Skip to content
Audio

Africa TTS

TTS для 92 африканских языков и английского

2 min read

Africa TTS от AfriSpeech — компактная TTS-модель на 110M параметров для 92 африканских языков и английского. Основана на 6-слойной Pocket TTS FlowLM, умеет клонировать голос по клипу 3–10 с и по умолчанию работает на CPU через audio.cpp быстрее реального времени.

Ключевые особенности

  • 92 африканских языка + английский; у каждого языка есть голос по умолчанию.
  • 110M параметров, 6-слойная Pocket TTS FlowLM; дистилляция из 24-слойной Africa TTS Base, 200k шагов.
  • CPU-движок audio.cpp: GGUF q8_0 на 144 MB, без PyTorch; первый поток аудио примерно через 0.1 с.
  • Клониров��ние голоса по 3–10-секундному клипу.
  • Форматы: safetensors fp32, GGUF q8_0 и int8-бандл sherpa-onnx для телефонов.
  • Обучена на 897 ч реальной речи; выход 24 kHz mono.

Пример

from africa_tts import AfricaTTS

tts = AfricaTTS.load()
audio = tts.generate('Akwaaba! Yɛfrɛ me Kofi.', lang='twi')
audio = tts.generate('Hello Ghana!', lang='eng', voice='my_voice.wav')
tts.save(audio, 'out.wav')
ПодробнееCER ниже, чем у MMS-TTS, в 21 из 27 общих языков: английский — 2.4 против 3.5, суахили — 1.2 против 2.6, киньяруанда — 3.2 против 6.3.
По умолчанию CPU-движок audio.cpp: GGUF q8_0 ≈144 MB, без PyTorch; первый поток аудио ≈0.1 с, загрузка модели ≈150 MB.
Обучена на 897 ч реальной речи; языки с менее ~2 ч данных слабые, часть G2P-таблиц не размечает тон. Лицензия CC-BY-NC-4.0, только некоммерческое использование; клонирование нельзя применять для выдачи себя за другого.
Discuss Telegram 0

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.4 GB · ready files0.9 GB
  • BF16weights 0.2 GB0.7 GB
  • GGUF Q8_0recommendedweights 0.1 GB · ready files0.6 GB
  • FP8 / INT8weights 0.1 GB0.6 GB
  • 4-bitweights 0.1 GB0.6 GB

Tip: go 8-bit.GGUF Q8_0 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: GGUF Q8_0 — from RTX 2060 · F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # AfriSpeech/africa-tts
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download AfriSpeech/africa-tts --local-dir ~/models/africa-tts
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to africa-tts
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of pocket-tts (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face

About the model

AfriSpeech/africa-tts
Author
AfriSpeech
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments