README.md

---
license: apache-2.0
pipeline_tag: text-to-speech
library_name: rlx
tags:
- gepard
- tts
- rlx
---

# Gepard-1.0 (RLX staging)

Gepard TTS + NanoCodec decoder weights for RLX.

| Field | Value |
|---|---|
| **Hub id** | [`eugenehp/gepard`](https://huggingface.co/eugenehp/gepard) |
| **Kind** | Staging redistrib of an upstream checkpoint for RLX runners. |
| **RLX crate** | [`rlx-gepard`](https://github.com/MIT-RLX/rlx-models/tree/main/crates/rlx-gepard) |
| **Upstream** | https://huggingface.co/nineninesix/gepard-1.0 |

## Quick start

```bash
hf download eugenehp/gepard --local-dir .
cargo run -p rlx-gepard --release -- --weights . --text "Hello." --out out.wav
```

## File highlights

- `gepard.f16.gguf` (1.4 GiB)
- `model.safetensors` (1.0 GiB)
- `nano_dec.safetensors` (400.4 MiB)
- `weights.safetensors` (169.1 MiB)
- `nano_dec_1.89kbps.safetensors` (120.4 MiB)
- `tokenizer.json` (19.1 MiB)
- `gepard_config.json` (4.3 KiB)
- `tokenizer_config.json` (1.1 KiB)

## 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:

- `gepard.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/gepard` (or pass the path explicitly), then:

```bash
cargo run -p rlx-gepard --release -- --weights . --text "Hello." --out out.wav
```

## License

Apache License 2.0 — see `LICENSE`. Inherit upstream terms when redistributing.

Original weights and authorship: https://huggingface.co/nineninesix/gepard-1.0

## 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.