CLIP kodu

clip_test.py

import torch
import clip
from PIL import Image
import gradio as gr

if __name__ == "__main__":
    device = "cuda" if torch.cuda.is_available() else "cpu"
    model, preprocess = clip.load("ViT-B/32", device=device)


    items = ["image of fire", "forest image", "image of smoke","image of sea","image of clouds"]

    def infer(image):
        out_labels = {}
        with torch.no_grad():
            image = preprocess(Image.fromarray(image)).unsqueeze(0).to(device)

            text = clip.tokenize(items).to(device)


            logits_per_image, logits_per_text = model(image, text)
            probs = logits_per_image.softmax(dim=-1).cpu().numpy()
            for item,prob in zip(items,probs[0].squeeze()):
                out_labels[item] = float(prob)

        return out_labels


    iface = gr.Interface(fn=infer, inputs="image", outputs="label")

    iface.launch()