app.py

import gradio as gr
from ultralyticsplus import YOLO, render_result
import cv2
import time

# Load model with automatic device detection
model = YOLO('foduucom/plant-leaf-detection-and-classification')

# Optimize model configuration
model.overrides.update({
    'conf': 0.25,
    'iou': 0.45,
    'imgsz': 640,
    'device': '0' if model.device.type != 'cpu' else 'cpu'
})

def detect_leaves(image):
    start_time = time.time()
    
    # Convert image format
    img = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
    
    # Predict with optimized settings
    results = model.predict(
        source=img,
        verbose=False,  # Disable unnecessary logging
        stream=False    # Disable streaming mode
    )
    
    # Process results
    num_leaves = len(results[0].boxes)
    rendered_img = render_result(model=model, image=img, result=results[0])
    
    print(f"Total processing time: {time.time() - start_time:.2f}s")
    return cv2.cvtColor(rendered_img, cv2.COLOR_BGR2RGB), num_leaves

# Create lightweight interface
interface = gr.Interface(
    fn=detect_leaves,
    inputs=gr.Image(label="Plant Image"),
    outputs=[
        gr.Image(label="Detection Result", width=600),
        gr.Number(label="Leaves Count")
    ],
    title="🍃 Leaf Detection",
    allow_flagging="never"
)

if __name__ == "__main__":
    interface.launch(
        server_port=7860,
        show_error=True,
        enable_queue=True
    )