app.py

import gradio as gr
from urllib.request import urlopen
from PIL import Image
import timm
import torch

# Load the model
model = timm.create_model("hf_hub:Marqo/nsfw-image-detection-384", pretrained=True)
model = model.eval()

# Prepare the data transformation
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

# Prediction function
def predict(image):
    with torch.no_grad():
        # Transform the image
        input_tensor = transforms(image).unsqueeze(0)
        # Run the model
        output = model(input_tensor).softmax(dim=-1).cpu()
        # Get class names
        class_names = model.pretrained_cfg["label_names"]
        # Create the result dictionary
        result = {class_names[i]: float(output[0, i]) for i in range(len(class_names))}
    return result

# Gradio interface
interface = gr.Interface(
    fn=predict,
    inputs=gr.Image(type="pil"),
    outputs=gr.Label(num_top_classes=3),
    title="NSFW Image Detection",
    description="Upload an image to detect if it is NSFW or Safe for Work."
)

if __name__ == "__main__":
    interface.launch()