Skip to content
Audio

Omnilingual ASR 1B CTC (int8 ONNX) - Speech Recognition for 1,600+ Languages

1 min readupdated

The Omnilingual ASR 1B CTC model from Meta, optimized and exported to the int8 ONNX format (a format for running models quickly across various devices) using the sherpa-onnx project. It is designed for automatic speech recognition across more than 1,600 languages. The repository includes ready-to-use weights and a tokenizer for developers. Ideal for building multilingual audio transcription services and generating subtitles.

Key features

  • Supports speech recognition for over 1,600 languages worldwide.
  • Provided in the int8 ONNX format for efficient inference (the model's operation process).
  • Built on the Meta Omnilingual ASR architecture (1 billion parameters).
  • Distributed under the open Apache 2.0 license.

What you can do

  • Integrate the model into applications for automatic audio recording transcription.
  • Set up offline voice data processing using the sherpa-onnx ecosystem.
  • Use it to analyze audio tracks and generate subtitles in a multilingual environment.

Example

To run speech recognition, the model is used together with the sherpa-onnx library in accordance with the project documentation.

Read moreDetails
Base model: Omnilingual ASR 1B CTC (Meta)
Format: int8 ONNX (k2-fsa / sherpa-onnx export)
Model file size: ~1.03 GB
License: Apache 2.0
Discuss Telegram 0

Will it run?

Runs in INT8

Fits without quantisation.

  • INT8recommendedweights 1.0 GB · ready files1.6 GB
  • 4-bitweights 0.5 GB1.1 GB

Tip: go 8-bit.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: INT8 — 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.

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

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

About the model

rostrumstudios/speech-languages
Author
rostrumstudios
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