---
license: cc-by-nc-sa-4.0
pipeline_tag: text-to-speech
library_name: rlx
tags:
- miratts
- tts
- rlx
---

# MiraTTS (RLX staging)

MiraTTS LM + ONNX decoders for RLX. Non-commercial license — see upstream.

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

## Quick start

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

## File highlights

- `model.safetensors` (966.4 MiB)
- `MiraTTS.Q4_K_M.gguf` (392.6 MiB)
- `decoders/detokenizer.onnx` (183.8 MiB)
- `decoders/q_encoder.onnx` (116.7 MiB)
- `decoders/processer.onnx` (83.4 MiB)
- `decoders/s_encoder.onnx` (22.7 MiB)
- `tokenizer.json` (13.4 MiB)
- `vocab.json` (2.6 MiB)
- `tokenizer_config.json` (2.4 MiB)
- `merges.txt` (1.6 MiB)
- `added_tokens.json` (460.6 KiB)
- `config.json` (1.3 KiB)
- `special_tokens_map.json` (613 B)
- `generation_config.json` (170 B)

## Run with RLX

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

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

## License

CC-BY-NC-SA-4.0 — see upstream license; non-commercial share-alike.

Original weights and authorship: https://huggingface.co/YatharthS/MiraTTS

## 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.