---
datasets:
- SparkAudio/voxbox
language:
- zh
- en
base_model:
- BlinkDL/rwkv7-g1
pipeline_tag: text-to-speech
library_name: transformers
---


Respark is a TTS system with RWKV-7 LM modeling the audio tokens with the BiCodec from SparkTTS. This repository will contain a serie of respark models with different size and training stage. Please checkout the sub directories for details.