import torch
from torchvision.models.detection import (
    fasterrcnn_resnet50_fpn,
    FasterRCNN_ResNet50_FPN_Weights
)
from PIL import Image
import torchvision.transforms as T

# Load pretrained model
weights = FasterRCNN_ResNet50_FPN_Weights.DEFAULT
model = fasterrcnn_resnet50_fpn(weights=weights)

model.eval()

# Load image
image = Image.open("image.jpg").convert("RGB")

# Convert image to tensor
transform = T.ToTensor()
image_tensor = transform(image)

# Run detection
with torch.no_grad():
    predictions = model([image_tensor])

prediction = predictions[0]

# COCO class names
categories = weights.meta["categories"]

# Print detected objects
for box, label, score in zip(
    prediction["boxes"],
    prediction["labels"],
    prediction["scores"]
):
    if score > 0.5:
        print(
            f"{categories[label]}: "
            f"{score:.2f}, "
            f"box={box.tolist()}"
        )