---
license: mit
pipeline_tag: text-to-speech
library_name: rlx
tags:
- openvoice
- onnx
- rlx
---

# OpenVoice ONNX v2 (RLX staging)

OpenVoice tone-color ONNX assets for RLX (pairs with TinyTTS).

| Field | Value |
|---|---|
| **Hub id** | [`eugenehp/openvoice`](https://huggingface.co/eugenehp/openvoice) |
| **Kind** | Staging redistrib of an upstream checkpoint for RLX runners. |
| **RLX crate** | [`rlx-openvoice`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-openvoice) |
| **Upstream** | https://huggingface.co/Hinotsuba/OpenVoice-ONNX-v2 |

## Quick start

```bash
hf download eugenehp/openvoice --local-dir .
cargo run -p rlx-openvoice --release -- --model-dir .
```

## File highlights

- `tone_color.onnx` (149.9 MiB)
- `tone_extract.onnx` (3.2 MiB)
- `weights.safetensors` (2.0 MiB)
- `openvoice.f16.gguf` (1020.6 KiB)
- `tone_config.json` (838 B)
- `manifest.json` (254 B)

## Do not use as the main runtime pack

These files may be present for historical / staging reasons but are **not** the supported RLX load path:

- `openvoice.f16.gguf`

Community `*.f16.gguf` / LM-only Q4 packs are format wraps — they are not drop-in replacements for the RLX primary file above.

## Run with RLX

Clone [rlx-models](https://github.com/MIT-RLX/rlx-models), place this repo under `weights/tts/openvoice` (or pass the path explicitly), then:

```bash
cargo run -p rlx-openvoice --release -- --model-dir .
```

## License

MIT — see `LICENSE`. Inherit upstream terms when redistributing.

Original weights and authorship: https://huggingface.co/Hinotsuba/OpenVoice-ONNX-v2

## Redistrib note

This Hub repo exists so RLX recipes have a stable fetch target. When you only need the upstream checkpoint, prefer the Upstream link above.