Skip to content
Audio

tiny-voca-ar

a compact speech synthesis model for Classical Arabic

2 min read

tiny-voca-ar is a port of the lightweight tiny-voca architecture adapted for Modern Standard Arabic. Trained from scratch on synthetic data without pre-trained weights, it packs roughly 4.5 million parameters and outputs mono audio at 24 kHz. The model is built for offline, local, or embedded deployments where hardware resources are tight.

Key features

  • Lightweight design: Just 4.5 million parameters for fast performance on low-end hardware.
  • Diacritic-driven: Works exclusively with fully diacritized (tashkeel) text.
  • Clean studio voice: Tuned for a male studio voice in the M2 slot.
  • Open license: Released under Apache-2.0 with CC BY 4.0 attribution.

What you can do

  • Narrate texts: Generate offline audio tracks for Classical Arabic texts.
  • Power bots: Add compact voice synthesis to local services and applications.
  • Export to ONNX: Run cross-platform inference using the ready-to-use ONNX export.

Example

Generate speech in Python using the tiny-voca package:

import soundfile as sf
from tiny_voca.infer import load_model, speak_text

model = load_model("tiny-voca-ar/tiny-voca-ar.pt")
wave, _ = speak_text(model, "اِتَّصِلْ بِي عَلَى هَاتِفِي 0501234567 فِي أَيِّ وَقْتٍ.", voice="M2")
sf.write("phone.wav", wave.numpy(), 24000)
Read moreDetails
• Architecture and training: trained for 79,955 steps on data from ClArTTS and the Arabic Speech Corpus.
• Metrics: achieves a Word Error Rate (WER) of 0.0532 evaluated with Whisper large-v3-turbo.
• Limitations: supports only a single male voice (the M2 slot; F2 is untrained), and input text must include diacritics.
Discuss Telegram 0 17

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 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: FP8 / INT8 — from RTX 2060 · BF16 — 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.

    # abdullah/tiny-voca-ar
    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 abdullah/tiny-voca-ar --local-dir ~/models/tiny-voca-ar
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to tiny-voca-ar
  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 tiny-voca-ar (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

abdullah/tiny-voca-ar
Author
abdullah
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

170

Ask AI about this model

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

Comments

Audio

SAVA

SAVA