Add link to Hugging Face paper page

README.md

@@ -1,12 +1,12 @@
 ---
-license: apache-2.0
 base_model:
 - Qwen/Qwen2.5-7B-Instruct
 - qqjz/WhisperPro
-library_name: pytorch
-pipeline_tag: audio-text-to-text
 language:
 - en
+library_name: pytorch
+license: apache-2.0
+pipeline_tag: audio-text-to-text
 tags:
 - textpro-slm
 - prosody
@@ -14,13 +14,14 @@ tags:
 - audio
 - original-checkpoint
 ---
+
 # TextPro-SLM-7B-Global
 
 **Your speech LLM can be a prosody-aware text LLM.**
 
 TextPro-SLM combines transcript tokens with a compact prosody embedding. Transcript tokens carry linguistic content, while prosody carries complementary acoustic information. This input-side factorization brings spoken input closer to the native input format of a text LLM.
 
-[Paper](https://openreview.net/forum?id=Jsjmo8eX29) · [GitHub project](https://github.com/dreamtheater123/TextPro-SLM) · [Paired checkpoint](https://huggingface.co/qqjz/WhisperPro)
+[Paper](https://huggingface.co/papers/2605.05927) · [OpenReview](https://openreview.net/forum?id=Jsjmo8eX29) · [GitHub project](https://github.com/dreamtheater123/TextPro-SLM) · [Paired checkpoint](https://huggingface.co/qqjz/WhisperPro)
 
 ## Release and pairing
 
@@ -86,4 +87,4 @@ Apache-2.0; see `LICENSE` and `NOTICE`. The LLM is derived from [Qwen2.5-7B-Inst
   year={2026},
   url={https://openreview.net/forum?id=Jsjmo8eX29}
 }
-```
+```