main.py

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

# Get the directory of the script
script_directory = os.path.dirname(os.path.realpath(__file__))
# Specify the directory where the cache will be stored (same folder as the script)
cache_directory = os.path.join(script_directory, "cache")
# Create the cache directory if it doesn't exist
os.makedirs(cache_directory, exist_ok=True)

# Load the CLIP processor and model
clip_processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32", cache_dir=cache_directory)
clip_model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32", cache_dir=cache_directory)

# Text description to generate image
text = "a cat sitting on a table"

# Tokenize text and get features
inputs = clip_processor(text, return_tensors="pt", padding=True)

with torch.no_grad():
  image_features = clip_model.get_image_features(**inputs)

# Generate image from features
generated_image = clip_model.generate_images(image_features)

# Save the generated image
output_image_path = "generated_image.png"
Image.fromarray(generated_image).save(output_image_path)

print("Image generated and saved as:", output_image_path)