A browser interface based on the Gradio library for OpenAI's Whisper model.
- Transcribes videos from YouTube, a video or audio file or a recording from your microphone.
- Output
- Text
- Markdown with Timestamps
- JSON with the text, start and end time
- JSON with all parameters
- Downloadable as file
- The system default dark theme is currently not supported in all browsers.
- To switch to dark mode manually add
/?__theme=dark
to the end of the URL.
- To switch to dark mode manually add
You can run this on your CPU but it is a lot slower than running it on an GPU.
See this for the GPU hardware requirements.
The models tiny
, base
and small
should run on any hardware.
- Install Python 3, checking "Add Python to PATH"
- Install git
- Install ffmpeg
- Reboot
- Download this repository, for example by running
git clone ..........
. - (optional) To use your GPU install PyTorch manually (see Use a GPU)
- Install the required Python dependencies with
pip3 install -r requirements.txt
orpip install -r requirements.txt
- Copy the
.env.example
file and rename the copied file to.env
- (optional) Change the parameters in the
.env
file - Run
py app.py
- Install dependencies:
# Debian-based: sudo apt update && sudo apt install git python3 python3-venv python3-pip ffmpeg # Red Hat-based: sudo dnf install git python3 python3-pip ffmpeg # Arch-based: sudo pacman -S git python3 python-pip ffmpeg
- Download this repository, for example by running
git clone ..........
. - (optional) To use your GPU install PyTorch manually (see Use a GPU)
- Install the required Python dependencies with
pip3 install -r requirements.txt
orpip install -r requirements.txt
- Copy the
.env.example
file and rename the copied file to.env
- (optional) Change the parameters in the
.env
file - Run
python3 app.py
orpython app.py
To use it with a GPU, install Nvidia's CUDA or AMD's ROCm drivers.
To install PyTorch with the right driver version see here: https://pytorch.org/get-started/locally/
Example: To install PyTorch for CUDA version 11.6: pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116
- OpenAI Whisper https://openai.com/blog/whisper/, https://github.com/openai/whisper