app.py

import gradio as gr
from PIL import Image
import torch
from transformers import CLIPProcessor, CLIPModel

# Load model and processor
model = CLIPModel.from_pretrained("laion/CLIP-ViT-H-14-laion2B-s32B-b79K")
processor = CLIPProcessor.from_pretrained("laion/CLIP-ViT-H-14-laion2B-s32B-b79K")

def calculate_similarity(image, text_prompt):
    # Process inputs
    inputs = processor(text=text_prompt, images=image, return_tensors="pt", padding=True)

    # Forward pass
    outputs = model(**inputs)

    # Normalize and calculate cosine similarity
    image_features = outputs.image_embeds / outputs.image_embeds.norm(dim=-1, keepdim=True)
    text_features = outputs.text_embeds / outputs.text_embeds.norm(dim=-1, keepdim=True)
    cosine_similarity = torch.nn.functional.cosine_similarity(image_features, text_features)

    return {"Cosine Similarity": cosine_similarity.item()}

# Set up Gradio interface
iface = gr.Interface(fn=calculate_similarity,
                     inputs=[gr.inputs.Image(type="pil"), gr.inputs.Textbox(label="Text Prompt")],
                     outputs=[gr.outputs.Label(label="Cosine Similarity")],
                     title="OpenClip Cosine Similarity Calculator",
                     description="Upload an image and provide a text prompt to calculate the cosine similarity.")

# Launch the interface locally for testing
iface.launch()