app.py

import gradio as gr
from gradio_client import Client, file
from PIL import Image
import requests
from io import BytesIO

# Initialize the Hugging Face API clients
captioning_client = Client("fancyfeast/joy-caption-pre-alpha")
generation_client = Client("black-forest-labs/FLUX.1-dev")

# Function to caption an image
def caption_image(image):
    caption = captioning_client.predict(
        input_image=image,
        api_name="/stream_chat"
    )
    return caption

# Function to generate an image from a text prompt using Hugging Face API
def generate_image_from_caption(caption):
    image = generation_client.predict(
        prompt=caption,
        seed=0,
        randomize_seed=True,
        width=1024,
        height=1024,
        guidance_scale=3.5,
        num_inference_steps=28,
        api_name="/infer"
    )
    return image

# Main function to handle the upload and generate images and captions in a loop
def process_image(image, iterations):
    generated_images = []
    captions = []
    
    current_image = image
    
    for i in range(iterations):
        # Caption the current image
        caption = caption_image(current_image)
        captions.append(caption)
        
        # Generate a new image based on the caption
        new_image = generate_image_from_caption(caption)
        generated_images.append(new_image)
        
        # Set the newly generated image as the current image for the next iteration
        current_image = new_image
    
    return generated_images, captions

# Gradio Interface
with gr.Blocks() as demo:
    with gr.Row():
        image_input = gr.Image(type="pil", label="Upload an Image")
        iterations_input = gr.Number(value=3, label="Number of Iterations")
    
    with gr.Row():
        output_images = gr.Gallery(label="Generated Images")
        output_captions = gr.Textbox(label="Generated Captions")
    
    generate_button = gr.Button("Generate")
    
    generate_button.click(
        fn=process_image,
        inputs=[image_input, iterations_input],
        outputs=[output_images, output_captions]
    )

# Launch the app
demo.launch()