Update README.md
Browse files
README.md
CHANGED
|
@@ -11,6 +11,32 @@ tags:
|
|
| 11 |
---
|
| 12 |
<h1 align="center">PrismAudio</h1>
|
| 13 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
---
|
| 15 |
|
| 16 |
**PrismAudio** is the first framework to integrate Reinforcement Learning into Video-to-Audio (V2A) generation with specialized Chain-of-Thought (CoT) planning. Building upon [ThinkSound](https://arxiv.org/pdf/2506.21448)'s pioneering CoT-based V2A framework, PrismAudio further decomposes monolithic reasoning into four specialized CoT modules (Semantic, Temporal, Aesthetic, and Spatial), each paired with targeted reward functions, enabling multi-dimensional RL optimization that jointly improves reasoning across all perceptual dimensions.
|
|
|
|
| 11 |
---
|
| 12 |
<h1 align="center">PrismAudio</h1>
|
| 13 |
|
| 14 |
+
<p align="center">
|
| 15 |
+
<img src="https://img.shields.io/badge/ICLR 2026-Main Conference-blue.svg" alt="ICLR 2026"/>
|
| 16 |
+
</p>
|
| 17 |
+
<a href="https://arxiv.org/abs/2511.18833">
|
| 18 |
+
<img src="https://img.shields.io/badge/arXiv-2511.18833-b31b1b.svg" alt="arXiv"/>
|
| 19 |
+
</a>
|
| 20 |
+
|
| 21 |
+
<a href="http://prismaudio-project.github.io/">
|
| 22 |
+
<img src="https://img.shields.io/badge/Online%20Demo-🌐-blue" alt="Online Demo"/>
|
| 23 |
+
</a>
|
| 24 |
+
|
| 25 |
+
<a href="https://huggingface.co/spaces/FunAudioLLM/PrismAudio">
|
| 26 |
+
<img src="https://img.shields.io/badge/HuggingFace-Spaces-orange?logo=huggingface" alt="Hugging Face"/>
|
| 27 |
+
</a>
|
| 28 |
+
|
| 29 |
+
<a href="https://www.modelscope.cn/studios/iic/PrismAudio">
|
| 30 |
+
<img src="https://img.shields.io/badge/ModelScope-在线体验-green" alt="ModelScope"/>
|
| 31 |
+
</a>
|
| 32 |
+
</p>
|
| 33 |
+
|
| 34 |
+
<p align="center">
|
| 35 |
+
If you find this project useful,<br>
|
| 36 |
+
a star ⭐ on GitHub would be greatly appreciated!
|
| 37 |
+
</p>
|
| 38 |
+
|
| 39 |
+
|
| 40 |
---
|
| 41 |
|
| 42 |
**PrismAudio** is the first framework to integrate Reinforcement Learning into Video-to-Audio (V2A) generation with specialized Chain-of-Thought (CoT) planning. Building upon [ThinkSound](https://arxiv.org/pdf/2506.21448)'s pioneering CoT-based V2A framework, PrismAudio further decomposes monolithic reasoning into four specialized CoT modules (Semantic, Temporal, Aesthetic, and Spatial), each paired with targeted reward functions, enabling multi-dimensional RL optimization that jointly improves reasoning across all perceptual dimensions.
|