app.py

import gradio as gr
from PIL import Image


def infer_segmentation(prompt, negative_prompt, image):
    # implement your inference function here
    im = Image.open("cat_image.jpeg")
    return im

def infer_canny(prompt, negative_prompt, image):
    # implement your inference function here
    im = Image.open("cat_image.jpeg")
    return im

with gr.Blocks() as demo:
    gr.Markdown("## Stable Diffusion with Different Controls")
    gr.Markdown("In this app, you can find different ControlNets with different filters. ")


    with gr.Tab("ControlNet on Canny Filter "):
        prompt_input_canny = gr.Textbox(label="Prompt")
        negative_prompt_canny = gr.Textbox(label="Negative Prompt")
        canny_input = gr.Image(label="Image")
        submit_btn = gr.Button("Submit")
        canny_inputs = [prompt_input_canny, negative_prompt_canny, canny_input]
        submit_btn.click(fn=write_to_dataset, inputs=canny_inputs, outputs="image")
    with gr.Tab("ControlNet with Semantic Segmentation"):
        prompt_input_seg = gr.Textbox(label="Prompt")
        negative_prompt_seg = gr.Textbox(label="Negative Prompt")
        seg_input = gr.Image(label="Image")
        submit_btn = gr.Button("Submit")
        seg_inputs = [prompt_input_seg, negative_prompt_seg, seg_input]
        submit_btn.click(fn=write_to_dataset, inputs=seg_inputs, outputs="image")

demo.launch()