Add metadata and correct paper link
Browse filesThis PR adds relevant metadata such as the pipeline tag, library name and license, and corrects the link to the technical paper. It also adds a sample usage.
README.md
CHANGED
@@ -1,13 +1,16 @@
|
|
1 |
---
|
2 |
-
license: apache-2.0
|
3 |
base_model:
|
4 |
- Qwen/Qwen2.5-32B-Instruct
|
|
|
|
|
|
|
5 |
---
|
|
|
6 |
# Light-R1: Surpassing R1-Distill from Scratch\* with \$1000 through Curriculum SFT & DPO
|
7 |
|
8 |
*\*from models without long COT*
|
9 |
|
10 |
-
[technical report](https://
|
11 |
|
12 |
[GitHub page](https://github.com/Qihoo360/Light-R1)
|
13 |
|
@@ -128,4 +131,4 @@ Training data are collected from various public sources.
|
|
128 |
archivePrefix={},
|
129 |
url={https://github.com/Qihoo360/Light-R1},
|
130 |
}
|
131 |
-
```
|
|
|
1 |
---
|
|
|
2 |
base_model:
|
3 |
- Qwen/Qwen2.5-32B-Instruct
|
4 |
+
license: apache-2.0
|
5 |
+
library_name: transformers
|
6 |
+
pipeline_tag: text-generation
|
7 |
---
|
8 |
+
|
9 |
# Light-R1: Surpassing R1-Distill from Scratch\* with \$1000 through Curriculum SFT & DPO
|
10 |
|
11 |
*\*from models without long COT*
|
12 |
|
13 |
+
[technical report](https://huggingface.co/papers/2503.10460)
|
14 |
|
15 |
[GitHub page](https://github.com/Qihoo360/Light-R1)
|
16 |
|
|
|
131 |
archivePrefix={},
|
132 |
url={https://github.com/Qihoo360/Light-R1},
|
133 |
}
|
134 |
+
```
|