# cilla-mini

> Gated, non-commercial (CC BY-NC 4.0): ask for access on this page. Cilla Mini is the 18M model for CPUs, phones and browsers: Cilla shrunk with Kyutai's small-model recipe (8 layers, a 64-dim interface, a one-step drifting sampler head, the 2.6M Mimi decoder), guidance baked in.

- **Organisation:** mehdi-hf
- **Category:** Text-to-Speech
- **Parameters:** 28M
- **Licence:** CC BY-NC 4.0 (Non-commercial)
- **Released:** 2026-11-16
- **Hugging Face:** 0 downloads (30 days), 66 likes

## Estimated memory

| Quant | ≈ GB |
|---|---|
| Q4_K_M | 0.52 |
| Q8_0 | 0.53 |
| F16 | 0.56 |

## About

Gated, non-commercial (CC BY-NC 4.0): ask for access on this page. Cilla Mini is the 18M model for CPUs, phones and browsers: Cilla shrunk with Kyutai's small-model recipe (8 layers, a 64-dim interface, a one-step drifting sampler head, the 2.6M Mimi decoder), guidance baked in. It speaks Persian in the voice of a short audio prompt. Built on Pocket TTS by Kyutai.

Try it in your browser · Run it on a Mac · Phonon runtime · Browser library · Python package

Audio made per second of work ("real time" is 1x):

speak_fa.py normalises the text, speaks one sentence at a time from the same prompt, trims the silence each piece opens with and joins them with a short pause. Do the same in your own code: cutting a sentence into 50-token pieces (the old Pocket TTS default) made these models repeat words, once the name «ساندر پیچای» twice in a row.

The released package is enough (pocket-tts 4.0.1 or newer, checked for this repo). config.yaml sets capitalize_first_letter, append_terminal_punctuation and pad_with_spaces_for_short_inputs to false: leave them, otherwise the package upper-cases or drops the first letters of Persian text.

The phonon/ folder has the model for the Phonon runtime (Rust, no PyTorch, no GPU) and the same Persian front-end. Apple silicon:

Details, options and notes: phonon/README.md.

## Alternatives

- [MBZUAI/LLMVoX](https://selfhost.directory/models/mbzuai/llmvox) (30M, CC BY-NC-SA 4.0, ≈0.57 GB): This repository contains the model as described in LLMVoX: Autoregressive Streaming Text-to-Speech Model for Any LLM.
- [hexgrad/Kokoro-82M](https://selfhost.directory/models/hexgrad/kokoro-82m) (82M, Apache 2.0, ≈0.68 GB): Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers comparable quality to larger models while being significantly faster and more cost-efficient.
- [zaakirio/kokoro-ru](https://selfhost.directory/models/zaakirio/kokoro-ru) (82M, OpenRAIL, ≈0.68 GB): Russian text-to-speech at 82M parameters, running 9.8x faster than realtime on a laptop CPU with no GPU. Three fixed voices from consented studio actors.
- [facebook/mms-tts-eng](https://selfhost.directory/models/facebook/mms-tts-eng) (36M, CC BY-NC 4.0, ≈0.58 GB): This repository contains the English (eng) language text-to-speech (TTS) model checkpoint.
- [KittenML/kitten-tts-2-mini](https://selfhost.directory/models/kittenml/kitten-tts-2-mini) (82M, Apache 2.0, ≈0.68 GB): 82M parameters, 16 languages, 118 voices, 24 kHz. ONNX with int8 weights (82.7 MB; runs on CPU, no PyTorch needed), and the float32 PyTorch weights.
- [OpenMOSS-Team/MOSS-TTS-Nano-100M](https://selfhost.directory/models/openmoss-team/moss-tts-nano-100m) (100M, Apache 2.0, ≈0.72 GB): MOSS-TTS-Nano is an open-source multilingual tiny speech generation model from MOSI.AI and the OpenMOSS team. With only 0.1B parameters, it is designed for realtime speech generation, can run directly on CPU without a GPU, and keeps the deployment stack simple enough for local demos, web serving…

More: https://selfhost.directory/models/mehdi-hf/cilla-mini/alternatives

## Runs well on

- [GeForce RTX 4060](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/rtx-4060): F16
- [Apple M4 (16 GB)](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/apple-m4-16gb): F16
- [Radeon RX 6700 XT](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/rx-6700-xt): F16
- [GeForce RTX 3060 12GB](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/rtx-3060): F16
- [Arc B580](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/arc-b580): F16
- [Tesla T4](https://selfhost.directory/models/mehdi-hf/cilla-mini/on/tesla-t4): F16

## Similar size

- [facebook/mms-tts-eng](https://selfhost.directory/models/facebook/mms-tts-eng) (36M)
- [facebook/mms-tts-vie](https://selfhost.directory/models/facebook/mms-tts-vie) (36M)
- [facebook/mms-tts-hin](https://selfhost.directory/models/facebook/mms-tts-hin) (36M)
- [facebook/mms-tts-spa](https://selfhost.directory/models/facebook/mms-tts-spa) (36M)
- [facebook/mms-tts-ara](https://selfhost.directory/models/facebook/mms-tts-ara) (36M)
- [facebook/mms-tts-rus](https://selfhost.directory/models/facebook/mms-tts-rus) (36M)

## FAQ

**How much VRAM does cilla-mini need?**
About 0.6 GB at FP16 / BF16.

**Can cilla-mini run on the RTX 4060?**
Yes. It fits at FP16 / BF16 (about 0.6 GB). It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.

**Can I use cilla-mini commercially?**
No. It is released under CC BY-NC 4.0. Non-commercial use only, with attribution.

## Links

- huggingface: https://huggingface.co/mehdi-hf/cilla-mini
- demo: https://huggingface.co/spaces/mehdi-hf/cilla-mini-demo

---
Source: https://selfhost.directory/models/mehdi-hf/cilla-mini