concept_1_multimodal.py

import torch
from PIL import Image
from transformers import CLIPProcessor, CLIPModel

def classify_image(image_path, text_labels):
    """
    Classifies an image based on a list of text labels using a CLIP model.
    """
    model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
    processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

    image = Image.open(image_path)
    inputs = processor(text=text_labels, images=image, return_tensors="pt", padding=True)

    with torch.no_grad():
        outputs = model(**inputs)
    
    logits_per_image = outputs.logits_per_image
    probs = logits_per_image.softmax(dim=1)
    
    return dict(zip(text_labels, probs.tolist()[0]))

if __name__ == '__main__':
    # Create a dummy image for testing
    dummy_image = Image.new('RGB', (100, 100), color = 'red')
    dummy_image.save("dummy_image.png")

    labels = ["a red square", "a blue circle", "a green triangle"]
    probabilities = classify_image("dummy_image.png", labels)
    print("Probabilities:", probabilities)
    print("Predicted label:", max(probabilities, key=probabilities.get))