Gradio is a Python library for quickly turning machine learning functions into interactive web interfaces. It helps you test, showcase, and share models without extensive frontend development. section 17.0 section 17.1.1
A basic Gradio workflow is:
- Define a function that accepts the model input and returns its output. section 17.1.1
- Create a gr.Interface with the function and input and output components. section 17.1.1
- Launch the application and open the generated interface in a browser. section 17.1.1
- Use share=True to create a public link for quick demonstrations. section 17.1.1
Text interface
A function such as hello_world(name) can be connected to text input and output components. Gradio generates the web UI, and demo.launch() starts it locally at port 7860. section 17.1.1
Image-generation demo
Wrap a pretrained Stable Diffusion pipeline in a function that accepts a text prompt and returns an image, then connect it to gr.Textbox and gr.Image components. section 17.1.1
Listing 17.1 Creating a simple Gradio interface with text input and output
import gradio as gr
def hello_world(name):
return "Hello, " + name + "!"
demo = gr.Interface(
fn=hello_world,
inputs=["text"],
outputs=["text"],
flagging_mode="auto",
)
demo.launch()
Inspect a complete text-input and text-output interface built with gr.Interface. section 17.1.1
