is a Python package that makes it very easy for developers to create machine learning apps that are powered by Anthropic's API.
- Clone this repo:
git clone https://github.com/AK391/anthropic-gradio.git
- Navigate into the folder that you cloned this repo into:
cd anthropic-gradio
- Install this package:
pip install -e .
That's it!
Just like if you were to use the anthropic
API, you should first save your Anthropic API key to this environment variable:
export ANTHROPIC_API_KEY=<your token>
Then in a Python file, write:
import gradio as gr
import anthropic_gradio
gr.load(
name='claude-3-opus-20240229',
src=anthropic_gradio.registry,
).launch()
Run the Python file, and you should see a Gradio Interface connected to the model on Anthropic!
Once you can create a Gradio UI from an Anthropic endpoint, you can customize it by setting your own input and output components, or any other arguments to gr.Interface
. For example:
import gradio as gr
import anthropic_gradio
gr.load(
name='claude-3-opus-20240229',
src=anthropic_gradio.registry,
title='Anthropic-Gradio Integration',
description="Chat with Claude 3 Opus model.",
examples=["Explain quantum gravity to a 5-year old.", "How many R are there in the word Strawberry?"]
).launch()
Or use your loaded Interface within larger Gradio Web UIs, e.g.
import gradio as gr
import anthropic_gradio
with gr.Blocks() as demo:
with gr.Tab("Claude 3 Opus"):
gr.load('claude-3-opus-20240229', src=anthropic_gradio.registry)
with gr.Tab("Claude 3 Sonnet"):
gr.load('claude-3-sonnet-20240229', src=anthropic_gradio.registry)
demo.launch()
The anthropic-gradio
Python library has two dependencies: anthropic
and gradio
. It defines a "registry" function anthropic_gradio.registry
, which takes in a model name and returns a Gradio app.
All chat API models supported by Anthropic are compatible with this integration. For a comprehensive list of available models and their specifications, please refer to the Anthropic Models documentation.
Note: if you are getting an authentication error, then the Anthropic API Client is not able to get the API token from the environment variable. This happened to me as well, in which case save it in your Python session, like this:
import os
os.environ["ANTHROPIC_API_KEY"] = ...