app.py

from typing import Union

from fastapi import FastAPI,File
from PIL import Image
from io import BytesIO
from transformers import DetrImageProcessor, DetrForObjectDetection
import torch
import requests

app = FastAPI(title="Object Detection",
    docs_url="/", 
    description="Object detection in Image")


processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")



@app.post('/image')
def read_image(image_file: bytes = File(...)):
    image = Image.open(BytesIO(image_file))
    # url = "http://images.cocodataset.org/val2017/000000039769.jpg"
    # image = Image.open(requests.get(url, stream=True).raw)

    inputs = processor(images=image, return_tensors="pt")
    print("image loaded")
    outputs = model(**inputs)
    target_sizes = torch.tensor([image.size[::-1]])
    results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]
    print("results pushed")
    response = {}
    
    for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
        box = [round(i, 2) for i in box.tolist()]
        # response['scores'] = model.config.id2label[label.item()]
        # response['labels'] = score.item()
        # response['boxes'] = box
        print(model.config.id2label[label.item()])
        print(score.item())
        print(box)
        # print(
        #         f"Detected {model.config.id2label[label.item()]} with confidence "
        #         f"{round(score.item(), 3)} at location {box}"
        # )
    
    return response