app.py

import os
from huggingface_hub import login
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import pytesseract



# Get Hugging Face Token from environment variable
hf_token = os.getenv('HF_AUTH_TOKEN')
if not hf_token:
    raise ValueError("Hugging Face token is not set in the environment variables.")
login(token=hf_token)

# Load the processor and model
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
import gradio as gr
from diffusers import DiffusionPipeline
import torch
import spaces  # Hugging Face Spaces module

# Initialize the model
pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-3.5-medium")
device = "cuda" if torch.cuda.is_available() else "cpu"
pipe.to(device)
model.to(device)



@spaces.GPU(duration=300)
def generate_caption_and_image(image):
    img = image.convert("RGB")

    
    extracted_text = pytesseract.image_to_string(img)
    # Process the image
    raw_image = image.convert("RGB")
      

    
    # Generate caption
    inputs = processor(raw_image, return_tensors="pt", padding=True, truncation=True, max_length=250)
    inputs = {key: val.to(device) for key, val in inputs.items()}
    out = model.generate(**inputs)
    caption = processor.decode(out[0], skip_special_tokens=True)
    prompt = f"Create a highly realistic design of a clothing item based on the following description: 'The design should incorporate elements from the extracted text: {extracted_text}, along with the caption: {caption}. The clothing should look realistic, modern, and stylish. Use high-quality fabric textures and realistic lighting to give the design a lifelike appearance. The colors, patterns, and materials should reflect the essence of the caption and extracted text.'"


    # Generate image based on the caption
    generated_image = pipe(prompt).images[0]

    return caption, generated_image

# Gradio UI
iface = gr.Interface(
    fn=generate_caption_and_image,
    inputs=gr.Image(type="pil", label="Upload Image"),
    outputs=[gr.Textbox(label="Generated Caption"), gr.Image(label="Generated Design")],
    live=True
)
iface.launch(share=True)