Skip to content

Command Line Arguments

jhj0517 edited this page Aug 24, 2024 · 18 revisions

How to use

You can run project with some command line arguments by running user-start-webui.bat file.

You can set the values in there and run it.

Or you can just run the project directly on command line with lower case arguments, for example:

python app.py --whisper_type insanely_fast_whisper --share --server_name 0.0.0.0 --theme author/example_theme

Arguments

Most arguments are related to Gradio's launch options. See the Gradio documentation for a detailed explanation.

Variable Description Example Value
--server_name Server name to set in Gradio. 0.0.0.0
--server_port Server port number to use when launching Gradio. 7861
--username Server username for Gradio. Both USERNAME and PASSWORD must be set simultaneously to enable authentication. myusername
--password Server password for Gradio. Both USERNAME and PASSWORD must be set simultaneously to enable authentication. mypassword
--share Enables the creation of a public link for the project. True
--root_path Gradio root path of the URL. If the application is served at "https://example.com/myapp", the root_path should be set to "/myapp". /myapp
--theme UI theme to apply in the browser. View available themes here: Theme Gallery author/example_theme
--whisper_type A type of the implementation of the whisper to use.
whisper : Normal whisper implementation by OpenAI
faster-whisper : Optimized for transcription speed & VRAM usage for whisper by SYSTRAN
insanely-fast-whisper : Optimized for transcription speed using some transformers options. Fastest implementation for now.

This project uses faster-whisper by default.
Enter the repository name as the value.
whisper or
faster-whisper or
insanely-fast-whisper
--faster_whisper_model_dir Use this if you want to use custom model directory path for faster-whisper. Make sure you enter the path "inside quotes". "C:\Whisper-WebUI\my_faster_whisper_model_path"
--insanely_fast_whisper_model_dir Use this if you want to use custom model directory path for insanely_fast_whisper. Make sure you enter the path "inside quotes". "C:\Whisper-WebUI\my_insanely_fast_whisper_model_path"
--whisper_model_dir Use this if you want to use custom model directory path for whisper. Make sure you enter the path "inside quotes". "C:\Whisper-WebUI\my_whisper_model_path"
--diarization_model_dir Use this if you want to use custom model directory path for speaker diarization models. Make sure you enter the path "inside quotes".
To download the model, you must manually go to https://huggingface.co/pyannote/speaker-diarization-3.1 and agree to their requirement.
"C:\Whisper-WebUI\model_dir"
--output_dir Use this if you want custom outputs directory "C:\Whisper-WebUI\outputs"
--colab Whether the envrionment is Google Colab or not. If it's true, "Input Folder Path (Optional)" tab will appear and you can enter the folder path instead of uploading via gradio.Files(). True
Clone this wiki locally