jim60105/docker-whisperX

Dockerfile for WhisperX: Automatic Speech Recognition with Word-Level Timestamps and Speaker Diarization (Dockerfile, CI image build and test)

What it solves

This project provides a set of optimized Docker images for WhisperX, an automatic speech recognition (ASR) system that provides word-level timestamps and speaker diarization. It simplifies the deployment of WhisperX by bundling pre-downloaded models, reducing the need for users to manually manage dependencies and model downloads during runtime.

How it works

The project uses a Dockerfile and a CI workflow to build and maintain a large matrix of pre-built images. Each image is tagged by the model size (e.g., large-v3) and the target language (e.g., en for English, zh for Chinese). Users can either pull a pre-built image from the GitHub Container Registry (GHCR) or build their own using build arguments for specific language and model combinations.

Who it’s for

It is designed for users who want to run WhisperX for speech-to-text transcription with GPU support on Windows (via WSL2), Linux, or OSX, without having to manually configure the environment or install the underlying ASR models.

Highlights

  • Pre-built Model Bundles: Offers a wide variety of images with models already baked in to speed up deployment.
  • GPU Support: Detailed configuration guides for NVIDIA GPUs across different operating systems.
  • Optimized Build Pipeline: Uses efficient Docker layer caching to manage the build of over 170 images weekly.
  • Specialized Models: Includes support for distilled variants (like distil-large-v3-en) and specialized regional models (like breeze-asr-26-zh for Taiwanese Hokkien).
  • Flexible Caching: Allows mounting a local volume to share alignment models across different containers.

Related

  • Project
    ahmetoner/whisper-asr-webservice

    A general-purpose speech recognition toolkit that wraps Whisper models into a REST API for easy deployment and integration.

  • Project
    m-bain/whisperX

    A fast automatic speech recognition system that enhances OpenAI's Whisper with batched inference, word-level timestamps, and speaker diarization.

  • Project
    eleiton/ollama-intel-arc

    A Docker-based deployment suite that enables Intel Arc GPU acceleration for Ollama, Stable Diffusion, and OpenAI Whisper via a unified interface.

  • Project
    collabora/WhisperLive

    WhisperLive is an open‑source, near‑real‑time transcription server built around OpenAI’s Whisper model. It supports multiple inference back‑ends (faster‑whisper, NVIDIA TensorRT, Intel OpenVINO, AMD ROCm), offers word‑level timestamps, hot‑word boosting, speaker diarization and optional translation, and provides both simple command‑line clients and a Python streaming API. Docker images are provided for GPU and CPU deployments, and browser/iOS extensions let you transcribe audio directly from a web page or mobile device.

  • Project
    Quantatirsk/qwen3-asr

    Qwen3‑ASR is a self‑hosted speech‑to‑text server that wraps Qwen3‑ASR models (0.6 B & 1.7 B). It auto‑selects a GPU‑vLLM backend or a CPU‑Rust backend, offers OpenAI‑ and Alibaba‑compatible HTTP/WebSocket APIs, supports speaker diarization, VAD‑based segmentation, batch inference, and can be deployed via Docker, custom GPU images, or offline tarballs. MIT‑licensed.