Add link to Github repository
#3
by
nielsr
HF Staff
- opened
README.md
CHANGED
|
@@ -17,6 +17,8 @@ DuoGuard-0.5B is built upon Qwen 2.5 (0.5B), a multilingual large language model
|
|
| 17 |
|
| 18 |
-It is presented in the paper [DuoGuard: A Two-Player RL-Driven Framework for Multilingual LLM Guardrails](https://huggingface.co/papers/2502.05163)
|
| 19 |
|
|
|
|
|
|
|
| 20 |
## How to Use
|
| 21 |
A quick code snippet or set of instructions on how to load and use the model in an application:
|
| 22 |
```python
|
|
@@ -72,7 +74,7 @@ category_names = [
|
|
| 72 |
prob_vector = probabilities[0].tolist() # shape: (12,)
|
| 73 |
|
| 74 |
predicted_labels = []
|
| 75 |
-
for cat_name, prob in zip(category_names, prob_vector):
|
| 76 |
label = 1 if prob > threshold else 0
|
| 77 |
predicted_labels.append(label)
|
| 78 |
|
|
|
|
| 17 |
|
| 18 |
-It is presented in the paper [DuoGuard: A Two-Player RL-Driven Framework for Multilingual LLM Guardrails](https://huggingface.co/papers/2502.05163)
|
| 19 |
|
| 20 |
+
For code, see https://github.com/yihedeng9/DuoGuard.
|
| 21 |
+
|
| 22 |
## How to Use
|
| 23 |
A quick code snippet or set of instructions on how to load and use the model in an application:
|
| 24 |
```python
|
|
|
|
| 74 |
prob_vector = probabilities[0].tolist() # shape: (12,)
|
| 75 |
|
| 76 |
predicted_labels = []
|
| 77 |
+
for cat_name, prob, label in zip(category_names, prob_vector):
|
| 78 |
label = 1 if prob > threshold else 0
|
| 79 |
predicted_labels.append(label)
|
| 80 |
|