Update MODEL_CARD.md
Browse files- MODEL_CARD.md +7 -7
MODEL_CARD.md
CHANGED
|
@@ -8,7 +8,7 @@ tags:
|
|
| 8 |
- outcome-reward-model
|
| 9 |
- pairwise-preference
|
| 10 |
datasets:
|
| 11 |
-
-
|
| 12 |
metrics:
|
| 13 |
- accuracy
|
| 14 |
pipeline_tag: text-classification
|
|
@@ -20,7 +20,7 @@ model-index:
|
|
| 20 |
name: Pairwise Preference Ranking
|
| 21 |
dataset:
|
| 22 |
name: ORM Pairwise Preference Pairs
|
| 23 |
-
type:
|
| 24 |
metrics:
|
| 25 |
- type: accuracy
|
| 26 |
value: 96.3
|
|
@@ -38,7 +38,7 @@ model-index:
|
|
| 38 |
|
| 39 |
[](link-to-arxiv)
|
| 40 |
[](https://huggingface.co/datasets/akleshmishra/orm-pairwise-preference-pairs)
|
| 41 |
-
[](
|
| 42 |
|
| 43 |
</div>
|
| 44 |
|
|
@@ -131,7 +131,7 @@ Each pair contains:
|
|
| 131 |
### Training Configuration
|
| 132 |
|
| 133 |
**Hyperparameters:**
|
| 134 |
-
- **Base Model**:
|
| 135 |
- **Trainable Parameters**: Scoring head only (~500K-1M params)
|
| 136 |
- **Optimizer**: AdamW
|
| 137 |
- Learning rate: 1e-4
|
|
@@ -343,15 +343,15 @@ If you use this model in your research, please cite:
|
|
| 343 |
## π Resources
|
| 344 |
|
| 345 |
- π **Paper**: [ArXiv](link-to-arxiv) (Coming soon)
|
| 346 |
-
- πΎ **Dataset**: [HuggingFace](https://huggingface.co/datasets/
|
| 347 |
-
- π» **Code**: [GitHub](
|
| 348 |
- π **Training Logs**: [Weights & Biases](wandb-link) (if available)
|
| 349 |
|
| 350 |
## π§ Contact
|
| 351 |
|
| 352 |
**Aklesh Mishra**
|
| 353 |
- Email: akleshmishra7@gmail.com
|
| 354 |
-
- GitHub: [@
|
| 355 |
|
| 356 |
## π License
|
| 357 |
|
|
|
|
| 8 |
- outcome-reward-model
|
| 9 |
- pairwise-preference
|
| 10 |
datasets:
|
| 11 |
+
- LossFunctionLover/orm-pairwise-preference-pairs
|
| 12 |
metrics:
|
| 13 |
- accuracy
|
| 14 |
pipeline_tag: text-classification
|
|
|
|
| 20 |
name: Pairwise Preference Ranking
|
| 21 |
dataset:
|
| 22 |
name: ORM Pairwise Preference Pairs
|
| 23 |
+
type: LossFunctionLover/orm-pairwise-preference-pairs
|
| 24 |
metrics:
|
| 25 |
- type: accuracy
|
| 26 |
value: 96.3
|
|
|
|
| 38 |
|
| 39 |
[](link-to-arxiv)
|
| 40 |
[](https://huggingface.co/datasets/akleshmishra/orm-pairwise-preference-pairs)
|
| 41 |
+
[](https://github.com/Coder-12)
|
| 42 |
|
| 43 |
</div>
|
| 44 |
|
|
|
|
| 131 |
### Training Configuration
|
| 132 |
|
| 133 |
**Hyperparameters:**
|
| 134 |
+
- **Base Model**: facebook/OPT1.3b
|
| 135 |
- **Trainable Parameters**: Scoring head only (~500K-1M params)
|
| 136 |
- **Optimizer**: AdamW
|
| 137 |
- Learning rate: 1e-4
|
|
|
|
| 343 |
## π Resources
|
| 344 |
|
| 345 |
- π **Paper**: [ArXiv](link-to-arxiv) (Coming soon)
|
| 346 |
+
- πΎ **Dataset**: [HuggingFace](https://huggingface.co/datasets/LossFunctionLover/orm-pairwise-preference-pairs)
|
| 347 |
+
- π» **Code**: [GitHub](https://github.com/Coder-12)
|
| 348 |
- π **Training Logs**: [Weights & Biases](wandb-link) (if available)
|
| 349 |
|
| 350 |
## π§ Contact
|
| 351 |
|
| 352 |
**Aklesh Mishra**
|
| 353 |
- Email: akleshmishra7@gmail.com
|
| 354 |
+
- GitHub: [@Coder-12](https://github.com/Coder-12)
|
| 355 |
|
| 356 |
## π License
|
| 357 |
|