diff --git a/yolov5/CMakeLists.txt b/yolov5/CMakeLists.txt
index cc95562..3cee96f 100644
--- a/yolov5/CMakeLists.txt
+++ b/yolov5/CMakeLists.txt
@@ -37,6 +37,13 @@ target_link_libraries(yolov5 cudart)
 target_link_libraries(yolov5 myplugins)
 target_link_libraries(yolov5 ${OpenCV_LIBS})
 
+add_executable(yolov5-cls calibrator.cpp yolov5_cls.cpp)
+
+target_link_libraries(yolov5-cls nvinfer)
+target_link_libraries(yolov5-cls cudart)
+target_link_libraries(yolov5-cls myplugins)
+target_link_libraries(yolov5-cls ${OpenCV_LIBS})
+
 if(UNIX)
 add_definitions(-O2 -pthread)
 endif(UNIX)
diff --git a/yolov5/gen_wts.py b/yolov5/gen_wts.py
index 5ae9497..c20fa50 100644
--- a/yolov5/gen_wts.py
+++ b/yolov5/gen_wts.py
@@ -8,8 +8,12 @@ from utils.torch_utils import select_device
 
 def parse_args():
     parser = argparse.ArgumentParser(description='Convert .pt file to .wts')
-    parser.add_argument('-w', '--weights', required=True, help='Input weights (.pt) file path (required)')
-    parser.add_argument('-o', '--output', help='Output (.wts) file path (optional)')
+    parser.add_argument('-w', '--weights', required=True,
+                        help='Input weights (.pt) file path (required)')
+    parser.add_argument(
+        '-o', '--output', help='Output (.wts) file path (optional)')
+    parser.add_argument(
+        '-t', '--type', type=str, default='', help='determines the model is detection/classification')
     args = parser.parse_args()
     if not os.path.isfile(args.weights):
         raise SystemExit('Invalid input file')
@@ -19,10 +23,10 @@ def parse_args():
         args.output = os.path.join(
             args.output,
             os.path.splitext(os.path.basename(args.weights))[0] + '.wts')
-    return args.weights, args.output
+    return args.weights, args.output, args.type
 
 
-pt_file, wts_file = parse_args()
+pt_file, wts_file, m_type = parse_args()
 
 # Initialize
 device = select_device('cpu')
@@ -30,11 +34,14 @@ device = select_device('cpu')
 model = torch.load(pt_file, map_location=device)  # load to FP32
 model = model['ema' if model.get('ema') else 'model'].float()
 
-# update anchor_grid info
-anchor_grid = model.model[-1].anchors * model.model[-1].stride[...,None,None]
-# model.model[-1].anchor_grid = anchor_grid
-delattr(model.model[-1], 'anchor_grid')  # model.model[-1] is detect layer
-model.model[-1].register_buffer("anchor_grid",anchor_grid) #The parameters are saved in the OrderDict through the "register_buffer" method, and then saved to the weight.
+if m_type == "detect":
+    # update anchor_grid info
+    anchor_grid = model.model[-1].anchors * \
+        model.model[-1].stride[..., None, None]
+    # model.model[-1].anchor_grid = anchor_grid
+    delattr(model.model[-1], 'anchor_grid')  # model.model[-1] is detect layer
+    # The parameters are saved in the OrderDict through the "register_buffer" method, and then saved to the weight.
+    model.model[-1].register_buffer("anchor_grid", anchor_grid)
 
 model.to(device).eval()
 
@@ -45,5 +52,5 @@ with open(wts_file, 'w') as f:
         f.write('{} {} '.format(k, len(vr)))
         for vv in vr:
             f.write(' ')
-            f.write(struct.pack('>f' ,float(vv)).hex())
+            f.write(struct.pack('>f', float(vv)).hex())
         f.write('\n')
diff --git a/yolov5/yolov5_cls.cpp b/yolov5/yolov5_cls.cpp
new file mode 100644
index 0000000..2949b88
--- /dev/null
+++ b/yolov5/yolov5_cls.cpp
@@ -0,0 +1,296 @@
+#include <iostream>
+#include <chrono>
+#include <cmath>
+#include "cuda_utils.h"
+#include "logging.h"
+#include "common.hpp"
+#include "utils.h"
+#include "calibrator.h"
+
+#define USE_FP32  // set USE_INT8 or USE_FP16 or USE_FP32
+#define DEVICE 0  // GPU id
+#define NMS_THRESH 0.4
+#define CONF_THRESH 0.5
+#define BATCH_SIZE 1
+#define MAX_IMAGE_INPUT_SIZE_THRESH 3000 * 3000 // ensure it exceed the maximum size in the input images !
+
+// stuff we know about the network and the input/output blobs
+static const int INPUT_H = 224;
+static const int INPUT_W = 224;
+static const int CLASS_NUM = 1000;
+
+static const int OUTPUT_SIZE = Yolo::MAX_OUTPUT_BBOX_COUNT * sizeof(Yolo::Detection) / sizeof(float) + 1;  // we assume the yololayer outputs no more than MAX_OUTPUT_BBOX_COUNT boxes that conf >= 0.1
+const char* INPUT_BLOB_NAME = "data";
+const char* OUTPUT_BLOB_NAME = "prob";
+static Logger gLogger;
+
+static int get_width(int x, float gw, int divisor = 8) {
+    return int(ceil((x * gw) / divisor)) * divisor;
+}
+
+static int get_depth(int x, float gd) {
+    if (x == 1) return 1;
+    int r = round(x * gd);
+    if (x * gd - int(x * gd) == 0.5 && (int(x * gd) % 2) == 0) {
+        --r;
+    }
+    return std::max<int>(r, 1);
+}
+
+ICudaEngine* build_engine(unsigned int maxBatchSize, IBuilder* builder, IBuilderConfig* config, DataType dt, float& gd, float& gw, std::string& wts_name) {
+    INetworkDefinition* network = builder->createNetworkV2(0U);
+
+    // Create input tensor of shape {3, INPUT_H, INPUT_W} with name INPUT_BLOB_NAME
+    ITensor* data = network->addInput(INPUT_BLOB_NAME, dt, Dims3{ 3, INPUT_H, INPUT_W });
+    assert(data);
+    std::map<std::string, Weights> weightMap = loadWeights(wts_name);
+    /* ------ yolov5 backbone------ */
+    auto conv0 = convBlock(network, weightMap, *data,  get_width(64, gw), 6, 2, 1,  "model.0");
+    assert(conv0);
+    auto conv1 = convBlock(network, weightMap, *conv0->getOutput(0), get_width(128, gw), 3, 2, 1, "model.1");
+    auto bottleneck_CSP2 = C3(network, weightMap, *conv1->getOutput(0), get_width(128, gw), get_width(128, gw), get_depth(3, gd), true, 1, 0.5, "model.2");
+    auto conv3 = convBlock(network, weightMap, *bottleneck_CSP2->getOutput(0), get_width(256, gw), 3, 2, 1, "model.3");
+    auto bottleneck_csp4 = C3(network, weightMap, *conv3->getOutput(0), get_width(256, gw), get_width(256, gw), get_depth(6, gd), true, 1, 0.5, "model.4");
+    auto conv5 = convBlock(network, weightMap, *bottleneck_csp4->getOutput(0), get_width(512, gw), 3, 2, 1, "model.5");
+    auto bottleneck_csp6 = C3(network, weightMap, *conv5->getOutput(0), get_width(512, gw), get_width(512, gw), get_depth(9, gd), true, 1, 0.5, "model.6");
+    auto conv7 = convBlock(network, weightMap, *bottleneck_csp6->getOutput(0), get_width(1024, gw), 3, 2, 1, "model.7");
+    auto bottleneck_csp8 = C3(network, weightMap, *conv7->getOutput(0), get_width(1024, gw), get_width(1024, gw), get_depth(3, gd), true, 1, 0.5, "model.8");
+
+    /* ------ yolov5 classification head ------ */
+    auto conv_class = convBlock(network, weightMap, *bottleneck_csp8->getOutput(0), 1280, 1, 1, 1, "model.9.conv");
+    IPoolingLayer* pool2 = network->addPoolingNd(*conv_class->getOutput(0), PoolingType::kAVERAGE, DimsHW{7, 7});
+    assert(pool2);
+    IFullyConnectedLayer* yolo = network->addFullyConnected(*pool2->getOutput(0), CLASS_NUM, weightMap["model.9.linear.weight"], weightMap["model.9.linear.bias"]);
+    assert(yolo);
+
+    yolo->getOutput(0)->setName(OUTPUT_BLOB_NAME);
+    network->markOutput(*yolo->getOutput(0));
+    // Build engine
+    builder->setMaxBatchSize(maxBatchSize);
+    config->setMaxWorkspaceSize(16 * (1 << 20));  // 16MB
+
+#if defined(USE_FP16)
+    config->setFlag(BuilderFlag::kFP16);
+#elif defined(USE_INT8)
+    std::cout << "Your platform support int8: " << (builder->platformHasFastInt8() ? "true" : "false") << std::endl;
+    assert(builder->platformHasFastInt8());
+    config->setFlag(BuilderFlag::kINT8);
+    Int8EntropyCalibrator2* calibrator = new Int8EntropyCalibrator2(1, INPUT_W, INPUT_H, "./coco_calib/", "int8calib.table", INPUT_BLOB_NAME);
+    config->setInt8Calibrator(calibrator);
+#endif
+
+    std::cout << "Building engine, please wait for a while..." << std::endl;
+    ICudaEngine* engine = builder->buildEngineWithConfig(*network, *config);
+    std::cout << "Build engine successfully!" << std::endl;
+
+    // Don't need the network any more
+    network->destroy();
+
+    // Release host memory
+    for (auto& mem : weightMap)
+    {
+        free((void*)(mem.second.values));
+    }
+
+    return engine;
+}
+
+
+void APIToModel(unsigned int maxBatchSize, IHostMemory** modelStream, float& gd, float& gw, std::string& wts_name) {
+    // Create builder
+    IBuilder* builder = createInferBuilder(gLogger);
+    IBuilderConfig* config = builder->createBuilderConfig();
+
+    // Create model to populate the network, then set the outputs and create an engine
+    ICudaEngine *engine = nullptr;
+
+    engine = build_engine(maxBatchSize, builder, config, DataType::kFLOAT, gd, gw, wts_name);
+    
+    assert(engine != nullptr);
+
+    // Serialize the engine
+    (*modelStream) = engine->serialize();
+
+    // Close everything down
+    engine->destroy();
+    builder->destroy();
+    config->destroy();
+}
+
+void doInference(IExecutionContext& context, cudaStream_t& stream, void **buffers, float* output, int batchSize) {
+    // infer on the batch asynchronously, and DMA output back to host
+    context.enqueue(batchSize, buffers, stream, nullptr);
+    CUDA_CHECK(cudaMemcpyAsync(output, buffers[1], batchSize * OUTPUT_SIZE * sizeof(float), cudaMemcpyDeviceToHost, stream));
+    cudaStreamSynchronize(stream);
+}
+
+bool parse_args(int argc, char** argv, std::string& wts, std::string& engine, float& gd, float& gw, std::string& img_dir) {
+    if (argc < 4) return false;
+    if (std::string(argv[1]) == "-s" && (argc == 5 || argc == 7)) {
+        wts = std::string(argv[2]);
+        engine = std::string(argv[3]);
+        auto net = std::string(argv[4]);
+        if (net[0] == 'n') {
+            gd = 0.33;
+            gw = 0.25;
+        } else if (net[0] == 's') {
+            gd = 0.33;
+            gw = 0.50;
+        } else if (net[0] == 'm') {
+            gd = 0.67;
+            gw = 0.75;
+        } else if (net[0] == 'l') {
+            gd = 1.0;
+            gw = 1.0;
+        } else if (net[0] == 'x') {
+            gd = 1.33;
+            gw = 1.25;
+        } else if (net[0] == 'c' && argc == 7) {
+            gd = atof(argv[5]);
+            gw = atof(argv[6]);
+        } else {
+            return false;
+        }
+    } else if (std::string(argv[1]) == "-d" && argc == 4) {
+        engine = std::string(argv[2]);
+        img_dir = std::string(argv[3]);
+    } else {
+        return false;
+    }
+    return true;
+}
+
+int main(int argc, char** argv) {
+    cudaSetDevice(DEVICE);
+
+    std::string wts_name = "";
+    std::string engine_name = "";
+    float gd = 0.0f, gw = 0.0f;
+    std::string img_dir;
+    if (!parse_args(argc, argv, wts_name, engine_name, gd, gw, img_dir)) {
+        std::cerr << "arguments not right!" << std::endl;
+        std::cerr << "./yolov5 -s [.wts] [.engine] [n/s/m/l/x or c gd gw]  // serialize model to plan file" << std::endl;
+        std::cerr << "./yolov5 -d [.engine] ../samples  // deserialize plan file and run inference" << std::endl;
+        return -1;
+    }
+
+    // create a model using the API directly and serialize it to a stream
+    if (!wts_name.empty()) {
+        IHostMemory* modelStream{ nullptr };
+        APIToModel(BATCH_SIZE, &modelStream, gd, gw, wts_name);
+        assert(modelStream != nullptr);
+        std::ofstream p(engine_name, std::ios::binary);
+        if (!p) {
+            std::cerr << "could not open plan output file" << std::endl;
+            return -1;
+        }
+        p.write(reinterpret_cast<const char*>(modelStream->data()), modelStream->size());
+        modelStream->destroy();
+        return 0;
+    }
+
+    // deserialize the .engine and run inference
+    std::ifstream file(engine_name, std::ios::binary);
+    if (!file.good()) {
+        std::cerr << "read " << engine_name << " error!" << std::endl;
+        return -1;
+    }
+    char *trtModelStream = nullptr;
+    size_t size = 0;
+    file.seekg(0, file.end);
+    size = file.tellg();
+    file.seekg(0, file.beg);
+    trtModelStream = new char[size];
+    assert(trtModelStream);
+    file.read(trtModelStream, size);
+    file.close();
+
+    std::vector<std::string> file_names;
+    if (read_files_in_dir(img_dir.c_str(), file_names) < 0) {
+        std::cerr << "read_files_in_dir failed." << std::endl;
+        return -1;
+    }
+
+    static float data[BATCH_SIZE * 3 * INPUT_H * INPUT_W];
+    static float prob[BATCH_SIZE * OUTPUT_SIZE];
+    IRuntime* runtime = createInferRuntime(gLogger);
+    assert(runtime != nullptr);
+    ICudaEngine* engine = runtime->deserializeCudaEngine(trtModelStream, size);
+    assert(engine != nullptr);
+    IExecutionContext* context = engine->createExecutionContext();
+    assert(context != nullptr);
+    delete[] trtModelStream;
+    assert(engine->getNbBindings() == 2);
+    float* buffers[2];
+    // In order to bind the buffers, we need to know the names of the input and output tensors.
+    // Note that indices are guaranteed to be less than IEngine::getNbBindings()
+    const int inputIndex = engine->getBindingIndex(INPUT_BLOB_NAME);
+    const int outputIndex = engine->getBindingIndex(OUTPUT_BLOB_NAME);
+    assert(inputIndex == 0);
+    assert(outputIndex == 1);
+    // Create GPU buffers on device
+    CUDA_CHECK(cudaMalloc((void**)&buffers[inputIndex], BATCH_SIZE * 3 * INPUT_H * INPUT_W * sizeof(float)));
+    CUDA_CHECK(cudaMalloc((void**)&buffers[outputIndex], BATCH_SIZE * OUTPUT_SIZE * sizeof(float)));
+
+    // Create stream
+    cudaStream_t stream;
+    CUDA_CHECK(cudaStreamCreate(&stream));
+    uint8_t* img_host = nullptr;
+    uint8_t* img_device = nullptr;
+    // prepare input data cache in pinned memory 
+    CUDA_CHECK(cudaMallocHost((void**)&img_host, MAX_IMAGE_INPUT_SIZE_THRESH * 3));
+    // prepare input data cache in device memory
+    CUDA_CHECK(cudaMalloc((void**)&img_device, MAX_IMAGE_INPUT_SIZE_THRESH * 3));
+    int fcount = 0;
+    std::vector<cv::Mat> imgs_buffer(BATCH_SIZE);
+    for (int f = 0; f < (int)file_names.size(); f++) {
+        fcount++;
+        if (fcount < BATCH_SIZE && f + 1 != (int)file_names.size()) continue;
+        //auto start = std::chrono::system_clock::now();
+        float* buffer_idx = (float*)buffers[inputIndex];
+        for (int b = 0; b < fcount; b++) {
+            cv::Mat img = cv::imread(img_dir + "/" + file_names[f - fcount + 1 + b]);
+            if (img.empty()) continue;
+            size_t  size_image = img.cols * img.rows * 3;
+            size_t  size_image_dst = INPUT_H * INPUT_W * 3;
+            cv::Mat pr_img;
+            cv::resize(img, pr_img, cv::Size(INPUT_W, INPUT_H));
+            int i = 0;
+            for (int row = 0; row < INPUT_H; ++row) {
+                uchar* uc_pixel = pr_img.data + row * pr_img.step;
+                for (int col = 0; col < INPUT_W; ++col) {
+                    data[b * 3 * INPUT_H * INPUT_W + i] = ((float)uc_pixel[2] / 255.0 - 0.485) / 0.229; // R-0.485
+                    data[b * 3 * INPUT_H * INPUT_W + i + INPUT_H * INPUT_W] = ((float)uc_pixel[1] / 255.0 - 0.456) / 0.224;
+                    data[b * 3 * INPUT_H * INPUT_W + i + 2 * INPUT_H * INPUT_W] = ((float)uc_pixel[0] / 255.0 - 0.406) / 0.225;
+                    uc_pixel += 3;
+                    ++i;
+                }
+            }
+            //copy data to pinned memory
+            memcpy(img_host,data,size_image);
+            //copy data to device memory
+            CUDA_CHECK(cudaMemcpyAsync(img_device,img_host,size_image,cudaMemcpyHostToDevice,stream));
+            buffer_idx += size_image_dst;
+        }
+        // Run inference
+        auto start = std::chrono::system_clock::now();
+        doInference(*context, stream, (void**)buffers, prob, BATCH_SIZE);
+        auto end = std::chrono::system_clock::now();
+        std::cout << "inference time: " << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count() << "ms" << std::endl;
+
+        fcount = 0;
+    }
+
+    // Release stream and buffers
+    cudaStreamDestroy(stream);
+    CUDA_CHECK(cudaFree(img_device));
+    CUDA_CHECK(cudaFreeHost(img_host));
+    CUDA_CHECK(cudaFree(buffers[inputIndex]));
+    CUDA_CHECK(cudaFree(buffers[outputIndex]));
+    // Destroy the engine
+    context->destroy();
+    engine->destroy();
+    runtime->destroy();
+
+    return 0;
+}
diff --git a/yolov5/yolov5_cls_trt.py b/yolov5/yolov5_cls_trt.py
new file mode 100644
index 0000000..6f5507b
--- /dev/null
+++ b/yolov5/yolov5_cls_trt.py
@@ -0,0 +1,249 @@
+"""
+An example that uses TensorRT's Python api to make inferences.
+"""
+import os
+import shutil
+import sys
+import threading
+import time
+import cv2
+import numpy as np
+import torch
+import pycuda.autoinit
+import pycuda.driver as cuda
+import tensorrt as trt
+
+
+def get_img_path_batches(batch_size, img_dir):
+    ret = []
+    batch = []
+    for root, dirs, files in os.walk(img_dir):
+        for name in files:
+            if len(batch) == batch_size:
+                ret.append(batch)
+                batch = []
+            batch.append(os.path.join(root, name))
+    if len(batch) > 0:
+        ret.append(batch)
+    return ret
+
+
+with open("imagenet_classes.txt") as f:
+    classes = [line.strip() for line in f.readlines()]
+
+
+class YoLov5TRT(object):
+    """
+    description: A YOLOv5 class that warps TensorRT ops, preprocess and postprocess ops.
+    """
+
+    def __init__(self, engine_file_path):
+        # Create a Context on this device,
+        self.ctx = cuda.Device(0).make_context()
+        stream = cuda.Stream()
+        TRT_LOGGER = trt.Logger(trt.Logger.INFO)
+        runtime = trt.Runtime(TRT_LOGGER)
+
+        # Deserialize the engine from file
+        with open(engine_file_path, "rb") as f:
+            engine = runtime.deserialize_cuda_engine(f.read())
+        context = engine.create_execution_context()
+
+        host_inputs = []
+        cuda_inputs = []
+        host_outputs = []
+        cuda_outputs = []
+        bindings = []
+        self.mean = (0.485, 0.456, 0.406)
+        self.std = (0.229, 0.224, 0.225)
+
+        for binding in engine:
+            print('binding:', binding, engine.get_binding_shape(binding))
+            size = trt.volume(engine.get_binding_shape(
+                binding)) * engine.max_batch_size
+            dtype = trt.nptype(engine.get_binding_dtype(binding))
+            # Allocate host and device buffers
+            host_mem = cuda.pagelocked_empty(size, dtype)
+            cuda_mem = cuda.mem_alloc(host_mem.nbytes)
+            # Append the device buffer to device bindings.
+            bindings.append(int(cuda_mem))
+            # Append to the appropriate list.
+            if engine.binding_is_input(binding):
+                self.input_w = engine.get_binding_shape(binding)[-1]
+                self.input_h = engine.get_binding_shape(binding)[-2]
+                host_inputs.append(host_mem)
+                cuda_inputs.append(cuda_mem)
+            else:
+                host_outputs.append(host_mem)
+                cuda_outputs.append(cuda_mem)
+
+        # Store
+        self.stream = stream
+        self.context = context
+        self.engine = engine
+        self.host_inputs = host_inputs
+        self.cuda_inputs = cuda_inputs
+        self.host_outputs = host_outputs
+        self.cuda_outputs = cuda_outputs
+        self.bindings = bindings
+        self.batch_size = engine.max_batch_size
+
+    def infer(self, raw_image_generator):
+        threading.Thread.__init__(self)
+        # Make self the active context, pushing it on top of the context stack.
+        self.ctx.push()
+        # Restore
+        stream = self.stream
+        context = self.context
+        engine = self.engine
+        host_inputs = self.host_inputs
+        cuda_inputs = self.cuda_inputs
+        host_outputs = self.host_outputs
+        cuda_outputs = self.cuda_outputs
+        bindings = self.bindings
+        # Do image preprocess
+        batch_image_raw = []
+        batch_input_image = np.empty(
+            shape=[self.batch_size, 3, self.input_h, self.input_w])
+        for i, image_raw in enumerate(raw_image_generator):
+            batch_image_raw.append(image_raw)
+            input_image = self.preprocess_cls_image(image_raw)
+            np.copyto(batch_input_image[i], input_image)
+        batch_input_image = np.ascontiguousarray(batch_input_image)
+
+        # Copy input image to host buffer
+        np.copyto(host_inputs[0], batch_input_image.ravel())
+        start = time.time()
+        # Transfer input data  to the GPU.
+        cuda.memcpy_htod_async(cuda_inputs[0], host_inputs[0], stream)
+        # Run inference.
+        context.execute_async(batch_size=self.batch_size,
+                              bindings=bindings, stream_handle=stream.handle)
+        # Transfer predictions back from the GPU.
+        cuda.memcpy_dtoh_async(host_outputs[0], cuda_outputs[0], stream)
+        # Synchronize the stream
+        stream.synchronize()
+        end = time.time()
+        # Remove any context from the top of the context stack, deactivating it.
+        self.ctx.pop()
+        # Here we use the first row of output in that batch_size = 1
+        output = host_outputs[0]
+        # Do postprocess
+        for i in range(self.batch_size):
+            classes_ls, predicted_conf_ls, category_id_ls = self.postprocess_cls(
+                output)
+            cv2.putText(batch_image_raw[i], str(
+                classes_ls), (10, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 1, cv2.LINE_AA)
+        return batch_image_raw, end - start
+
+    def destroy(self):
+        # Remove any context from the top of the context stack, deactivating it.
+        self.ctx.pop()
+
+    def get_raw_image(self, image_path_batch):
+        """
+        description: Read an image from image path
+        """
+        for img_path in image_path_batch:
+            yield cv2.imread(img_path)
+
+    def get_raw_image_zeros(self, image_path_batch=None):
+        """
+        description: Ready data for warmup
+        """
+        for _ in range(self.batch_size):
+            yield np.zeros([self.input_h, self.input_w, 3], dtype=np.uint8)
+
+    def preprocess_cls_image(self, input_img):
+        im = cv2.cvtColor(input_img, cv2.COLOR_BGR2RGB)
+        im = cv2.resize(im, (self.input_h, self.input_w))
+        im = np.float32(im)
+        im /= 255.0
+        im -= self.mean
+        im /= self.std
+        im = im.transpose(2, 0, 1)
+        # prepare batch
+        batch_data = np.expand_dims(im, axis=0)
+        return batch_data
+
+    def postprocess_cls(self, output_data):
+        classes_ls = []
+        predicted_conf_ls = []
+        category_id_ls = []
+        output_data = output_data.reshape(self.batch_size, -1)
+        output_data = torch.Tensor(output_data)
+        p = torch.nn.functional.softmax(output_data, dim=1)
+        score, index = torch.topk(output_data, 3)
+        for ind in range(index.shape[0]):
+            input_category_id = index[ind][0].item()  # 716
+            category_id_ls.append(input_category_id)
+            predicted_confidence = score[ind][0].item()
+            predicted_conf_ls.append(predicted_confidence)
+            classes_ls.append(classes[input_category_id])
+        return classes_ls, predicted_conf_ls, category_id_ls
+
+
+class inferThread(threading.Thread):
+    def __init__(self, yolov5_wrapper, image_path_batch):
+        threading.Thread.__init__(self)
+        self.yolov5_wrapper = yolov5_wrapper
+        self.image_path_batch = image_path_batch
+
+    def run(self):
+        batch_image_raw, use_time = self.yolov5_wrapper.infer(
+            self.yolov5_wrapper.get_raw_image(self.image_path_batch))
+        for i, img_path in enumerate(self.image_path_batch):
+            parent, filename = os.path.split(img_path)
+            save_name = os.path.join('output', filename)
+            # Save image
+            cv2.imwrite(save_name, batch_image_raw[i])
+        print('input->{}, time->{:.2f}ms, saving into output/'.format(
+            self.image_path_batch, use_time * 1000))
+
+
+class warmUpThread(threading.Thread):
+    def __init__(self, yolov5_wrapper):
+        threading.Thread.__init__(self)
+        self.yolov5_wrapper = yolov5_wrapper
+
+    def run(self):
+        batch_image_raw, use_time = self.yolov5_wrapper.infer(
+            self.yolov5_wrapper.get_raw_image_zeros())
+        print(
+            'warm_up->{}, time->{:.2f}ms'.format(batch_image_raw[0].shape, use_time * 1000))
+
+
+if __name__ == "__main__":
+    # load custom plugin and engine
+    engine_file_path = "build/yolov5s_cls.engine"
+
+    if len(sys.argv) > 1:
+        engine_file_path = sys.argv[1]
+    if len(sys.argv) > 2:
+        PLUGIN_LIBRARY = sys.argv[2]
+
+    if os.path.exists('output/'):
+        shutil.rmtree('output/')
+    os.makedirs('output/')
+    # a YoLov5TRT instance
+    yolov5_wrapper = YoLov5TRT(engine_file_path)
+    try:
+        print('batch size is', yolov5_wrapper.batch_size)
+
+        image_dir = "samples/"
+        image_path_batches = get_img_path_batches(
+            yolov5_wrapper.batch_size, image_dir)
+
+        for i in range(10):
+            # create a new thread to do warm_up
+            thread1 = warmUpThread(yolov5_wrapper)
+            thread1.start()
+            thread1.join()
+        for batch in image_path_batches:
+            # create a new thread to do inference
+            thread1 = inferThread(yolov5_wrapper, batch)
+            thread1.start()
+            thread1.join()
+    finally:
+        # destroy the instance
+        yolov5_wrapper.destroy()