-
Update README.md
Fix error -
Update README.md
push mode weights
-
Upload hcodec
Upload hcodec -
Upload read me again
Upload read me again
-
Upload weights_small_data_20wh.pt
Upload 20wh trainning dataset ckpt -
Upload HCodec.jpg
upload arch
-
Upload Hcodec readme
 -
Update README.md
# QuarkAudio-HCodec: A Unified Discrete Audio Tokenizer for High-Fidelity, Multitask Audio Generation
<p align="center">
<a href="https://arxiv.org/abs/2510.26372">
<img src="https://img.shields.io/badge/Paper-ArXiv-red.svg" alt="Paper">
</a>
<a href="https://huggingface.co/spaces/QuarkAudio/">
<img src="https://img.shields.io/badge/Model-Hugging%20Face-yellow.svg" alt="Hugging Face">
</a>
</p>
<p align="center">
<a href="https://arxiv.org/abs/2510.26372"><img src="HCodec.jpg" width="70%" /></a>
</p>
> 🔊 **H-Codec**: *A Unified, Dual-Stream Neural Audio Codec with Adaptive Frame Rate and 48kHz Support*
> Enabling high-fidelity, efficient, and semantically rich audio tokenization for next-generation LLM-based audio generation.
🚀 **Key Highlights**:
- ✅ **Dual-Stream Tokenization**: Separately quantizes acoustic and semantic features into independent codebooks — preserving both signal fidelity and linguistic content.
- 🔄 **Dynamic Frame Rate (H-Codec-1.5)**: In… -
initial commit
