run_maskrcnn.py

import numpy as np
import cv2

import torch
import torchvision

in_size = 300

input_shape = (1, 3, in_size, in_size)


def download(url, path):
    import urllib.request as urllib2
    urllib2.urlretrieve(url, path)


def get_input():
    img_path = "test_street_small.jpg"
    img_url = (
        "https://raw.githubusercontent.com/dmlc/web-data/" "master/gluoncv/detection/street_small.jpg"
    )
    download(img_url, img_path)

    img = cv2.imread(img_path).astype("float32")
    img = cv2.resize(img, (in_size, in_size))
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    img = np.transpose(img / 255.0, [2, 0, 1])
    img = np.expand_dims(img, axis=0)
    return img


model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)
model.eval()

img = get_input()
inp = torch.from_numpy(img)

with torch.no_grad():
    out = model(inp)