Run inference using TensorRT using .engine file of .trt file

I have no idea how do will run inference on single image without DS. I have use code mentioned in TensorRT seminar and i want that type output. code sample is below

import numpy as np
import imageio
import tensorrt as trt
import matplotlib.pyplot as plt
import utils

ENGINE_PATH = ‘resnet10.trt’
CLASSES = [‘Bike’]

CROP_SIZE = (224,224)
DATA_TYPE = trt.infer.DataType.FLOAT #or HALF

engine = trt.lite.Engine(PLAN = ENGINE_PATH, data_type = DATA_TYPE)

INPUT_IMAGE_PATH = ‘sample_720p.jpg’

def prepare_image(image_in,crop_size,data_type):
img = utils.resize_and_crop(image_in,crop_size)
if data_type == trt.infer.DataType.HALF:
img = img.astype(np.float16)
elif data_type == trt.infer.DataType.FLOAT:
img = img.astype(np.float32)
img = img.transpose(2,0,1) #to CHW

img = imageio.imread(INPUT_IMAGE_PATH, pilmode=‘RGB’)
img = prepare_image(img,CROP_SIZE,DATA_TYPE)


out = engine.infer(img)

print(‘Prediction : {}’.format(CLASSES[np.argmax(out[0])]))

please help me out.
if there is any sample code by which i will able to get prediction result on resnet10.caffemodel_b1_int8.engine and resnet18_detector.trt like detected class their b-box and precision for detector, and detected class for classifier.