Skip to content

Latest commit

 

History

History

gradio

README.md

Gradio

Gradio is an open-source Python library that enables creating customizable and shareable user interfaces for machine learning models, making it easier to demo your applications with minimal coding effort. It offers features like text inputs, image uploads, data tables, and real-time feedback outputs, allowing users to interactively explore ML models in their web browsers.

This configuration includes the following prebuilt apps:

Config

make config

Choose a host name for the gradio instance:

GRADIO_TRAEFIK_HOST: Enter the gradio domain name (eg. gradio.example.com)
: gradio.example.com

Choose the app to install:

? Choose the gradio app to install
  chatbot
> hello

Enable sentry authorization:

? Do you want to enable sentry authorization in front of this app (effectively making the entire site private)?  
  No
> Yes, with HTTP Basic Authentication
  Yes, with Oauth2
  Yes, with Mutual TLS (mTLS)
  
Enter the username for HTTP Basic Authentication
: private
Enter the passphrase for private (leave blank to generate a random passphrase)
: hunter2

> Would you like to create additional usernames (for the same access privilege)? No

> Would you like to export the usernames and cleartext passwords to the file passwords.json? No

Install

make install wait

Open

make open

or open your web browser to the GRADIO_TRAEFIK_HOST you set, and enter the username and password when prompted.

Chatbot app

If you choose the chatbot app in the config:

? Choose the gradio app to install  
> chatbot
  hello

You must set the remote URL for your LM Studio instance.

  • Download LM Studio on a machine that has a GPU.
    • The Docker server must have a network route to the LM Studio machine.
  • In LM studio, go to the Developer tab.
  • Enter a server port: 1234.
  • Enable Serve on Local Network.
  • Enable Verbose logging.
  • Click Start.

Enter the URL in the config, using the LAN IP address of the machine running LM Studio:

http://192.168.1.2:1234/v1/chat/completions

This will let Gradio use your LM Studio instance as its backend agent.

Development

You can create new applications in the gradio sub-directory. To enable automatic reloading on source file changes, run:

make dev

This relies upon having a Rust + Cargo enviornment, and it will automatically install cargo-watch