|
| 957 |
-
"rewards/accuracies": 0.
|
| 958 |
-
"rewards/chosen": -1.
|
| 959 |
-
"rewards/margins": 1.
|
| 960 |
-
"rewards/rejected": -
|
| 961 |
"step": 680
|
| 962 |
},
|
| 963 |
{
|
| 964 |
-
"epoch": 1.
|
| 965 |
-
"learning_rate": 2.
|
| 966 |
-
"logits/chosen":
|
| 967 |
-
"logits/rejected": 0.
|
| 968 |
-
"logps/chosen": -
|
| 969 |
-
"logps/rejected": -
|
| 970 |
-
"loss": 0.
|
| 971 |
-
"rewards/accuracies": 0.
|
| 972 |
-
"rewards/chosen": -1.
|
| 973 |
-
"rewards/margins": 1.
|
| 974 |
-
"rewards/rejected": -
|
| 975 |
"step": 690
|
| 976 |
},
|
| 977 |
{
|
| 978 |
-
"epoch": 1.
|
| 979 |
-
"learning_rate":
|
| 980 |
-
"logits/chosen": 0.
|
| 981 |
-
"logits/rejected": 0.
|
| 982 |
-
"logps/chosen": -
|
| 983 |
-
"logps/rejected": -
|
| 984 |
-
"loss": 0.
|
| 985 |
-
"rewards/accuracies": 0.
|
| 986 |
-
"rewards/chosen": -1.
|
| 987 |
-
"rewards/margins":
|
| 988 |
-
"rewards/rejected": -
|
| 989 |
"step": 700
|
| 990 |
},
|
| 991 |
{
|
| 992 |
-
"epoch":
|
| 993 |
-
"
|
| 994 |
-
"
|
| 995 |
-
"
|
| 996 |
-
"
|
| 997 |
-
"
|
| 998 |
-
"
|
| 999 |
-
"
|
| 1000 |
-
"
|
| 1001 |
-
"
|
| 1002 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1003 |
"step": 710
|
| 1004 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1005 |
{
|
| 1006 |
"epoch": 2.0,
|
| 1007 |
-
"step":
|
| 1008 |
"total_flos": 0.0,
|
| 1009 |
-
"train_loss": 0.
|
| 1010 |
-
"train_runtime":
|
| 1011 |
-
"train_samples_per_second": 8
|
| 1012 |
-
"train_steps_per_second": 0.
|
| 1013 |
}
|
| 1014 |
],
|
| 1015 |
"logging_steps": 10,
|
| 1016 |
-
"max_steps":
|
| 1017 |
"num_train_epochs": 2,
|
| 1018 |
-
"save_steps":
|
| 1019 |
"total_flos": 0.0,
|
| 1020 |
"trial_name": null,
|
| 1021 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9968602825745683,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 954,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.0416666666666667e-07,
|
| 14 |
+
"logits/chosen": 0.294111967086792,
|
| 15 |
+
"logits/rejected": 0.3186231553554535,
|
| 16 |
+
"logps/chosen": -295.23431396484375,
|
| 17 |
+
"logps/rejected": -290.9120178222656,
|
| 18 |
+
"loss": 0.1859,
|
| 19 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 20 |
+
"rewards/chosen": 6.947966903680936e-05,
|
| 21 |
+
"rewards/margins": 0.0006955973221920431,
|
| 22 |
+
"rewards/rejected": -0.0006261176313273609,
|
| 23 |
"step": 10
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.04,
|
| 27 |
+
"learning_rate": 2.0833333333333333e-07,
|
| 28 |
+
"logits/chosen": 0.2527770400047302,
|
| 29 |
+
"logits/rejected": 0.3252316415309906,
|
| 30 |
+
"logps/chosen": -318.26434326171875,
|
| 31 |
+
"logps/rejected": -289.712158203125,
|
| 32 |
+
"loss": 0.1867,
|
| 33 |
+
"rewards/accuracies": 0.53125,
|
| 34 |
+
"rewards/chosen": -0.0002512992941774428,
|
| 35 |
+
"rewards/margins": 0.0006638198392465711,
|
| 36 |
+
"rewards/rejected": -0.0009151192498393357,
|
| 37 |
"step": 20
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.06,
|
| 41 |
+
"learning_rate": 3.1249999999999997e-07,
|
| 42 |
+
"logits/chosen": 0.33200520277023315,
|
| 43 |
+
"logits/rejected": 0.386562705039978,
|
| 44 |
+
"logps/chosen": -296.428466796875,
|
| 45 |
+
"logps/rejected": -283.0581970214844,
|
| 46 |
+
"loss": 0.1751,
|
| 47 |
+
"rewards/accuracies": 0.606249988079071,
|
| 48 |
+
"rewards/chosen": -0.0013968495186418295,
|
| 49 |
+
"rewards/margins": 0.002019281964749098,
|
| 50 |
+
"rewards/rejected": -0.003416131716221571,
|
| 51 |
"step": 30
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.08,
|
| 55 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 56 |
+
"logits/chosen": 0.22856274247169495,
|
| 57 |
+
"logits/rejected": 0.32988241314888,
|
| 58 |
+
"logps/chosen": -305.99603271484375,
|
| 59 |
+
"logps/rejected": -293.2794189453125,
|
| 60 |
+
"loss": 0.1853,
|
| 61 |
+
"rewards/accuracies": 0.59375,
|
| 62 |
+
"rewards/chosen": -0.0037044710479676723,
|
| 63 |
+
"rewards/margins": 0.006597139872610569,
|
| 64 |
+
"rewards/rejected": -0.010301610454916954,
|
| 65 |
"step": 40
|
| 66 |
},
|
| 67 |
{
|
| 68 |
+
"epoch": 0.1,
|
| 69 |
+
"learning_rate": 5.208333333333334e-07,
|
| 70 |
+
"logits/chosen": 0.2711636424064636,
|
| 71 |
+
"logits/rejected": 0.3201292157173157,
|
| 72 |
+
"logps/chosen": -314.5457458496094,
|
| 73 |
+
"logps/rejected": -306.9486083984375,
|
| 74 |
+
"loss": 0.1804,
|
| 75 |
+
"rewards/accuracies": 0.643750011920929,
|
| 76 |
+
"rewards/chosen": -0.008643127046525478,
|
| 77 |
+
"rewards/margins": 0.0160285122692585,
|
| 78 |
+
"rewards/rejected": -0.024671640247106552,
|
| 79 |
"step": 50
|
| 80 |
},
|
| 81 |
{
|
| 82 |
+
"epoch": 0.13,
|
| 83 |
+
"learning_rate": 6.249999999999999e-07,
|
| 84 |
+
"logits/chosen": 0.318474680185318,
|
| 85 |
+
"logits/rejected": 0.4364282488822937,
|
| 86 |
+
"logps/chosen": -294.63751220703125,
|
| 87 |
+
"logps/rejected": -272.2162780761719,
|
| 88 |
+
"loss": 0.1829,
|
| 89 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 90 |
+
"rewards/chosen": -0.02497897669672966,
|
| 91 |
+
"rewards/margins": 0.03536849468946457,
|
| 92 |
+
"rewards/rejected": -0.06034746766090393,
|
| 93 |
"step": 60
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.15,
|
| 97 |
+
"learning_rate": 7.291666666666666e-07,
|
| 98 |
+
"logits/chosen": 0.3239240050315857,
|
| 99 |
+
"logits/rejected": 0.3422989249229431,
|
| 100 |
+
"logps/chosen": -304.72418212890625,
|
| 101 |
+
"logps/rejected": -310.6500549316406,
|
| 102 |
+
"loss": 0.1733,
|
| 103 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 104 |
+
"rewards/chosen": -0.05714759975671768,
|
| 105 |
+
"rewards/margins": 0.04348129406571388,
|
| 106 |
+
"rewards/rejected": -0.10062889009714127,
|
| 107 |
"step": 70
|
| 108 |
},
|
| 109 |
{
|
| 110 |
+
"epoch": 0.17,
|
| 111 |
+
"learning_rate": 8.333333333333333e-07,
|
| 112 |
+
"logits/chosen": 0.35590413212776184,
|
| 113 |
+
"logits/rejected": 0.4268670678138733,
|
| 114 |
+
"logps/chosen": -306.149169921875,
|
| 115 |
+
"logps/rejected": -279.54949951171875,
|
| 116 |
+
"loss": 0.1565,
|
| 117 |
+
"rewards/accuracies": 0.65625,
|
| 118 |
+
"rewards/chosen": -0.10662279278039932,
|
| 119 |
+
"rewards/margins": 0.10379795730113983,
|
| 120 |
+
"rewards/rejected": -0.21042077243328094,
|
| 121 |
"step": 80
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"epoch": 0.19,
|
| 125 |
+
"learning_rate": 9.374999999999999e-07,
|
| 126 |
+
"logits/chosen": 0.35440078377723694,
|
| 127 |
+
"logits/rejected": 0.46889105439186096,
|
| 128 |
+
"logps/chosen": -340.5415954589844,
|
| 129 |
+
"logps/rejected": -296.3750915527344,
|
| 130 |
+
"loss": 0.1275,
|
| 131 |
"rewards/accuracies": 0.606249988079071,
|
| 132 |
+
"rewards/chosen": -0.2491822987794876,
|
| 133 |
+
"rewards/margins": 0.09365006536245346,
|
| 134 |
+
"rewards/rejected": -0.34283238649368286,
|
| 135 |
"step": 90
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.21,
|
| 139 |
+
"learning_rate": 9.999463737538052e-07,
|
| 140 |
+
"logits/chosen": 0.4465979039669037,
|
| 141 |
+
"logits/rejected": 0.4796988070011139,
|
| 142 |
+
"logps/chosen": -326.7769470214844,
|
| 143 |
+
"logps/rejected": -353.57916259765625,
|
| 144 |
+
"loss": 0.1135,
|
| 145 |
+
"rewards/accuracies": 0.606249988079071,
|
| 146 |
+
"rewards/chosen": -0.39248064160346985,
|
| 147 |
+
"rewards/margins": 0.1683400571346283,
|
| 148 |
+
"rewards/rejected": -0.5608206987380981,
|
| 149 |
"step": 100
|
| 150 |
},
|
| 151 |
{
|
| 152 |
+
"epoch": 0.21,
|
| 153 |
+
"eval_logits/chosen": 0.5124487280845642,
|
| 154 |
+
"eval_logits/rejected": 0.5669243931770325,
|
| 155 |
+
"eval_logps/chosen": -344.704833984375,
|
| 156 |
+
"eval_logps/rejected": -387.95428466796875,
|
| 157 |
+
"eval_loss": 0.10603953152894974,
|
| 158 |
+
"eval_rewards/accuracies": 0.70703125,
|
| 159 |
+
"eval_rewards/chosen": -0.4112018942832947,
|
| 160 |
+
"eval_rewards/margins": 0.35208994150161743,
|
| 161 |
+
"eval_rewards/rejected": -0.7632918357849121,
|
| 162 |
+
"eval_runtime": 72.7637,
|
| 163 |
+
"eval_samples_per_second": 27.486,
|
| 164 |
+
"eval_steps_per_second": 0.44,
|
| 165 |
+
"step": 100
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.23,
|
| 169 |
+
"learning_rate": 9.993432105822034e-07,
|
| 170 |
+
"logits/chosen": 0.46603718400001526,
|
| 171 |
+
"logits/rejected": 0.5153345465660095,
|
| 172 |
+
"logps/chosen": -368.1249694824219,
|
| 173 |
+
"logps/rejected": -372.76104736328125,
|
| 174 |
+
"loss": 0.0948,
|
| 175 |
"rewards/accuracies": 0.643750011920929,
|
| 176 |
+
"rewards/chosen": -0.546645998954773,
|
| 177 |
+
"rewards/margins": 0.3340264856815338,
|
| 178 |
+
"rewards/rejected": -0.8806723356246948,
|
| 179 |
"step": 110
|
| 180 |
},
|
| 181 |
{
|
| 182 |
+
"epoch": 0.25,
|
| 183 |
+
"learning_rate": 9.980706626858607e-07,
|
| 184 |
+
"logits/chosen": 0.3831257224082947,
|
| 185 |
+
"logits/rejected": 0.4766886830329895,
|
| 186 |
+
"logps/chosen": -413.6620178222656,
|
| 187 |
+
"logps/rejected": -443.69097900390625,
|
| 188 |
+
"loss": 0.0872,
|
| 189 |
"rewards/accuracies": 0.65625,
|
| 190 |
+
"rewards/chosen": -0.6011692881584167,
|
| 191 |
+
"rewards/margins": 0.3687607944011688,
|
| 192 |
+
"rewards/rejected": -0.9699299931526184,
|
| 193 |
"step": 120
|
| 194 |
},
|
| 195 |
{
|
| 196 |
+
"epoch": 0.27,
|
| 197 |
+
"learning_rate": 9.961304359538434e-07,
|
| 198 |
+
"logits/chosen": 0.2602660059928894,
|
| 199 |
+
"logits/rejected": 0.38597235083580017,
|
| 200 |
+
"logps/chosen": -382.05316162109375,
|
| 201 |
+
"logps/rejected": -372.43804931640625,
|
| 202 |
+
"loss": 0.088,
|
| 203 |
+
"rewards/accuracies": 0.637499988079071,
|
| 204 |
+
"rewards/chosen": -0.5223852396011353,
|
| 205 |
+
"rewards/margins": 0.23098063468933105,
|
| 206 |
+
"rewards/rejected": -0.7533659338951111,
|
| 207 |
"step": 130
|
| 208 |
},
|
| 209 |
{
|
| 210 |
+
"epoch": 0.29,
|
| 211 |
+
"learning_rate": 9.935251313189563e-07,
|
| 212 |
+
"logits/chosen": 0.30170053243637085,
|
| 213 |
+
"logits/rejected": 0.34471625089645386,
|
| 214 |
+
"logps/chosen": -358.23101806640625,
|
| 215 |
+
"logps/rejected": -367.53643798828125,
|
| 216 |
+
"loss": 0.0912,
|
| 217 |
+
"rewards/accuracies": 0.706250011920929,
|
| 218 |
+
"rewards/chosen": -0.44500336050987244,
|
| 219 |
+
"rewards/margins": 0.29020196199417114,
|
| 220 |
+
"rewards/rejected": -0.7352052927017212,
|
| 221 |
"step": 140
|
| 222 |
},
|
| 223 |
{
|
| 224 |
+
"epoch": 0.31,
|
| 225 |
+
"learning_rate": 9.902582412711118e-07,
|
| 226 |
+
"logits/chosen": 0.2502862811088562,
|
| 227 |
+
"logits/rejected": 0.41286569833755493,
|
| 228 |
+
"logps/chosen": -331.702392578125,
|
| 229 |
+
"logps/rejected": -343.1724548339844,
|
| 230 |
+
"loss": 0.0805,
|
| 231 |
+
"rewards/accuracies": 0.643750011920929,
|
| 232 |
+
"rewards/chosen": -0.5062417984008789,
|
| 233 |
+
"rewards/margins": 0.3571576774120331,
|
| 234 |
+
"rewards/rejected": -0.8633993864059448,
|
| 235 |
"step": 150
|
| 236 |
},
|
| 237 |
{
|
| 238 |
+
"epoch": 0.33,
|
| 239 |
+
"learning_rate": 9.86334145175542e-07,
|
| 240 |
+
"logits/chosen": 0.2876195013523102,
|
| 241 |
+
"logits/rejected": 0.40448102355003357,
|
| 242 |
+
"logps/chosen": -360.23065185546875,
|
| 243 |
+
"logps/rejected": -405.818603515625,
|
| 244 |
+
"loss": 0.0739,
|
| 245 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 246 |
+
"rewards/chosen": -0.6673613786697388,
|
| 247 |
+
"rewards/margins": 0.4857359826564789,
|
| 248 |
+
"rewards/rejected": -1.15309739112854,
|
| 249 |
"step": 160
|
| 250 |
},
|
| 251 |
{
|
| 252 |
+
"epoch": 0.36,
|
| 253 |
+
"learning_rate": 9.817581034021272e-07,
|
| 254 |
+
"logits/chosen": 0.2106766402721405,
|
| 255 |
+
"logits/rejected": 0.3229225277900696,
|
| 256 |
+
"logps/chosen": -428.03643798828125,
|
| 257 |
+
"logps/rejected": -429.11322021484375,
|
| 258 |
+
"loss": 0.0617,
|
| 259 |
+
"rewards/accuracies": 0.668749988079071,
|
| 260 |
+
"rewards/chosen": -0.6919075846672058,
|
| 261 |
+
"rewards/margins": 0.339050829410553,
|
| 262 |
+
"rewards/rejected": -1.0309584140777588,
|
| 263 |
"step": 170
|
| 264 |
},
|
| 265 |
{
|
| 266 |
+
"epoch": 0.38,
|
| 267 |
+
"learning_rate": 9.765362502737097e-07,
|
| 268 |
+
"logits/chosen": 0.2592887878417969,
|
| 269 |
+
"logits/rejected": 0.32125118374824524,
|
| 270 |
+
"logps/chosen": -373.13519287109375,
|
| 271 |
+
"logps/rejected": -411.79132080078125,
|
| 272 |
+
"loss": 0.0567,
|
| 273 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 274 |
+
"rewards/chosen": -0.795512855052948,
|
| 275 |
+
"rewards/margins": 0.5069063901901245,
|
| 276 |
+
"rewards/rejected": -1.3024193048477173,
|
| 277 |
"step": 180
|
| 278 |
},
|
| 279 |
{
|
| 280 |
+
"epoch": 0.4,
|
| 281 |
+
"learning_rate": 9.706755858428485e-07,
|
| 282 |
+
"logits/chosen": 0.3405285179615021,
|
| 283 |
+
"logits/rejected": 0.4521329998970032,
|
| 284 |
+
"logps/chosen": -348.65020751953125,
|
| 285 |
+
"logps/rejected": -362.8314208984375,
|
| 286 |
+
"loss": 0.0642,
|
| 287 |
"rewards/accuracies": 0.6499999761581421,
|
| 288 |
+
"rewards/chosen": -0.8077024221420288,
|
| 289 |
+
"rewards/margins": 0.4007204473018646,
|
| 290 |
+
"rewards/rejected": -1.2084228992462158,
|
| 291 |
"step": 190
|
| 292 |
},
|
| 293 |
{
|
| 294 |
+
"epoch": 0.42,
|
| 295 |
+
"learning_rate": 9.641839665080363e-07,
|
| 296 |
+
"logits/chosen": 0.2780831456184387,
|
| 297 |
+
"logits/rejected": 0.38184481859207153,
|
| 298 |
+
"logps/chosen": -348.57135009765625,
|
| 299 |
+
"logps/rejected": -405.402099609375,
|
| 300 |
+
"loss": 0.0648,
|
| 301 |
+
"rewards/accuracies": 0.6875,
|
| 302 |
+
"rewards/chosen": -0.730074405670166,
|
| 303 |
+
"rewards/margins": 0.5448073148727417,
|
| 304 |
+
"rewards/rejected": -1.2748818397521973,
|
| 305 |
"step": 200
|
| 306 |
},
|
| 307 |
{
|
| 308 |
+
"epoch": 0.42,
|
| 309 |
+
"eval_logits/chosen": 0.31048065423965454,
|
| 310 |
+
"eval_logits/rejected": 0.3866545557975769,
|
| 311 |
+
"eval_logps/chosen": -372.07177734375,
|
| 312 |
+
"eval_logps/rejected": -442.6331787109375,
|
| 313 |
+
"eval_loss": 0.06636561453342438,
|
| 314 |
+
"eval_rewards/accuracies": 0.71875,
|
| 315 |
+
"eval_rewards/chosen": -0.6848713755607605,
|
| 316 |
+
"eval_rewards/margins": 0.6252094507217407,
|
| 317 |
+
"eval_rewards/rejected": -1.3100807666778564,
|
| 318 |
+
"eval_runtime": 74.8918,
|
| 319 |
+
"eval_samples_per_second": 26.705,
|
| 320 |
+
"eval_steps_per_second": 0.427,
|
| 321 |
+
"step": 200
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.44,
|
| 325 |
+
"learning_rate": 9.570700944819582e-07,
|
| 326 |
+
"logits/chosen": 0.23630718886852264,
|
| 327 |
+
"logits/rejected": 0.31974127888679504,
|
| 328 |
+
"logps/chosen": -394.2319641113281,
|
| 329 |
+
"logps/rejected": -433.3797912597656,
|
| 330 |
+
"loss": 0.0565,
|
| 331 |
+
"rewards/accuracies": 0.731249988079071,
|
| 332 |
+
"rewards/chosen": -0.7185484170913696,
|
| 333 |
+
"rewards/margins": 0.5530284643173218,
|
| 334 |
+
"rewards/rejected": -1.271577000617981,
|
| 335 |
"step": 210
|
| 336 |
},
|
| 337 |
{
|
| 338 |
+
"epoch": 0.46,
|
| 339 |
+
"learning_rate": 9.493435061259129e-07,
|
| 340 |
+
"logits/chosen": 0.13368266820907593,
|
| 341 |
+
"logits/rejected": 0.32444554567337036,
|
| 342 |
+
"logps/chosen": -409.3749084472656,
|
| 343 |
+
"logps/rejected": -413.44097900390625,
|
| 344 |
+
"loss": 0.0558,
|
| 345 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 346 |
+
"rewards/chosen": -0.8843868374824524,
|
| 347 |
+
"rewards/margins": 0.42033252120018005,
|
| 348 |
+
"rewards/rejected": -1.3047194480895996,
|
| 349 |
"step": 220
|
| 350 |
},
|
| 351 |
{
|
| 352 |
+
"epoch": 0.48,
|
| 353 |
+
"learning_rate": 9.4101455916603e-07,
|
| 354 |
+
"logits/chosen": 0.23291189968585968,
|
| 355 |
+
"logits/rejected": 0.3359973430633545,
|
| 356 |
+
"logps/chosen": -385.23681640625,
|
| 357 |
+
"logps/rejected": -439.33294677734375,
|
| 358 |
+
"loss": 0.0557,
|
| 359 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 360 |
+
"rewards/chosen": -1.0321977138519287,
|
| 361 |
+
"rewards/margins": 0.5443316102027893,
|
| 362 |
+
"rewards/rejected": -1.5765292644500732,
|
| 363 |
"step": 230
|
| 364 |
},
|
| 365 |
{
|
| 366 |
+
"epoch": 0.5,
|
| 367 |
+
"learning_rate": 9.320944188084241e-07,
|
| 368 |
+
"logits/chosen": 0.1577446311712265,
|
| 369 |
+
"logits/rejected": 0.2382678985595703,
|
| 370 |
+
"logps/chosen": -437.01275634765625,
|
| 371 |
+
"logps/rejected": -492.4737243652344,
|
| 372 |
+
"loss": 0.0618,
|
| 373 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 374 |
+
"rewards/chosen": -0.8377536535263062,
|
| 375 |
+
"rewards/margins": 0.61639404296875,
|
| 376 |
+
"rewards/rejected": -1.4541476964950562,
|
| 377 |
"step": 240
|
| 378 |
},
|
| 379 |
{
|
| 380 |
+
"epoch": 0.52,
|
| 381 |
+
"learning_rate": 9.225950427718974e-07,
|
| 382 |
+
"logits/chosen": 0.19211766123771667,
|
| 383 |
+
"logits/rejected": 0.27316194772720337,
|
| 384 |
+
"logps/chosen": -396.7606201171875,
|
| 385 |
+
"logps/rejected": -429.4735412597656,
|
| 386 |
+
"loss": 0.0599,
|
| 387 |
+
"rewards/accuracies": 0.65625,
|
| 388 |
+
"rewards/chosen": -0.8097313046455383,
|
| 389 |
+
"rewards/margins": 0.5714440941810608,
|
| 390 |
+
"rewards/rejected": -1.3811752796173096,
|
| 391 |
"step": 250
|
| 392 |
},
|
| 393 |
{
|
| 394 |
+
"epoch": 0.54,
|
| 395 |
+
"learning_rate": 9.125291652582547e-07,
|
| 396 |
+
"logits/chosen": 0.2013740837574005,
|
| 397 |
+
"logits/rejected": 0.3595736622810364,
|
| 398 |
+
"logps/chosen": -440.74432373046875,
|
| 399 |
+
"logps/rejected": -463.736572265625,
|
| 400 |
+
"loss": 0.0633,
|
| 401 |
+
"rewards/accuracies": 0.675000011920929,
|
| 402 |
+
"rewards/chosen": -0.9895216822624207,
|
| 403 |
+
"rewards/margins": 0.7094276547431946,
|
| 404 |
+
"rewards/rejected": -1.6989490985870361,
|
| 405 |
"step": 260
|
| 406 |
},
|
| 407 |
{
|
| 408 |
+
"epoch": 0.57,
|
| 409 |
+
"learning_rate": 9.019102798817195e-07,
|
| 410 |
+
"logits/chosen": 0.2775947153568268,
|
| 411 |
+
"logits/rejected": 0.3985646963119507,
|
| 412 |
+
"logps/chosen": -395.5704345703125,
|
| 413 |
+
"logps/rejected": -408.7200622558594,
|
| 414 |
+
"loss": 0.0775,
|
| 415 |
+
"rewards/accuracies": 0.71875,
|
| 416 |
+
"rewards/chosen": -0.6699169874191284,
|
| 417 |
+
"rewards/margins": 0.4529454708099365,
|
| 418 |
+
"rewards/rejected": -1.122862458229065,
|
| 419 |
"step": 270
|
| 420 |
},
|
| 421 |
{
|
| 422 |
+
"epoch": 0.59,
|
| 423 |
+
"learning_rate": 8.90752621580335e-07,
|
| 424 |
+
"logits/chosen": 0.25832900404930115,
|
| 425 |
+
"logits/rejected": 0.3815799355506897,
|
| 426 |
+
"logps/chosen": -392.74462890625,
|
| 427 |
+
"logps/rejected": -462.65399169921875,
|
| 428 |
+
"loss": 0.0621,
|
| 429 |
+
"rewards/accuracies": 0.706250011920929,
|
| 430 |
+
"rewards/chosen": -0.8548792004585266,
|
| 431 |
+
"rewards/margins": 0.5883964896202087,
|
| 432 |
+
"rewards/rejected": -1.4432756900787354,
|
| 433 |
"step": 280
|
| 434 |
},
|
| 435 |
{
|
| 436 |
+
"epoch": 0.61,
|
| 437 |
+
"learning_rate": 8.79071147533597e-07,
|
| 438 |
+
"logits/chosen": 0.27000856399536133,
|
| 439 |
+
"logits/rejected": 0.3684462904930115,
|
| 440 |
+
"logps/chosen": -373.5545654296875,
|
| 441 |
+
"logps/rejected": -402.52203369140625,
|
| 442 |
+
"loss": 0.0565,
|
| 443 |
+
"rewards/accuracies": 0.65625,
|
| 444 |
+
"rewards/chosen": -0.7661363482475281,
|
| 445 |
+
"rewards/margins": 0.4899563789367676,
|
| 446 |
+
"rewards/rejected": -1.2560927867889404,
|
| 447 |
"step": 290
|
| 448 |
},
|
| 449 |
{
|
| 450 |
+
"epoch": 0.63,
|
| 451 |
+
"learning_rate": 8.668815171119019e-07,
|
| 452 |
+
"logits/chosen": 0.20499548316001892,
|
| 453 |
+
"logits/rejected": 0.4121769070625305,
|
| 454 |
+
"logps/chosen": -405.9000549316406,
|
| 455 |
+
"logps/rejected": -392.7608337402344,
|
| 456 |
+
"loss": 0.0565,
|
| 457 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 458 |
+
"rewards/chosen": -0.9424744844436646,
|
| 459 |
+
"rewards/margins": 0.392583429813385,
|
| 460 |
+
"rewards/rejected": -1.3350577354431152,
|
| 461 |
"step": 300
|
| 462 |
},
|
| 463 |
{
|
| 464 |
+
"epoch": 0.63,
|
| 465 |
+
"eval_logits/chosen": 0.28342685103416443,
|
| 466 |
+
"eval_logits/rejected": 0.3624545931816101,
|
| 467 |
+
"eval_logps/chosen": -379.5645751953125,
|
| 468 |
+
"eval_logps/rejected": -461.2485656738281,
|
| 469 |
+
"eval_loss": 0.05852030962705612,
|
| 470 |
+
"eval_rewards/accuracies": 0.73046875,
|
| 471 |
+
"eval_rewards/chosen": -0.7597990036010742,
|
| 472 |
+
"eval_rewards/margins": 0.7364358305931091,
|
| 473 |
+
"eval_rewards/rejected": -1.4962347745895386,
|
| 474 |
+
"eval_runtime": 74.348,
|
| 475 |
+
"eval_samples_per_second": 26.901,
|
| 476 |
+
"eval_steps_per_second": 0.43,
|
| 477 |
+
"step": 300
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.65,
|
| 481 |
+
"learning_rate": 8.54200070884685e-07,
|
| 482 |
+
"logits/chosen": 0.2647777199745178,
|
| 483 |
+
"logits/rejected": 0.28958767652511597,
|
| 484 |
+
"logps/chosen": -424.25604248046875,
|
| 485 |
+
"logps/rejected": -472.11639404296875,
|
| 486 |
+
"loss": 0.0603,
|
| 487 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 488 |
+
"rewards/chosen": -1.0271676778793335,
|
| 489 |
+
"rewards/margins": 0.4712137281894684,
|
| 490 |
+
"rewards/rejected": -1.4983813762664795,
|
| 491 |
"step": 310
|
| 492 |
},
|
| 493 |
{
|
| 494 |
+
"epoch": 0.67,
|
| 495 |
+
"learning_rate": 8.410438087153911e-07,
|
| 496 |
+
"logits/chosen": 0.18836882710456848,
|
| 497 |
+
"logits/rejected": 0.3168850541114807,
|
| 498 |
+
"logps/chosen": -439.59716796875,
|
| 499 |
+
"logps/rejected": -433.8974609375,
|
| 500 |
+
"loss": 0.0608,
|
| 501 |
"rewards/accuracies": 0.625,
|
| 502 |
+
"rewards/chosen": -0.8473286628723145,
|
| 503 |
+
"rewards/margins": 0.38924068212509155,
|
| 504 |
+
"rewards/rejected": -1.2365691661834717,
|
| 505 |
"step": 320
|
| 506 |
},
|
| 507 |
{
|
| 508 |
+
"epoch": 0.69,
|
| 509 |
+
"learning_rate": 8.274303669726426e-07,
|
| 510 |
+
"logits/chosen": 0.23949918150901794,
|
| 511 |
+
"logits/rejected": 0.25133609771728516,
|
| 512 |
+
"logps/chosen": -380.25909423828125,
|
| 513 |
+
"logps/rejected": -497.5677185058594,
|
| 514 |
+
"loss": 0.0576,
|
| 515 |
+
"rewards/accuracies": 0.737500011920929,
|
| 516 |
+
"rewards/chosen": -0.7643330097198486,
|
| 517 |
+
"rewards/margins": 0.8182552456855774,
|
| 518 |
+
"rewards/rejected": -1.5825883150100708,
|
| 519 |
"step": 330
|
| 520 |
},
|
| 521 |
{
|
| 522 |
+
"epoch": 0.71,
|
| 523 |
+
"learning_rate": 8.133779948881513e-07,
|
| 524 |
+
"logits/chosen": 0.2764629125595093,
|
| 525 |
+
"logits/rejected": 0.2793533205986023,
|
| 526 |
+
"logps/chosen": -400.74652099609375,
|
| 527 |
+
"logps/rejected": -499.9100036621094,
|
| 528 |
+
"loss": 0.0488,
|
| 529 |
+
"rewards/accuracies": 0.643750011920929,
|
| 530 |
+
"rewards/chosen": -1.11314058303833,
|
| 531 |
+
"rewards/margins": 0.7436734437942505,
|
| 532 |
+
"rewards/rejected": -1.8568140268325806,
|
| 533 |
"step": 340
|
| 534 |
},
|
| 535 |
{
|
| 536 |
+
"epoch": 0.73,
|
| 537 |
+
"learning_rate": 7.989055300930704e-07,
|
| 538 |
+
"logits/chosen": 0.321544349193573,
|
| 539 |
+
"logits/rejected": 0.34274429082870483,
|
| 540 |
+
"logps/chosen": -420.0064392089844,
|
| 541 |
+
"logps/rejected": -518.2902221679688,
|
| 542 |
+
"loss": 0.047,
|
| 543 |
+
"rewards/accuracies": 0.731249988079071,
|
| 544 |
+
"rewards/chosen": -1.1256664991378784,
|
| 545 |
+
"rewards/margins": 0.7915542125701904,
|
| 546 |
+
"rewards/rejected": -1.9172207117080688,
|
| 547 |
"step": 350
|
| 548 |
},
|
| 549 |
{
|
| 550 |
+
"epoch": 0.75,
|
| 551 |
+
"learning_rate": 7.840323733655778e-07,
|
| 552 |
+
"logits/chosen": 0.25519177317619324,
|
| 553 |
+
"logits/rejected": 0.36787793040275574,
|
| 554 |
+
"logps/chosen": -438.910400390625,
|
| 555 |
+
"logps/rejected": -482.37774658203125,
|
| 556 |
+
"loss": 0.0508,
|
| 557 |
+
"rewards/accuracies": 0.737500011920929,
|
| 558 |
+
"rewards/chosen": -0.9125250577926636,
|
| 559 |
+
"rewards/margins": 0.8073097467422485,
|
| 560 |
+
"rewards/rejected": -1.7198346853256226,
|
| 561 |
"step": 360
|
| 562 |
},
|
| 563 |
{
|
| 564 |
+
"epoch": 0.77,
|
| 565 |
+
"learning_rate": 7.687784626235447e-07,
|
| 566 |
+
"logits/chosen": 0.24404089152812958,
|
| 567 |
+
"logits/rejected": 0.3457106649875641,
|
| 568 |
+
"logps/chosen": -447.19036865234375,
|
| 569 |
+
"logps/rejected": -464.69091796875,
|
| 570 |
+
"loss": 0.0524,
|
| 571 |
+
"rewards/accuracies": 0.706250011920929,
|
| 572 |
+
"rewards/chosen": -0.9422460794448853,
|
| 573 |
+
"rewards/margins": 0.6395829916000366,
|
| 574 |
+
"rewards/rejected": -1.5818291902542114,
|
| 575 |
"step": 370
|
| 576 |
},
|
| 577 |
{
|
| 578 |
+
"epoch": 0.8,
|
| 579 |
+
"learning_rate": 7.531642461971514e-07,
|
| 580 |
+
"logits/chosen": 0.13880091905593872,
|
| 581 |
+
"logits/rejected": 0.28412070870399475,
|
| 582 |
+
"logps/chosen": -437.6341247558594,
|
| 583 |
+
"logps/rejected": -500.48712158203125,
|
| 584 |
+
"loss": 0.0526,
|
| 585 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 586 |
+
"rewards/chosen": -1.0558099746704102,
|
| 587 |
+
"rewards/margins": 0.8256821632385254,
|
| 588 |
+
"rewards/rejected": -1.881492018699646,
|
| 589 |
"step": 380
|
| 590 |
},
|
| 591 |
{
|
| 592 |
+
"epoch": 0.82,
|
| 593 |
+
"learning_rate": 7.372106554172801e-07,
|
| 594 |
+
"logits/chosen": 0.19525602459907532,
|
| 595 |
+
"logits/rejected": 0.29236602783203125,
|
| 596 |
+
"logps/chosen": -405.10064697265625,
|
| 597 |
+
"logps/rejected": -449.4366149902344,
|
| 598 |
+
"loss": 0.06,
|
| 599 |
+
"rewards/accuracies": 0.706250011920929,
|
| 600 |
+
"rewards/chosen": -0.8169919848442078,
|
| 601 |
+
"rewards/margins": 0.6195858120918274,
|
| 602 |
+
"rewards/rejected": -1.4365779161453247,
|
| 603 |
"step": 390
|
| 604 |
},
|
| 605 |
{
|
| 606 |
+
"epoch": 0.84,
|
| 607 |
+
"learning_rate": 7.209390765564318e-07,
|
| 608 |
+
"logits/chosen": 0.157417893409729,
|
| 609 |
+
"logits/rejected": 0.24023254215717316,
|
| 610 |
+
"logps/chosen": -378.41204833984375,
|
| 611 |
+
"logps/rejected": -445.98272705078125,
|
| 612 |
+
"loss": 0.0643,
|
| 613 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 614 |
+
"rewards/chosen": -0.813235878944397,
|
| 615 |
+
"rewards/margins": 0.5227171182632446,
|
| 616 |
+
"rewards/rejected": -1.3359529972076416,
|
| 617 |
"step": 400
|
| 618 |
},
|
| 619 |
{
|
| 620 |
+
"epoch": 0.84,
|
| 621 |
+
"eval_logits/chosen": 0.18623091280460358,
|
| 622 |
+
"eval_logits/rejected": 0.2607361972332001,
|
| 623 |
+
"eval_logps/chosen": -370.18780517578125,
|
| 624 |
+
"eval_logps/rejected": -451.6427917480469,
|
| 625 |
+
"eval_loss": 0.06227795034646988,
|
| 626 |
+
"eval_rewards/accuracies": 0.73046875,
|
| 627 |
+
"eval_rewards/chosen": -0.6660312414169312,
|
| 628 |
+
"eval_rewards/margins": 0.7341459393501282,
|
| 629 |
+
"eval_rewards/rejected": -1.400177240371704,
|
| 630 |
+
"eval_runtime": 75.0562,
|
| 631 |
+
"eval_samples_per_second": 26.647,
|
| 632 |
+
"eval_steps_per_second": 0.426,
|
| 633 |
+
"step": 400
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.86,
|
| 637 |
+
"learning_rate": 7.043713221597773e-07,
|
| 638 |
+
"logits/chosen": 0.13670489192008972,
|
| 639 |
+
"logits/rejected": 0.21374602615833282,
|
| 640 |
+
"logps/chosen": -379.88201904296875,
|
| 641 |
+
"logps/rejected": -425.9203186035156,
|
| 642 |
+
"loss": 0.0629,
|
| 643 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 644 |
+
"rewards/chosen": -0.6690796613693237,
|
| 645 |
+
"rewards/margins": 0.6432248950004578,
|
| 646 |
+
"rewards/rejected": -1.3123046159744263,
|
| 647 |
"step": 410
|
| 648 |
},
|
| 649 |
{
|
| 650 |
+
"epoch": 0.88,
|
| 651 |
+
"learning_rate": 6.875296018047809e-07,
|
| 652 |
+
"logits/chosen": 0.15494689345359802,
|
| 653 |
+
"logits/rejected": 0.22296158969402313,
|
| 654 |
+
"logps/chosen": -377.009765625,
|
| 655 |
+
"logps/rejected": -412.5980529785156,
|
| 656 |
+
"loss": 0.0605,
|
| 657 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 658 |
+
"rewards/chosen": -0.6854912042617798,
|
| 659 |
+
"rewards/margins": 0.6449334025382996,
|
| 660 |
+
"rewards/rejected": -1.3304245471954346,
|
| 661 |
"step": 420
|
| 662 |
},
|
| 663 |
{
|
| 664 |
+
"epoch": 0.9,
|
| 665 |
+
"learning_rate": 6.704364923285857e-07,
|
| 666 |
+
"logits/chosen": 0.10000725835561752,
|
| 667 |
+
"logits/rejected": 0.20355455577373505,
|
| 668 |
+
"logps/chosen": -442.444580078125,
|
| 669 |
+
"logps/rejected": -475.17901611328125,
|
| 670 |
+
"loss": 0.0633,
|
| 671 |
+
"rewards/accuracies": 0.699999988079071,
|
| 672 |
+
"rewards/chosen": -0.920007050037384,
|
| 673 |
+
"rewards/margins": 0.6362106204032898,
|
| 674 |
+
"rewards/rejected": -1.5562175512313843,
|
| 675 |
"step": 430
|
| 676 |
},
|
| 677 |
{
|
| 678 |
+
"epoch": 0.92,
|
| 679 |
+
"learning_rate": 6.531149075630796e-07,
|
| 680 |
+
"logits/chosen": 0.18681207299232483,
|
| 681 |
+
"logits/rejected": 0.24837341904640198,
|
| 682 |
+
"logps/chosen": -394.078857421875,
|
| 683 |
+
"logps/rejected": -446.2527770996094,
|
| 684 |
+
"loss": 0.0585,
|
| 685 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 686 |
+
"rewards/chosen": -0.9974417686462402,
|
| 687 |
+
"rewards/margins": 0.7502747774124146,
|
| 688 |
+
"rewards/rejected": -1.7477165460586548,
|
| 689 |
"step": 440
|
| 690 |
},
|
| 691 |
{
|
| 692 |
+
"epoch": 0.94,
|
| 693 |
+
"learning_rate": 6.355880676182085e-07,
|
| 694 |
+
"logits/chosen": 0.1681515872478485,
|
| 695 |
+
"logits/rejected": 0.34068578481674194,
|
| 696 |
+
"logps/chosen": -438.5791015625,
|
| 697 |
+
"logps/rejected": -468.8702087402344,
|
| 698 |
+
"loss": 0.056,
|
| 699 |
+
"rewards/accuracies": 0.65625,
|
| 700 |
+
"rewards/chosen": -0.959572434425354,
|
| 701 |
+
"rewards/margins": 0.6984410881996155,
|
| 702 |
+
"rewards/rejected": -1.6580133438110352,
|
| 703 |
"step": 450
|
| 704 |
},
|
| 705 |
{
|
| 706 |
+
"epoch": 0.96,
|
| 707 |
+
"learning_rate": 6.178794677547137e-07,
|
| 708 |
+
"logits/chosen": 0.22829516232013702,
|
| 709 |
+
"logits/rejected": 0.33127135038375854,
|
| 710 |
+
"logps/chosen": -417.236572265625,
|
| 711 |
+
"logps/rejected": -454.47930908203125,
|
| 712 |
+
"loss": 0.0488,
|
| 713 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 714 |
+
"rewards/chosen": -1.1504933834075928,
|
| 715 |
+
"rewards/margins": 0.6518447399139404,
|
| 716 |
+
"rewards/rejected": -1.8023380041122437,
|
| 717 |
"step": 460
|
| 718 |
},
|
| 719 |
{
|
| 720 |
+
"epoch": 0.98,
|
| 721 |
+
"learning_rate": 6.000128468880222e-07,
|
| 722 |
+
"logits/chosen": 0.1735696941614151,
|
| 723 |
+
"logits/rejected": 0.23917797207832336,
|
| 724 |
+
"logps/chosen": -441.33526611328125,
|
| 725 |
+
"logps/rejected": -507.3721618652344,
|
| 726 |
+
"loss": 0.0469,
|
| 727 |
+
"rewards/accuracies": 0.699999988079071,
|
| 728 |
+
"rewards/chosen": -1.0348496437072754,
|
| 729 |
+
"rewards/margins": 0.6962344646453857,
|
| 730 |
+
"rewards/rejected": -1.7310841083526611,
|
| 731 |
"step": 470
|
| 732 |
},
|
| 733 |
{
|
| 734 |
+
"epoch": 1.0,
|
| 735 |
+
"learning_rate": 5.820121557655108e-07,
|
| 736 |
+
"logits/chosen": 0.16151778399944305,
|
| 737 |
+
"logits/rejected": 0.30070850253105164,
|
| 738 |
+
"logps/chosen": -389.2116394042969,
|
| 739 |
+
"logps/rejected": -456.37554931640625,
|
| 740 |
+
"loss": 0.0468,
|
| 741 |
+
"rewards/accuracies": 0.731249988079071,
|
| 742 |
+
"rewards/chosen": -0.9426639676094055,
|
| 743 |
+
"rewards/margins": 0.6416986584663391,
|
| 744 |
+
"rewards/rejected": -1.5843626260757446,
|
| 745 |
"step": 480
|
| 746 |
},
|
| 747 |
{
|
| 748 |
+
"epoch": 1.03,
|
| 749 |
+
"learning_rate": 5.639015248598023e-07,
|
| 750 |
+
"logits/chosen": 0.2655644118785858,
|
| 751 |
+
"logits/rejected": 0.3393844962120056,
|
| 752 |
+
"logps/chosen": -394.0737609863281,
|
| 753 |
+
"logps/rejected": -481.3250427246094,
|
| 754 |
+
"loss": 0.0334,
|
| 755 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 756 |
+
"rewards/chosen": -0.886431097984314,
|
| 757 |
+
"rewards/margins": 1.0793970823287964,
|
| 758 |
+
"rewards/rejected": -1.9658282995224,
|
| 759 |
"step": 490
|
| 760 |
},
|
| 761 |
{
|
| 762 |
+
"epoch": 1.05,
|
| 763 |
+
"learning_rate": 5.457052320211339e-07,
|
| 764 |
+
"logits/chosen": 0.30765271186828613,
|
| 765 |
+
"logits/rejected": 0.33542585372924805,
|
| 766 |
+
"logps/chosen": -416.736083984375,
|
| 767 |
+
"logps/rejected": -526.0009765625,
|
| 768 |
+
"loss": 0.0214,
|
| 769 |
+
"rewards/accuracies": 0.75,
|
| 770 |
+
"rewards/chosen": -1.1540133953094482,
|
| 771 |
+
"rewards/margins": 1.152156114578247,
|
| 772 |
+
"rewards/rejected": -2.3061697483062744,
|
| 773 |
"step": 500
|
| 774 |
},
|
| 775 |
{
|
| 776 |
+
"epoch": 1.05,
|
| 777 |
+
"eval_logits/chosen": 0.4214042127132416,
|
| 778 |
+
"eval_logits/rejected": 0.4999687075614929,
|
| 779 |
+
"eval_logps/chosen": -460.466796875,
|
| 780 |
+
"eval_logps/rejected": -587.046875,
|
| 781 |
+
"eval_loss": 0.030972089618444443,
|
| 782 |
+
"eval_rewards/accuracies": 0.7421875,
|
| 783 |
+
"eval_rewards/chosen": -1.5688213109970093,
|
| 784 |
+
"eval_rewards/margins": 1.185396432876587,
|
| 785 |
+
"eval_rewards/rejected": -2.7542178630828857,
|
| 786 |
+
"eval_runtime": 74.4978,
|
| 787 |
+
"eval_samples_per_second": 26.846,
|
| 788 |
+
"eval_steps_per_second": 0.43,
|
| 789 |
+
"step": 500
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 1.07,
|
| 793 |
+
"learning_rate": 5.274476699321637e-07,
|
| 794 |
+
"logits/chosen": 0.35821476578712463,
|
| 795 |
+
"logits/rejected": 0.46395426988601685,
|
| 796 |
+
"logps/chosen": -462.271484375,
|
| 797 |
+
"logps/rejected": -581.0950927734375,
|
| 798 |
+
"loss": 0.0148,
|
| 799 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 800 |
+
"rewards/chosen": -1.6695778369903564,
|
| 801 |
+
"rewards/margins": 1.4044713973999023,
|
| 802 |
+
"rewards/rejected": -3.074049472808838,
|
| 803 |
"step": 510
|
| 804 |
},
|
| 805 |
{
|
| 806 |
+
"epoch": 1.09,
|
| 807 |
+
"learning_rate": 5.091533134088387e-07,
|
| 808 |
+
"logits/chosen": 0.3840331435203552,
|
| 809 |
+
"logits/rejected": 0.5413318872451782,
|
| 810 |
+
"logps/chosen": -466.8184509277344,
|
| 811 |
+
"logps/rejected": -561.7572631835938,
|
| 812 |
+
"loss": 0.0145,
|
| 813 |
+
"rewards/accuracies": 0.78125,
|
| 814 |
+
"rewards/chosen": -1.635196328163147,
|
| 815 |
+
"rewards/margins": 1.1625127792358398,
|
| 816 |
+
"rewards/rejected": -2.7977089881896973,
|
| 817 |
"step": 520
|
| 818 |
},
|
| 819 |
{
|
| 820 |
+
"epoch": 1.11,
|
| 821 |
+
"learning_rate": 4.908466865911614e-07,
|
| 822 |
+
"logits/chosen": 0.3811998963356018,
|
| 823 |
+
"logits/rejected": 0.5391031503677368,
|
| 824 |
+
"logps/chosen": -462.8243103027344,
|
| 825 |
+
"logps/rejected": -599.1219482421875,
|
| 826 |
+
"loss": 0.0153,
|
| 827 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 828 |
+
"rewards/chosen": -1.5375772714614868,
|
| 829 |
+
"rewards/margins": 1.5291885137557983,
|
| 830 |
+
"rewards/rejected": -3.066765785217285,
|
| 831 |
"step": 530
|
| 832 |
},
|
| 833 |
{
|
| 834 |
+
"epoch": 1.13,
|
| 835 |
+
"learning_rate": 4.7255233006783624e-07,
|
| 836 |
+
"logits/chosen": 0.29220065474510193,
|
| 837 |
+
"logits/rejected": 0.5385439395904541,
|
| 838 |
+
"logps/chosen": -494.8135681152344,
|
| 839 |
+
"logps/rejected": -553.2826538085938,
|
| 840 |
+
"loss": 0.0152,
|
| 841 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 842 |
+
"rewards/chosen": -1.4295642375946045,
|
| 843 |
+
"rewards/margins": 1.265754222869873,
|
| 844 |
+
"rewards/rejected": -2.6953186988830566,
|
| 845 |
"step": 540
|
| 846 |
},
|
| 847 |
{
|
| 848 |
+
"epoch": 1.15,
|
| 849 |
+
"learning_rate": 4.5429476797886617e-07,
|
| 850 |
+
"logits/chosen": 0.36708998680114746,
|
| 851 |
+
"logits/rejected": 0.4897306561470032,
|
| 852 |
+
"logps/chosen": -459.0501403808594,
|
| 853 |
+
"logps/rejected": -559.3533935546875,
|
| 854 |
+
"loss": 0.0156,
|
| 855 |
+
"rewards/accuracies": 0.706250011920929,
|
| 856 |
+
"rewards/chosen": -1.6044361591339111,
|
| 857 |
+
"rewards/margins": 1.0187304019927979,
|
| 858 |
+
"rewards/rejected": -2.623166561126709,
|
| 859 |
"step": 550
|
| 860 |
},
|
| 861 |
{
|
| 862 |
+
"epoch": 1.17,
|
| 863 |
+
"learning_rate": 4.3609847514019763e-07,
|
| 864 |
+
"logits/chosen": 0.401399701833725,
|
| 865 |
+
"logits/rejected": 0.5879624485969543,
|
| 866 |
+
"logps/chosen": -483.4132385253906,
|
| 867 |
+
"logps/rejected": -575.8525390625,
|
| 868 |
+
"loss": 0.0143,
|
| 869 |
"rewards/accuracies": 0.800000011920929,
|
| 870 |
+
"rewards/chosen": -1.5152244567871094,
|
| 871 |
+
"rewards/margins": 1.3872954845428467,
|
| 872 |
+
"rewards/rejected": -2.902519702911377,
|
| 873 |
"step": 560
|
| 874 |
},
|
| 875 |
{
|
| 876 |
+
"epoch": 1.19,
|
| 877 |
+
"learning_rate": 4.179878442344892e-07,
|
| 878 |
+
"logits/chosen": 0.3999864161014557,
|
| 879 |
+
"logits/rejected": 0.5597594976425171,
|
| 880 |
+
"logps/chosen": -466.9307556152344,
|
| 881 |
+
"logps/rejected": -579.4730834960938,
|
| 882 |
+
"loss": 0.0142,
|
| 883 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 884 |
+
"rewards/chosen": -1.7734342813491821,
|
| 885 |
+
"rewards/margins": 1.3825814723968506,
|
| 886 |
+
"rewards/rejected": -3.1560158729553223,
|
| 887 |
"step": 570
|
| 888 |
},
|
| 889 |
{
|
| 890 |
+
"epoch": 1.21,
|
| 891 |
+
"learning_rate": 3.9998715311197783e-07,
|
| 892 |
+
"logits/chosen": 0.43183040618896484,
|
| 893 |
+
"logits/rejected": 0.49680033326148987,
|
| 894 |
+
"logps/chosen": -505.7025451660156,
|
| 895 |
+
"logps/rejected": -659.0694580078125,
|
| 896 |
+
"loss": 0.0109,
|
| 897 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 898 |
+
"rewards/chosen": -2.024803638458252,
|
| 899 |
+
"rewards/margins": 1.5267921686172485,
|
| 900 |
+
"rewards/rejected": -3.5515952110290527,
|
| 901 |
"step": 580
|
| 902 |
},
|
| 903 |
{
|
| 904 |
+
"epoch": 1.23,
|
| 905 |
+
"learning_rate": 3.821205322452863e-07,
|
| 906 |
+
"logits/chosen": 0.45015770196914673,
|
| 907 |
+
"logits/rejected": 0.6397759318351746,
|
| 908 |
+
"logps/chosen": -516.25830078125,
|
| 909 |
+
"logps/rejected": -631.279052734375,
|
| 910 |
+
"loss": 0.0115,
|
| 911 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 912 |
+
"rewards/chosen": -1.9833446741104126,
|
| 913 |
+
"rewards/margins": 1.3147289752960205,
|
| 914 |
+
"rewards/rejected": -3.2980740070343018,
|
| 915 |
"step": 590
|
| 916 |
},
|
| 917 |
{
|
| 918 |
+
"epoch": 1.26,
|
| 919 |
+
"learning_rate": 3.6441193238179146e-07,
|
| 920 |
+
"logits/chosen": 0.46848899126052856,
|
| 921 |
+
"logits/rejected": 0.5270673036575317,
|
| 922 |
+
"logps/chosen": -509.9579162597656,
|
| 923 |
+
"logps/rejected": -689.2035522460938,
|
| 924 |
+
"loss": 0.0145,
|
| 925 |
+
"rewards/accuracies": 0.824999988079071,
|
| 926 |
+
"rewards/chosen": -1.8605167865753174,
|
| 927 |
+
"rewards/margins": 1.5842502117156982,
|
| 928 |
+
"rewards/rejected": -3.4447669982910156,
|
| 929 |
"step": 600
|
| 930 |
},
|
| 931 |
{
|
| 932 |
+
"epoch": 1.26,
|
| 933 |
+
"eval_logits/chosen": 0.5416067838668823,
|
| 934 |
+
"eval_logits/rejected": 0.63544762134552,
|
| 935 |
+
"eval_logps/chosen": -482.76092529296875,
|
| 936 |
+
"eval_logps/rejected": -615.6082763671875,
|
| 937 |
+
"eval_loss": 0.024462632834911346,
|
| 938 |
+
"eval_rewards/accuracies": 0.734375,
|
| 939 |
+
"eval_rewards/chosen": -1.7917625904083252,
|
| 940 |
+
"eval_rewards/margins": 1.2480695247650146,
|
| 941 |
+
"eval_rewards/rejected": -3.039832353591919,
|
| 942 |
+
"eval_runtime": 75.4471,
|
| 943 |
+
"eval_samples_per_second": 26.509,
|
| 944 |
+
"eval_steps_per_second": 0.424,
|
| 945 |
+
"step": 600
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 1.28,
|
| 949 |
+
"learning_rate": 3.4688509243692034e-07,
|
| 950 |
+
"logits/chosen": 0.47085338830947876,
|
| 951 |
+
"logits/rejected": 0.5862728357315063,
|
| 952 |
+
"logps/chosen": -508.70098876953125,
|
| 953 |
+
"logps/rejected": -643.3338623046875,
|
| 954 |
+
"loss": 0.0121,
|
| 955 |
+
"rewards/accuracies": 0.831250011920929,
|
| 956 |
+
"rewards/chosen": -1.7418628931045532,
|
| 957 |
+
"rewards/margins": 1.5565435886383057,
|
| 958 |
+
"rewards/rejected": -3.2984061241149902,
|
| 959 |
"step": 610
|
| 960 |
},
|
| 961 |
{
|
| 962 |
+
"epoch": 1.3,
|
| 963 |
+
"learning_rate": 3.295635076714144e-07,
|
| 964 |
+
"logits/chosen": 0.3978613018989563,
|
| 965 |
+
"logits/rejected": 0.5926721096038818,
|
| 966 |
+
"logps/chosen": -494.2681579589844,
|
| 967 |
+
"logps/rejected": -604.7338256835938,
|
| 968 |
+
"loss": 0.0114,
|
| 969 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 970 |
+
"rewards/chosen": -1.8917449712753296,
|
| 971 |
+
"rewards/margins": 1.2589935064315796,
|
| 972 |
+
"rewards/rejected": -3.1507387161254883,
|
| 973 |
"step": 620
|
| 974 |
},
|
| 975 |
{
|
| 976 |
+
"epoch": 1.32,
|
| 977 |
+
"learning_rate": 3.12470398195219e-07,
|
| 978 |
+
"logits/chosen": 0.4904108941555023,
|
| 979 |
+
"logits/rejected": 0.6586324572563171,
|
| 980 |
+
"logps/chosen": -502.24322509765625,
|
| 981 |
+
"logps/rejected": -618.8930053710938,
|
| 982 |
+
"loss": 0.0112,
|
| 983 |
+
"rewards/accuracies": 0.737500011920929,
|
| 984 |
+
"rewards/chosen": -1.9003562927246094,
|
| 985 |
+
"rewards/margins": 1.4167784452438354,
|
| 986 |
+
"rewards/rejected": -3.3171348571777344,
|
| 987 |
"step": 630
|
| 988 |
},
|
| 989 |
{
|
| 990 |
+
"epoch": 1.34,
|
| 991 |
+
"learning_rate": 2.956286778402226e-07,
|
| 992 |
+
"logits/chosen": 0.4602645933628082,
|
| 993 |
+
"logits/rejected": 0.5689934492111206,
|
| 994 |
+
"logps/chosen": -449.55718994140625,
|
| 995 |
+
"logps/rejected": -578.3916625976562,
|
| 996 |
+
"loss": 0.0118,
|
| 997 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 998 |
+
"rewards/chosen": -1.755163550376892,
|
| 999 |
+
"rewards/margins": 1.4266632795333862,
|
| 1000 |
+
"rewards/rejected": -3.1818270683288574,
|
| 1001 |
"step": 640
|
| 1002 |
},
|
| 1003 |
{
|
| 1004 |
+
"epoch": 1.36,
|
| 1005 |
+
"learning_rate": 2.7906092344356826e-07,
|
| 1006 |
+
"logits/chosen": 0.4582684636116028,
|
| 1007 |
+
"logits/rejected": 0.5123009085655212,
|
| 1008 |
+
"logps/chosen": -464.2190856933594,
|
| 1009 |
+
"logps/rejected": -599.2928466796875,
|
| 1010 |
+
"loss": 0.0127,
|
| 1011 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1012 |
+
"rewards/chosen": -1.6637153625488281,
|
| 1013 |
+
"rewards/margins": 1.4032952785491943,
|
| 1014 |
+
"rewards/rejected": -3.0670104026794434,
|
| 1015 |
"step": 650
|
| 1016 |
},
|
| 1017 |
{
|
| 1018 |
+
"epoch": 1.38,
|
| 1019 |
+
"learning_rate": 2.6278934458271996e-07,
|
| 1020 |
+
"logits/chosen": 0.3585265576839447,
|
| 1021 |
+
"logits/rejected": 0.5152703523635864,
|
| 1022 |
+
"logps/chosen": -539.8528442382812,
|
| 1023 |
+
"logps/rejected": -647.3905029296875,
|
| 1024 |
+
"loss": 0.0106,
|
| 1025 |
+
"rewards/accuracies": 0.78125,
|
| 1026 |
+
"rewards/chosen": -1.8331562280654907,
|
| 1027 |
+
"rewards/margins": 1.5139731168746948,
|
| 1028 |
+
"rewards/rejected": -3.3471293449401855,
|
| 1029 |
"step": 660
|
| 1030 |
},
|
| 1031 |
{
|
| 1032 |
+
"epoch": 1.4,
|
| 1033 |
+
"learning_rate": 2.468357538028487e-07,
|
| 1034 |
+
"logits/chosen": 0.4794502854347229,
|
| 1035 |
+
"logits/rejected": 0.6796830892562866,
|
| 1036 |
+
"logps/chosen": -531.4798583984375,
|
| 1037 |
+
"logps/rejected": -636.687255859375,
|
| 1038 |
+
"loss": 0.0106,
|
| 1039 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1040 |
+
"rewards/chosen": -2.1033949851989746,
|
| 1041 |
+
"rewards/margins": 1.3440332412719727,
|
| 1042 |
+
"rewards/rejected": -3.4474282264709473,
|
| 1043 |
"step": 670
|
| 1044 |
},
|
| 1045 |
{
|
| 1046 |
+
"epoch": 1.42,
|
| 1047 |
+
"learning_rate": 2.312215373764551e-07,
|
| 1048 |
+
"logits/chosen": 0.48690152168273926,
|
| 1049 |
+
"logits/rejected": 0.6909142136573792,
|
| 1050 |
+
"logps/chosen": -482.9384765625,
|
| 1051 |
+
"logps/rejected": -627.9971313476562,
|
| 1052 |
+
"loss": 0.0105,
|
| 1053 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1054 |
+
"rewards/chosen": -1.8502315282821655,
|
| 1055 |
+
"rewards/margins": 1.5311414003372192,
|
| 1056 |
+
"rewards/rejected": -3.381373167037964,
|
| 1057 |
"step": 680
|
| 1058 |
},
|
| 1059 |
{
|
| 1060 |
+
"epoch": 1.44,
|
| 1061 |
+
"learning_rate": 2.1596762663442213e-07,
|
| 1062 |
+
"logits/chosen": 0.40491142868995667,
|
| 1063 |
+
"logits/rejected": 0.6269996166229248,
|
| 1064 |
+
"logps/chosen": -517.865234375,
|
| 1065 |
+
"logps/rejected": -638.0469970703125,
|
| 1066 |
+
"loss": 0.0118,
|
| 1067 |
+
"rewards/accuracies": 0.800000011920929,
|
| 1068 |
+
"rewards/chosen": -1.894121766090393,
|
| 1069 |
+
"rewards/margins": 1.5707839727401733,
|
| 1070 |
+
"rewards/rejected": -3.4649059772491455,
|
| 1071 |
"step": 690
|
| 1072 |
},
|
| 1073 |
{
|
| 1074 |
+
"epoch": 1.47,
|
| 1075 |
+
"learning_rate": 2.0109446990692963e-07,
|
| 1076 |
+
"logits/chosen": 0.4644289016723633,
|
| 1077 |
+
"logits/rejected": 0.5091214179992676,
|
| 1078 |
+
"logps/chosen": -506.99554443359375,
|
| 1079 |
+
"logps/rejected": -711.6946411132812,
|
| 1080 |
+
"loss": 0.0098,
|
| 1081 |
+
"rewards/accuracies": 0.793749988079071,
|
| 1082 |
+
"rewards/chosen": -1.9687172174453735,
|
| 1083 |
+
"rewards/margins": 1.7158300876617432,
|
| 1084 |
+
"rewards/rejected": -3.684546709060669,
|
| 1085 |
"step": 700
|
| 1086 |
},
|
| 1087 |
{
|
| 1088 |
+
"epoch": 1.47,
|
| 1089 |
+
"eval_logits/chosen": 0.5851482152938843,
|
| 1090 |
+
"eval_logits/rejected": 0.6870884299278259,
|
| 1091 |
+
"eval_logps/chosen": -507.2100830078125,
|
| 1092 |
+
"eval_logps/rejected": -646.5032958984375,
|
| 1093 |
+
"eval_loss": 0.02104853093624115,
|
| 1094 |
+
"eval_rewards/accuracies": 0.73046875,
|
| 1095 |
+
"eval_rewards/chosen": -2.0362541675567627,
|
| 1096 |
+
"eval_rewards/margins": 1.3125278949737549,
|
| 1097 |
+
"eval_rewards/rejected": -3.3487820625305176,
|
| 1098 |
+
"eval_runtime": 74.9024,
|
| 1099 |
+
"eval_samples_per_second": 26.701,
|
| 1100 |
+
"eval_steps_per_second": 0.427,
|
| 1101 |
+
"step": 700
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.49,
|
| 1105 |
+
"learning_rate": 1.8662200511184872e-07,
|
| 1106 |
+
"logits/chosen": 0.5449516773223877,
|
| 1107 |
+
"logits/rejected": 0.7054388523101807,
|
| 1108 |
+
"logps/chosen": -489.23712158203125,
|
| 1109 |
+
"logps/rejected": -618.7713623046875,
|
| 1110 |
+
"loss": 0.0106,
|
| 1111 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 1112 |
+
"rewards/chosen": -1.8367626667022705,
|
| 1113 |
+
"rewards/margins": 1.529603362083435,
|
| 1114 |
+
"rewards/rejected": -3.366366147994995,
|
| 1115 |
"step": 710
|
| 1116 |
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.51,
|
| 1119 |
+
"learning_rate": 1.725696330273575e-07,
|
| 1120 |
+
"logits/chosen": 0.4658689498901367,
|
| 1121 |
+
"logits/rejected": 0.5483459234237671,
|
| 1122 |
+
"logps/chosen": -499.9456481933594,
|
| 1123 |
+
"logps/rejected": -656.5256958007812,
|
| 1124 |
+
"loss": 0.0115,
|
| 1125 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1126 |
+
"rewards/chosen": -1.7960010766983032,
|
| 1127 |
+
"rewards/margins": 1.747243881225586,
|
| 1128 |
+
"rewards/rejected": -3.5432448387145996,
|
| 1129 |
+
"step": 720
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.53,
|
| 1133 |
+
"learning_rate": 1.589561912846089e-07,
|
| 1134 |
+
"logits/chosen": 0.5471647381782532,
|
| 1135 |
+
"logits/rejected": 0.7508308291435242,
|
| 1136 |
+
"logps/chosen": -486.8019104003906,
|
| 1137 |
+
"logps/rejected": -648.6491088867188,
|
| 1138 |
+
"loss": 0.012,
|
| 1139 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1140 |
+
"rewards/chosen": -1.7438154220581055,
|
| 1141 |
+
"rewards/margins": 1.5669682025909424,
|
| 1142 |
+
"rewards/rejected": -3.310783863067627,
|
| 1143 |
+
"step": 730
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.55,
|
| 1147 |
+
"learning_rate": 1.4579992911531496e-07,
|
| 1148 |
+
"logits/chosen": 0.43005961179733276,
|
| 1149 |
+
"logits/rejected": 0.5700550079345703,
|
| 1150 |
+
"logps/chosen": -507.7278747558594,
|
| 1151 |
+
"logps/rejected": -604.3984375,
|
| 1152 |
+
"loss": 0.0111,
|
| 1153 |
+
"rewards/accuracies": 0.78125,
|
| 1154 |
+
"rewards/chosen": -1.9238866567611694,
|
| 1155 |
+
"rewards/margins": 1.2898902893066406,
|
| 1156 |
+
"rewards/rejected": -3.2137768268585205,
|
| 1157 |
+
"step": 740
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.57,
|
| 1161 |
+
"learning_rate": 1.3311848288809813e-07,
|
| 1162 |
+
"logits/chosen": 0.5181199312210083,
|
| 1163 |
+
"logits/rejected": 0.7587286233901978,
|
| 1164 |
+
"logps/chosen": -517.6824951171875,
|
| 1165 |
+
"logps/rejected": -604.0447387695312,
|
| 1166 |
+
"loss": 0.0114,
|
| 1167 |
+
"rewards/accuracies": 0.75,
|
| 1168 |
+
"rewards/chosen": -1.8500522375106812,
|
| 1169 |
+
"rewards/margins": 1.3126447200775146,
|
| 1170 |
+
"rewards/rejected": -3.1626968383789062,
|
| 1171 |
+
"step": 750
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.59,
|
| 1175 |
+
"learning_rate": 1.209288524664029e-07,
|
| 1176 |
+
"logits/chosen": 0.38817617297172546,
|
| 1177 |
+
"logits/rejected": 0.6967793703079224,
|
| 1178 |
+
"logps/chosen": -534.35888671875,
|
| 1179 |
+
"logps/rejected": -617.2965087890625,
|
| 1180 |
+
"loss": 0.0116,
|
| 1181 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1182 |
+
"rewards/chosen": -1.964895248413086,
|
| 1183 |
+
"rewards/margins": 1.3300648927688599,
|
| 1184 |
+
"rewards/rejected": -3.2949604988098145,
|
| 1185 |
+
"step": 760
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 1.61,
|
| 1189 |
+
"learning_rate": 1.0924737841966497e-07,
|
| 1190 |
+
"logits/chosen": 0.657262921333313,
|
| 1191 |
+
"logits/rejected": 0.8421579599380493,
|
| 1192 |
+
"logps/chosen": -479.6363220214844,
|
| 1193 |
+
"logps/rejected": -551.2875366210938,
|
| 1194 |
+
"loss": 0.0107,
|
| 1195 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1196 |
+
"rewards/chosen": -1.988033652305603,
|
| 1197 |
+
"rewards/margins": 1.0586892366409302,
|
| 1198 |
+
"rewards/rejected": -3.046722888946533,
|
| 1199 |
+
"step": 770
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.63,
|
| 1203 |
+
"learning_rate": 9.808972011828054e-08,
|
| 1204 |
+
"logits/chosen": 0.534065842628479,
|
| 1205 |
+
"logits/rejected": 0.7022881507873535,
|
| 1206 |
+
"logps/chosen": -485.6368713378906,
|
| 1207 |
+
"logps/rejected": -590.6656494140625,
|
| 1208 |
+
"loss": 0.0112,
|
| 1209 |
+
"rewards/accuracies": 0.6875,
|
| 1210 |
+
"rewards/chosen": -1.9203450679779053,
|
| 1211 |
+
"rewards/margins": 1.1289684772491455,
|
| 1212 |
+
"rewards/rejected": -3.04931378364563,
|
| 1213 |
+
"step": 780
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.65,
|
| 1217 |
+
"learning_rate": 8.747083474174527e-08,
|
| 1218 |
+
"logits/chosen": 0.6039692163467407,
|
| 1219 |
+
"logits/rejected": 0.7411171197891235,
|
| 1220 |
+
"logps/chosen": -460.4100646972656,
|
| 1221 |
+
"logps/rejected": -635.5348510742188,
|
| 1222 |
+
"loss": 0.01,
|
| 1223 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 1224 |
+
"rewards/chosen": -1.8810646533966064,
|
| 1225 |
+
"rewards/margins": 1.626961350440979,
|
| 1226 |
+
"rewards/rejected": -3.508026123046875,
|
| 1227 |
+
"step": 790
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 1.67,
|
| 1231 |
+
"learning_rate": 7.740495722810269e-08,
|
| 1232 |
+
"logits/chosen": 0.35127943754196167,
|
| 1233 |
+
"logits/rejected": 0.535417914390564,
|
| 1234 |
+
"logps/chosen": -558.7343139648438,
|
| 1235 |
+
"logps/rejected": -669.262451171875,
|
| 1236 |
+
"loss": 0.0101,
|
| 1237 |
+
"rewards/accuracies": 0.78125,
|
| 1238 |
+
"rewards/chosen": -2.1135106086730957,
|
| 1239 |
+
"rewards/margins": 1.4616328477859497,
|
| 1240 |
+
"rewards/rejected": -3.575143337249756,
|
| 1241 |
+
"step": 800
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.67,
|
| 1245 |
+
"eval_logits/chosen": 0.5936872959136963,
|
| 1246 |
+
"eval_logits/rejected": 0.6981104016304016,
|
| 1247 |
+
"eval_logps/chosen": -502.8233642578125,
|
| 1248 |
+
"eval_logps/rejected": -638.039794921875,
|
| 1249 |
+
"eval_loss": 0.021807128563523293,
|
| 1250 |
+
"eval_rewards/accuracies": 0.734375,
|
| 1251 |
+
"eval_rewards/chosen": -1.992387056350708,
|
| 1252 |
+
"eval_rewards/margins": 1.2717599868774414,
|
| 1253 |
+
"eval_rewards/rejected": -3.2641468048095703,
|
| 1254 |
+
"eval_runtime": 73.4364,
|
| 1255 |
+
"eval_samples_per_second": 27.234,
|
| 1256 |
+
"eval_steps_per_second": 0.436,
|
| 1257 |
+
"step": 800
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.7,
|
| 1261 |
+
"learning_rate": 6.790558119157597e-08,
|
| 1262 |
+
"logits/chosen": 0.514393150806427,
|
| 1263 |
+
"logits/rejected": 0.7291263341903687,
|
| 1264 |
+
"logps/chosen": -525.2340698242188,
|
| 1265 |
+
"logps/rejected": -616.1982421875,
|
| 1266 |
+
"loss": 0.0106,
|
| 1267 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1268 |
+
"rewards/chosen": -1.9875078201293945,
|
| 1269 |
+
"rewards/margins": 1.2109524011611938,
|
| 1270 |
+
"rewards/rejected": -3.198460102081299,
|
| 1271 |
+
"step": 810
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.72,
|
| 1275 |
+
"learning_rate": 5.898544083397e-08,
|
| 1276 |
+
"logits/chosen": 0.5488702654838562,
|
| 1277 |
+
"logits/rejected": 0.6949459910392761,
|
| 1278 |
+
"logps/chosen": -483.0582580566406,
|
| 1279 |
+
"logps/rejected": -598.3394775390625,
|
| 1280 |
+
"loss": 0.0115,
|
| 1281 |
+
"rewards/accuracies": 0.71875,
|
| 1282 |
+
"rewards/chosen": -1.958407998085022,
|
| 1283 |
+
"rewards/margins": 1.3388407230377197,
|
| 1284 |
+
"rewards/rejected": -3.2972488403320312,
|
| 1285 |
+
"step": 820
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.74,
|
| 1289 |
+
"learning_rate": 5.065649387408705e-08,
|
| 1290 |
+
"logits/chosen": 0.5013612508773804,
|
| 1291 |
+
"logits/rejected": 0.5702673196792603,
|
| 1292 |
+
"logps/chosen": -528.1060180664062,
|
| 1293 |
+
"logps/rejected": -625.3968505859375,
|
| 1294 |
+
"loss": 0.012,
|
| 1295 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1296 |
+
"rewards/chosen": -2.1246533393859863,
|
| 1297 |
+
"rewards/margins": 1.261916995048523,
|
| 1298 |
+
"rewards/rejected": -3.3865699768066406,
|
| 1299 |
+
"step": 830
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.76,
|
| 1303 |
+
"learning_rate": 4.292990551804171e-08,
|
| 1304 |
+
"logits/chosen": 0.42700353264808655,
|
| 1305 |
+
"logits/rejected": 0.6631879806518555,
|
| 1306 |
+
"logps/chosen": -502.4617614746094,
|
| 1307 |
+
"logps/rejected": -618.699462890625,
|
| 1308 |
+
"loss": 0.0114,
|
| 1309 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1310 |
+
"rewards/chosen": -1.8777034282684326,
|
| 1311 |
+
"rewards/margins": 1.4012324810028076,
|
| 1312 |
+
"rewards/rejected": -3.2789359092712402,
|
| 1313 |
+
"step": 840
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.78,
|
| 1317 |
+
"learning_rate": 3.581603349196371e-08,
|
| 1318 |
+
"logits/chosen": 0.5486747026443481,
|
| 1319 |
+
"logits/rejected": 0.7163097262382507,
|
| 1320 |
+
"logps/chosen": -501.1570739746094,
|
| 1321 |
+
"logps/rejected": -640.0488891601562,
|
| 1322 |
+
"loss": 0.0102,
|
| 1323 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 1324 |
+
"rewards/chosen": -1.7478796243667603,
|
| 1325 |
+
"rewards/margins": 1.5438367128372192,
|
| 1326 |
+
"rewards/rejected": -3.2917160987854004,
|
| 1327 |
+
"step": 850
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.8,
|
| 1331 |
+
"learning_rate": 2.9324414157151367e-08,
|
| 1332 |
+
"logits/chosen": 0.4644288420677185,
|
| 1333 |
+
"logits/rejected": 0.6574305891990662,
|
| 1334 |
+
"logps/chosen": -473.5594787597656,
|
| 1335 |
+
"logps/rejected": -598.8511962890625,
|
| 1336 |
+
"loss": 0.0108,
|
| 1337 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1338 |
+
"rewards/chosen": -1.788891077041626,
|
| 1339 |
+
"rewards/margins": 1.2945648431777954,
|
| 1340 |
+
"rewards/rejected": -3.083456039428711,
|
| 1341 |
+
"step": 860
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 1.82,
|
| 1345 |
+
"learning_rate": 2.3463749726290284e-08,
|
| 1346 |
+
"logits/chosen": 0.49835458397865295,
|
| 1347 |
+
"logits/rejected": 0.6767226457595825,
|
| 1348 |
+
"logps/chosen": -532.7442016601562,
|
| 1349 |
+
"logps/rejected": -652.2874755859375,
|
| 1350 |
+
"loss": 0.0114,
|
| 1351 |
+
"rewards/accuracies": 0.75,
|
| 1352 |
+
"rewards/chosen": -1.964927315711975,
|
| 1353 |
+
"rewards/margins": 1.482194185256958,
|
| 1354 |
+
"rewards/rejected": -3.4471213817596436,
|
| 1355 |
+
"step": 870
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 1.84,
|
| 1359 |
+
"learning_rate": 1.824189659787284e-08,
|
| 1360 |
+
"logits/chosen": 0.432788610458374,
|
| 1361 |
+
"logits/rejected": 0.6490308046340942,
|
| 1362 |
+
"logps/chosen": -524.7098388671875,
|
| 1363 |
+
"logps/rejected": -685.0862426757812,
|
| 1364 |
+
"loss": 0.0119,
|
| 1365 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1366 |
+
"rewards/chosen": -1.886285424232483,
|
| 1367 |
+
"rewards/margins": 1.6881020069122314,
|
| 1368 |
+
"rewards/rejected": -3.5743870735168457,
|
| 1369 |
+
"step": 880
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.86,
|
| 1373 |
+
"learning_rate": 1.3665854824458035e-08,
|
| 1374 |
+
"logits/chosen": 0.45631036162376404,
|
| 1375 |
+
"logits/rejected": 0.6264600157737732,
|
| 1376 |
+
"logps/chosen": -514.7445068359375,
|
| 1377 |
+
"logps/rejected": -619.1950073242188,
|
| 1378 |
+
"loss": 0.011,
|
| 1379 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1380 |
+
"rewards/chosen": -1.918135643005371,
|
| 1381 |
+
"rewards/margins": 1.308524250984192,
|
| 1382 |
+
"rewards/rejected": -3.2266597747802734,
|
| 1383 |
+
"step": 890
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 1.88,
|
| 1387 |
+
"learning_rate": 9.741758728888217e-09,
|
| 1388 |
+
"logits/chosen": 0.4965108036994934,
|
| 1389 |
+
"logits/rejected": 0.7444375157356262,
|
| 1390 |
+
"logps/chosen": -533.3218994140625,
|
| 1391 |
+
"logps/rejected": -607.4019165039062,
|
| 1392 |
+
"loss": 0.0103,
|
| 1393 |
+
"rewards/accuracies": 0.768750011920929,
|
| 1394 |
+
"rewards/chosen": -2.0954043865203857,
|
| 1395 |
+
"rewards/margins": 1.2425765991210938,
|
| 1396 |
+
"rewards/rejected": -3.3379809856414795,
|
| 1397 |
+
"step": 900
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 1.88,
|
| 1401 |
+
"eval_logits/chosen": 0.610427975654602,
|
| 1402 |
+
"eval_logits/rejected": 0.7138903141021729,
|
| 1403 |
+
"eval_logps/chosen": -503.7669677734375,
|
| 1404 |
+
"eval_logps/rejected": -638.1763305664062,
|
| 1405 |
+
"eval_loss": 0.021610137075185776,
|
| 1406 |
+
"eval_rewards/accuracies": 0.734375,
|
| 1407 |
+
"eval_rewards/chosen": -2.0018229484558105,
|
| 1408 |
+
"eval_rewards/margins": 1.2636897563934326,
|
| 1409 |
+
"eval_rewards/rejected": -3.2655129432678223,
|
| 1410 |
+
"eval_runtime": 74.571,
|
| 1411 |
+
"eval_samples_per_second": 26.82,
|
| 1412 |
+
"eval_steps_per_second": 0.429,
|
| 1413 |
+
"step": 900
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.9,
|
| 1417 |
+
"learning_rate": 6.474868681043577e-09,
|
| 1418 |
+
"logits/chosen": 0.5612128973007202,
|
| 1419 |
+
"logits/rejected": 0.8626712560653687,
|
| 1420 |
+
"logps/chosen": -493.08551025390625,
|
| 1421 |
+
"logps/rejected": -595.2066650390625,
|
| 1422 |
+
"loss": 0.0104,
|
| 1423 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 1424 |
+
"rewards/chosen": -1.881819486618042,
|
| 1425 |
+
"rewards/margins": 1.3230464458465576,
|
| 1426 |
+
"rewards/rejected": -3.2048659324645996,
|
| 1427 |
+
"step": 910
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.93,
|
| 1431 |
+
"learning_rate": 3.869564046156459e-09,
|
| 1432 |
+
"logits/chosen": 0.33559730648994446,
|
| 1433 |
+
"logits/rejected": 0.5526424646377563,
|
| 1434 |
+
"logps/chosen": -545.949951171875,
|
| 1435 |
+
"logps/rejected": -630.3507080078125,
|
| 1436 |
+
"loss": 0.0106,
|
| 1437 |
+
"rewards/accuracies": 0.737500011920929,
|
| 1438 |
+
"rewards/chosen": -2.0992321968078613,
|
| 1439 |
+
"rewards/margins": 1.3220652341842651,
|
| 1440 |
+
"rewards/rejected": -3.421297550201416,
|
| 1441 |
+
"step": 920
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.95,
|
| 1445 |
+
"learning_rate": 1.929337314139412e-09,
|
| 1446 |
+
"logits/chosen": 0.3952362537384033,
|
| 1447 |
+
"logits/rejected": 0.6363921165466309,
|
| 1448 |
+
"logps/chosen": -528.9135131835938,
|
| 1449 |
+
"logps/rejected": -657.9598388671875,
|
| 1450 |
+
"loss": 0.0099,
|
| 1451 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1452 |
+
"rewards/chosen": -1.9300628900527954,
|
| 1453 |
+
"rewards/margins": 1.5338274240493774,
|
| 1454 |
+
"rewards/rejected": -3.4638900756835938,
|
| 1455 |
+
"step": 930
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 1.97,
|
| 1459 |
+
"learning_rate": 6.567894177967325e-10,
|
| 1460 |
+
"logits/chosen": 0.44664573669433594,
|
| 1461 |
+
"logits/rejected": 0.6714277267456055,
|
| 1462 |
+
"logps/chosen": -544.5101928710938,
|
| 1463 |
+
"logps/rejected": -673.3298950195312,
|
| 1464 |
+
"loss": 0.0113,
|
| 1465 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1466 |
+
"rewards/chosen": -2.0046563148498535,
|
| 1467 |
+
"rewards/margins": 1.3531763553619385,
|
| 1468 |
+
"rewards/rejected": -3.35783314704895,
|
| 1469 |
+
"step": 940
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.99,
|
| 1473 |
+
"learning_rate": 5.3626246194704575e-11,
|
| 1474 |
+
"logits/chosen": 0.5509423017501831,
|
| 1475 |
+
"logits/rejected": 0.6711171865463257,
|
| 1476 |
+
"logps/chosen": -456.3231506347656,
|
| 1477 |
+
"logps/rejected": -598.7819213867188,
|
| 1478 |
+
"loss": 0.0095,
|
| 1479 |
+
"rewards/accuracies": 0.78125,
|
| 1480 |
+
"rewards/chosen": -1.7993762493133545,
|
| 1481 |
+
"rewards/margins": 1.4040424823760986,
|
| 1482 |
+
"rewards/rejected": -3.203418731689453,
|
| 1483 |
+
"step": 950
|
| 1484 |
+
},
|
| 1485 |
{
|
| 1486 |
"epoch": 2.0,
|
| 1487 |
+
"step": 954,
|
| 1488 |
"total_flos": 0.0,
|
| 1489 |
+
"train_loss": 0.048426837674039463,
|
| 1490 |
+
"train_runtime": 8860.0868,
|
| 1491 |
+
"train_samples_per_second": 13.8,
|
| 1492 |
+
"train_steps_per_second": 0.108
|
| 1493 |
}
|
| 1494 |
],
|
| 1495 |
"logging_steps": 10,
|
| 1496 |
+
"max_steps": 954,
|
| 1497 |
"num_train_epochs": 2,
|
| 1498 |
+
"save_steps": 1000,
|
| 1499 |
"total_flos": 0.0,
|
| 1500 |
"trial_name": null,
|
| 1501 |
"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 6648
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ebe8424565caecb3908e9e9e2a372a3ad6da08c02b93398d33a8308351a9f5c
|
| 3 |
size 6648
|