diff --git a/research/cv/sknet/README.md b/research/cv/sknet/README.md
index 65549807e56a101ecabb440651f67597e8c752ed..3d5d3042934b6b62d209f898c4d24786d8854600 100644
--- a/research/cv/sknet/README.md
+++ b/research/cv/sknet/README.md
@@ -1,21 +1,38 @@
 # Contents
 
-- [SK-Net Description](#sK-net-description)
+- [Contents](#contents)
+- [SK-Net Description](#sk-net-description)
+- [Description](#description)
+- [Paper](#paper)
 - [Model Architecture](#model-architecture)
 - [Dataset](#dataset)
 - [Features](#features)
-    - [Mixed Precision](#mixed-precision)
+- [Mixed Precision](#mixed-precision)
 - [Environment Requirements](#environment-requirements)
 - [Quick Start](#quick-start)
 - [Script Description](#script-description)
-    - [Script and Sample Code](#script-and-sample-code)
-    - [Script Parameters](#script-parameters)
-    - [Training Process](#training-process)
-    - [Evaluation Process](#evaluation-process)
+- [Script and Sample Code](#script-and-sample-code)
+- [Script Parameters](#script-parameters)
+- [Training Process](#training-process)
+- [Usage](#usage)
+- [Running on Ascend](#running-on-ascend)
+- [Result](#result)
+- [Evaluation Process](#evaluation-process)
+- [Usage](#usage-1)
+- [Running on Ascend](#running-on-ascend-1)
+- [Result](#result-1)
+- [Inference Process](#inference-process)
+- [Export MindIR](#export-mindir)
+- [Infer on Ascend310](#infer-on-ascend310)
+- [Result](#result-2)
 - [Model Description](#model-description)
-    - [Performance](#performance)
-        - [Evaluation Performance](#evaluation-performance)
-        - [Inference Performance](#inference-performance)
+- [Performance](#performance)
+- [Evaluation Performance](#evaluation-performance)
+- [SKNet50 on CIFRA10](#sknet50-on-cifra10)
+- [Inference Performance](#inference-performance)
+- [SKNet50 on CIFAR10](#sknet50-on-cifar10)
+- [310 Inference Performance](#310-inference-performance)
+- [SKNet50 on CIFAR10](#sknet50-on-cifar10-1)
 - [Description of Random Situation](#description-of-random-situation)
 - [ModelZoo Homepage](#modelzoo-homepage)
 
@@ -93,10 +110,16 @@ python eval.py --checkpoint_path=/resnet/sknet_90.ckpt --dataset_path=/data/cifa
 ```text
 鈹斺攢鈹€SK-Net
   鈹溾攢鈹€ README.md
+  鈹溾攢鈹€ ascend310_infer
+    鈹溾攢鈹€ inc
+    鈹溾攢鈹€ src
+    鈹溾攢鈹€ build.sh                         # make process  
+    鈹溾攢鈹€ CMakeLists.txt                   # cmake configuration  
   鈹溾攢鈹€ scripts
     鈹溾攢鈹€ run_distribute_train.sh            # launch ascend distributed training(8 pcs)
     鈹溾攢鈹€ run_eval.sh                        # launch ascend evaluation
     鈹溾攢鈹€ run_standalone_train.sh            # launch ascend standalone training(1 pcs)
+    鈹溾攢鈹€ run_infer_310.sh                   # launch 310 infer  
   鈹溾攢鈹€ src
     鈹溾攢鈹€ config.py                          # parameter configuration
     鈹溾攢鈹€ CrossEntropySmooth.py              # loss definition
@@ -108,6 +131,8 @@ python eval.py --checkpoint_path=/resnet/sknet_90.ckpt --dataset_path=/data/cifa
   鈹溾攢鈹€ export.py                            # export model for inference
   鈹溾攢鈹€ eval.py                              # eval net
   鈹斺攢鈹€ train.py                             # train net
+  鈹溾攢鈹€ preprocess.py                        # preprocess scripts
+  鈹溾攢鈹€ postprocess.py                       # postprocess scripts
 ```
 
 ## [Script Parameters](#contents)
@@ -187,6 +212,41 @@ bash run_eval.sh [DATASET_PATH]  [CHECKPOINT_PATH]
 result: {'top_5_accuracy': 0.9982972756410257, 'top_1_accuracy': 0.9449118589743589}
 ```
 
+## [Inference Process](#contents)
+
+### Export MindIR
+
+```bash
+python export.py --ckpt_file [CKPT_PATH] --file_name [FILE_NAME] --file_format [FILE_FORMAT]
+```
+
+The ckpt_file parameter is required,
+`FILE_NAME` is the name of the AIR/ONNX/MINDIR file.
+`FILE_FORMAT` should be in ["AIR","ONNX", "MINDIR"]
+
+### Infer on Ascend310
+
+Before performing inference, the mindir file must be exported by `export.py` script. We only provide an example of inference using MINDIR model.
+
+```shell
+# Ascend310 inference
+bash run_infer_310.sh [MINDIR_PATH] [DATASET_NAME] [DATASET_PATH] [NEED PREPROCESS] [DEVICE_ID]
+```
+
+- DATASET_NAME can choose from ['cifar10', 'imagenet2012'].
+- NEED_PREPROCESS means weather need preprocess or not, it's value is 'y' or 'n'.
+- DEVICE_ID is optional, it can be set by environment variable device_id, otherwise the value is zero"
+- `DVPP` is mandatory, and must choose from ["DVPP", "CPU"], it's case-insensitive. SE-net only support CPU mode.
+- `DEVICE_ID` is optional, default value is 0.
+
+### Result
+
+Inference result is saved in current path, you can find result like this in acc.log file.
+
+```bash
+result: {'top_1_accuracy': 0.9449118589743589}
+```
+
 # [Model Description](#contents)
 
 ## [Performance](#contents)
@@ -227,6 +287,20 @@ result: {'top_5_accuracy': 0.9982972756410257, 'top_1_accuracy': 0.9449118589743
 | batch_size          | 32                          |
 | Accuracy            | 94.49%                      |
 
+### 310 Inference Performance
+
+#### SKNet50 on CIFAR10
+
+| Parameters          | Ascend                      |
+| ------------------- | --------------------------- |
+| Model Version       | SKNet50                 |
+| Resource            | Ascend 310                  |
+| Uploaded Date       | 09/23/2021 (month/day/year) |
+| MindSpore Version   | 1.3.0                 |
+| Dataset             | CIFAR10                |
+| batch_size          | 32                          |
+| Accuracy            | 95.49%                      |
+
 # [Description of Random Situation](#contents)
 
 In dataset.py, we set the seed inside "create_dataset" function. We also use random seed in train.py.
diff --git a/research/cv/sknet/ascend310_infer/CMakeLists.txt b/research/cv/sknet/ascend310_infer/CMakeLists.txt
new file mode 100644
index 0000000000000000000000000000000000000000..49478f950647bd15851ae9b931c04ed190ef9cbf
--- /dev/null
+++ b/research/cv/sknet/ascend310_infer/CMakeLists.txt
@@ -0,0 +1,16 @@
+cmake_minimum_required(VERSION 3.14.1)
+project(Ascend310Infer)
+find_package(OpenCV 2 REQUIRED)
+find_package(gflags REQUIRED)
+add_compile_definitions(_GLIBCXX_USE_CXX11_ABI=0)
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O0 -g -std=c++17 -Werror -Wall -fPIE -Wl,--allow-shlib-undefined")
+set(PROJECT_SRC_ROOT ${CMAKE_CURRENT_LIST_DIR}/)
+option(MINDSPORE_PATH "mindspore install path" "")
+include_directories(${OpenCV_INCLUDE_DIRS})
+include_directories(${MINDSPORE_PATH})
+include_directories(${MINDSPORE_PATH}/include)
+include_directories(${PROJECT_SRC_ROOT})
+find_library(MS_LIB libmindspore.so ${MINDSPORE_PATH}/lib)
+file(GLOB_RECURSE MD_LIB ${MINDSPORE_PATH}/_c_dataengine*)
+add_executable(main src/main.cc src/utils.cc)
+target_link_libraries(main  ${MS_LIB} ${MD_LIB} ${OpenCV_LIBS} gflags)
diff --git a/research/cv/sknet/ascend310_infer/build.sh b/research/cv/sknet/ascend310_infer/build.sh
new file mode 100644
index 0000000000000000000000000000000000000000..922df6cb4d87095802306ba30c7741fe64ca79f1
--- /dev/null
+++ b/research/cv/sknet/ascend310_infer/build.sh
@@ -0,0 +1,28 @@
+#!/bin/bash
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+
+if [ ! -d out ]; then 
+  mkdir out
+fi
+
+cd out || exit
+
+if [ -f "Makefile" ]; then
+  make clean
+fi
+
+cmake ..  -DMINDSPORE_PATH="`pip3.7 show mindspore-ascend | grep Location | awk '{print $2"/mindspore"}' | xargs realpath`"
+make
diff --git a/research/cv/sknet/ascend310_infer/inc/utils.h b/research/cv/sknet/ascend310_infer/inc/utils.h
new file mode 100644
index 0000000000000000000000000000000000000000..f8ae1e5b473d869b77af8d725a280d7c7665527c
--- /dev/null
+++ b/research/cv/sknet/ascend310_infer/inc/utils.h
@@ -0,0 +1,35 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ * 
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ * 
+ * http://www.apache.org/licenses/LICENSE-2.0
+ * 
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef MINDSPORE_INFERENCE_UTILS_H_
+#define MINDSPORE_INFERENCE_UTILS_H_
+
+#include <sys/stat.h>
+#include <dirent.h>
+#include <vector>
+#include <string>
+#include <memory>
+#include "include/api/types.h"
+
+std::vector<std::string> GetAllFiles(std::string_view dirName);
+DIR *OpenDir(std::string_view dirName);
+std::string RealPath(std::string_view path);
+mindspore::MSTensor ReadFileToTensor(const std::string &file);
+int WriteResult(const std::string& imageFile, const std::vector<mindspore::MSTensor> &outputs);
+std::vector<std::string> GetAllFiles(std::string dir_name);
+std::vector<std::vector<std::string>> GetAllInputData(std::string dir_name);
+
+#endif
diff --git a/research/cv/sknet/ascend310_infer/src/main.cc b/research/cv/sknet/ascend310_infer/src/main.cc
new file mode 100644
index 0000000000000000000000000000000000000000..9862fcfb6b731c18a1b6d5ec6077c9b248e1461e
--- /dev/null
+++ b/research/cv/sknet/ascend310_infer/src/main.cc
@@ -0,0 +1,148 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ * 
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ * 
+ * http://www.apache.org/licenses/LICENSE-2.0
+ * 
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+#include <sys/time.h>
+#include <gflags/gflags.h>
+#include <dirent.h>
+#include <iostream>
+#include <string>
+#include <algorithm>
+#include <iosfwd>
+#include <vector>
+#include <fstream>
+#include <sstream>
+
+#include "include/api/model.h"
+#include "include/api/context.h"
+#include "include/api/types.h"
+#include "include/api/serialization.h"
+#include "include/dataset/vision_ascend.h"
+#include "include/dataset/execute.h"
+#include "include/dataset/transforms.h"
+#include "include/dataset/vision.h"
+#include "inc/utils.h"
+
+using mindspore::Context;
+using mindspore::Serialization;
+using mindspore::Model;
+using mindspore::Status;
+using mindspore::ModelType;
+using mindspore::GraphCell;
+using mindspore::kSuccess;
+using mindspore::MSTensor;
+using mindspore::dataset::Execute;
+using mindspore::dataset::vision::Decode;
+using mindspore::dataset::vision::Resize;
+using mindspore::dataset::vision::CenterCrop;
+using mindspore::dataset::vision::Normalize;
+using mindspore::dataset::vision::HWC2CHW;
+
+
+DEFINE_string(mindir_path, "", "mindir path");
+DEFINE_string(dataset_name, "cifar10", "['cifar10', 'imagenet2012']");
+DEFINE_string(input0_path, ".", "input0 path");
+DEFINE_int32(device_id, 0, "device id");
+
+int load_model(Model *model, std::vector<MSTensor> *model_inputs, std::string mindir_path, int device_id) {
+  if (RealPath(mindir_path).empty()) {
+    std::cout << "Invalid mindir" << std::endl;
+    return 1;
+  }
+  std::cout << "good mindir" << std::endl;
+  auto context = std::make_shared<Context>();
+  auto ascend310 = std::make_shared<mindspore::Ascend310DeviceInfo>();
+  ascend310->SetDeviceID(device_id);
+  std::cout << "find device" << std::endl;
+  context->MutableDeviceInfo().push_back(ascend310);
+  mindspore::Graph graph;
+  Serialization::Load(mindir_path, ModelType::kMindIR, &graph);
+
+  Status ret = model->Build(GraphCell(graph), context);
+  if (ret != kSuccess) {
+    std::cout << "ERROR: Build failed." << std::endl;
+    return 1;
+  }
+
+  *model_inputs = model->GetInputs();
+  if (model_inputs->empty()) {
+    std::cout << "Invalid model, inputs is empty." << std::endl;
+    return 1;
+  }
+  return 0;
+}
+
+int main(int argc, char **argv) {
+  gflags::ParseCommandLineFlags(&argc, &argv, true);
+  Model model;
+  std::vector<MSTensor> model_inputs;
+  load_model(&model, &model_inputs, FLAGS_mindir_path, FLAGS_device_id);
+  std::cout << "load model success" << std::endl;
+  std::map<double, double> costTime_map;
+  struct timeval start = {0};
+  struct timeval end = {0};
+  if (FLAGS_dataset_name == "cifar10") {
+    auto input0_files = GetAllFiles(FLAGS_input0_path);
+    if (input0_files.empty()) {
+      std::cout << "ERROR: no input data." << std::endl;
+      return 1;
+    }
+    std::cout << "find input data" << std::endl;
+    size_t size = input0_files.size();
+    for (size_t i = 0; i < size; ++i) {
+      std::vector<MSTensor> inputs;
+      std::vector<MSTensor> outputs;
+      double startTimeMs;
+      double endTimeMs;
+      std::cout << "Start predict input files:" << input0_files[i] <<std::endl;
+      auto input0 = ReadFileToTensor(input0_files[i]);
+      inputs.emplace_back(model_inputs[0].Name(), model_inputs[0].DataType(), model_inputs[0].Shape(),
+                          input0.Data().get(), input0.DataSize());
+
+      gettimeofday(&start, nullptr);
+      Status ret = model.Predict(inputs, &outputs);
+      gettimeofday(&end, nullptr);
+      if (ret != kSuccess) {
+        std::cout << "Predict " << input0_files[i] << " failed." << std::endl;
+        return 1;
+      }
+      startTimeMs = (1.0 * start.tv_sec * 1000000 + start.tv_usec) / 1000;
+      endTimeMs = (1.0 * end.tv_sec * 1000000 + end.tv_usec) / 1000;
+      costTime_map.insert(std::pair<double, double>(startTimeMs, endTimeMs));
+      int rst = WriteResult(input0_files[i], outputs);
+      if (rst != 0) {
+          std::cout << "write result failed." << std::endl;
+          return rst;
+      }
+    }
+  }
+  double average = 0.0;
+  int inferCount = 0;
+  for (auto iter = costTime_map.begin(); iter != costTime_map.end(); iter++) {
+    double diff = 0.0;
+    diff = iter->second - iter->first;
+    average += diff;
+    inferCount++;
+  }
+  average = average / inferCount;
+  std::stringstream timeCost;
+  timeCost << "NN inference cost average time: "<< average << " ms of infer_count " << inferCount << std::endl;
+  std::cout << "NN inference cost average time: "<< average << "ms of infer_count " << inferCount << std::endl;
+  std::string fileName = "./time_Result" + std::string("/test_perform_static.txt");
+  std::ofstream fileStream(fileName.c_str(), std::ios::trunc);
+  fileStream << timeCost.str();
+  fileStream.close();
+  costTime_map.clear();
+  return 0;
+}
diff --git a/research/cv/sknet/ascend310_infer/src/utils.cc b/research/cv/sknet/ascend310_infer/src/utils.cc
new file mode 100644
index 0000000000000000000000000000000000000000..d71f388b83d23c2813d8bfc883dbcf2e7e0e4ef0
--- /dev/null
+++ b/research/cv/sknet/ascend310_infer/src/utils.cc
@@ -0,0 +1,185 @@
+/**
+ * Copyright 2021 Huawei Technologies Co., Ltd
+ * 
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ * 
+ * http://www.apache.org/licenses/LICENSE-2.0
+ * 
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <fstream>
+#include <algorithm>
+#include <iostream>
+#include "inc/utils.h"
+
+using mindspore::MSTensor;
+using mindspore::DataType;
+
+
+std::vector<std::vector<std::string>> GetAllInputData(std::string dir_name) {
+  std::vector<std::vector<std::string>> ret;
+
+  DIR *dir = OpenDir(dir_name);
+  if (dir == nullptr) {
+    return {};
+  }
+  struct dirent *filename;
+  /* read all the files in the dir ~ */
+  std::vector<std::string> sub_dirs;
+  while ((filename = readdir(dir)) != nullptr) {
+    std::string d_name = std::string(filename->d_name);
+    // get rid of "." and ".."
+    if (d_name == "." || d_name == ".." || d_name.empty()) {
+      continue;
+    }
+    std::string dir_path = RealPath(std::string(dir_name) + "/" + filename->d_name);
+    struct stat s;
+    lstat(dir_path.c_str(), &s);
+    if (!S_ISDIR(s.st_mode)) {
+      continue;
+    }
+
+    sub_dirs.emplace_back(dir_path);
+  }
+  std::sort(sub_dirs.begin(), sub_dirs.end());
+
+  (void)std::transform(sub_dirs.begin(), sub_dirs.end(), std::back_inserter(ret),
+                       [](const std::string &d) { return GetAllFiles(d); });
+
+  return ret;
+}
+
+
+std::vector<std::string> GetAllFiles(std::string dir_name) {
+  struct dirent *filename;
+  DIR *dir = OpenDir(dir_name);
+  if (dir == nullptr) {
+    return {};
+  }
+
+  std::vector<std::string> res;
+  while ((filename = readdir(dir)) != nullptr) {
+    std::string d_name = std::string(filename->d_name);
+    if (d_name == "." || d_name == ".." || d_name.size() <= 3) {
+      continue;
+    }
+    res.emplace_back(std::string(dir_name) + "/" + filename->d_name);
+  }
+  std::sort(res.begin(), res.end());
+
+  return res;
+}
+
+
+std::vector<std::string> GetAllFiles(std::string_view dirName) {
+  struct dirent *filename;
+  DIR *dir = OpenDir(dirName);
+  if (dir == nullptr) {
+    return {};
+  }
+  std::vector<std::string> res;
+  while ((filename = readdir(dir)) != nullptr) {
+    std::string dName = std::string(filename->d_name);
+    if (dName == "." || dName == ".." || filename->d_type != DT_REG) {
+      continue;
+    }
+    res.emplace_back(std::string(dirName) + "/" + filename->d_name);
+  }
+  std::sort(res.begin(), res.end());
+  for (auto &f : res) {
+    std::cout << "image file: " << f << std::endl;
+  }
+  return res;
+}
+
+
+int WriteResult(const std::string& imageFile, const std::vector<MSTensor> &outputs) {
+  std::string homePath = "./result_Files";
+  for (size_t i = 0; i < outputs.size(); ++i) {
+    size_t outputSize;
+    std::shared_ptr<const void> netOutput;
+    netOutput = outputs[i].Data();
+    outputSize = outputs[i].DataSize();
+    int pos = imageFile.rfind('/');
+    std::string fileName(imageFile, pos + 1);
+    fileName.replace(fileName.find('.'), fileName.size() - fileName.find('.'), '_' + std::to_string(i) + ".bin");
+    std::string outFileName = homePath + "/" + fileName;
+    FILE *outputFile = fopen(outFileName.c_str(), "wb");
+    fwrite(netOutput.get(), outputSize, sizeof(char), outputFile);
+    fclose(outputFile);
+    outputFile = nullptr;
+  }
+  return 0;
+}
+
+mindspore::MSTensor ReadFileToTensor(const std::string &file) {
+  if (file.empty()) {
+    std::cout << "Pointer file is nullptr" << std::endl;
+    return mindspore::MSTensor();
+  }
+
+  std::ifstream ifs(file);
+  if (!ifs.good()) {
+    std::cout << "File: " << file << " is not exist" << std::endl;
+    return mindspore::MSTensor();
+  }
+
+  if (!ifs.is_open()) {
+    std::cout << "File: " << file << "open failed" << std::endl;
+    return mindspore::MSTensor();
+  }
+
+  ifs.seekg(0, std::ios::end);
+  size_t size = ifs.tellg();
+  mindspore::MSTensor buffer(file, mindspore::DataType::kNumberTypeUInt8, {static_cast<int64_t>(size)}, nullptr, size);
+
+  ifs.seekg(0, std::ios::beg);
+  ifs.read(reinterpret_cast<char *>(buffer.MutableData()), size);
+  ifs.close();
+
+  return buffer;
+}
+
+
+DIR *OpenDir(std::string_view dirName) {
+  if (dirName.empty()) {
+    std::cout << " dirName is null ! " << std::endl;
+    return nullptr;
+  }
+  std::string realPath = RealPath(dirName);
+  struct stat s;
+  lstat(realPath.c_str(), &s);
+  if (!S_ISDIR(s.st_mode)) {
+    std::cout << "dirName is not a valid directory !" << std::endl;
+    return nullptr;
+  }
+  DIR *dir;
+  dir = opendir(realPath.c_str());
+  if (dir == nullptr) {
+    std::cout << "Can not open dir " << dirName << std::endl;
+    return nullptr;
+  }
+  std::cout << "Successfully opened the dir " << dirName << std::endl;
+  return dir;
+}
+
+std::string RealPath(std::string_view path) {
+  char realPathMem[PATH_MAX] = {0};
+  char *realPathRet = nullptr;
+  realPathRet = realpath(path.data(), realPathMem);
+  if (realPathRet == nullptr) {
+    std::cout << "File: " << path << " is not exist.";
+    return "";
+  }
+
+  std::string realPath(realPathMem);
+  std::cout << path << " realpath is: " << realPath << std::endl;
+  return realPath;
+}
diff --git a/research/cv/sknet/eval.py b/research/cv/sknet/eval.py
index 6cbd6b3715dcb982923abbecce561929a8278323..073ec53bc0db85f44b40103588b6d1d57ff8c1f8 100644
--- a/research/cv/sknet/eval.py
+++ b/research/cv/sknet/eval.py
@@ -37,7 +37,7 @@ if __name__ == '__main__':
         from src.sknet50 import sknet50 as sknet
         if args_opt.dataset == "cifar10":
             from src.config import config1 as config
-            from src.dataset import create_dataset1 as create_dataset
+            from src.dataset import create_dataset_cifar10 as create_dataset
 
     target = args_opt.device_target
     # init context
diff --git a/research/cv/sknet/postprocess.py b/research/cv/sknet/postprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..98bc55da6ce07acfc2eecb548f5ad77a7dafa257
--- /dev/null
+++ b/research/cv/sknet/postprocess.py
@@ -0,0 +1,38 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""postprocess for 310 inference"""
+import os
+import argparse
+import numpy as np
+from mindspore.nn import Top1CategoricalAccuracy
+parser = argparse.ArgumentParser(description="postprocess")
+label_path = "./preprocess_Result/cifar10_label_ids.npy"
+parser.add_argument("--result_dir", type=str, default="./result_Files", help="result files path.")
+parser.add_argument("--label_dir", type=str, default=label_path, help="image file path.")
+args = parser.parse_args()
+
+def calcul_acc(lab, preds):
+    return sum(1 for x, y in zip(lab, preds) if x == y) / len(lab)
+if __name__ == '__main__':
+    top1_acc = Top1CategoricalAccuracy()
+    rst_path = args.result_dir
+    labels = np.load(args.label_dir, allow_pickle=True)
+    batch_size = 32
+    for idx, label in enumerate(labels):
+        f_name = os.path.join(rst_path, "sknet_data_bs" + str(batch_size) + "_" + str(idx) + "_0.bin")
+        pred = np.fromfile(f_name, np.float32)
+        pred = pred.reshape(batch_size, int(pred.shape[0] / batch_size))
+        top1_acc.update(pred, labels[idx])
+    print("acc: ", top1_acc.eval())
diff --git a/research/cv/sknet/preprocess.py b/research/cv/sknet/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..6b2133f4acddd5a7e1ed4452976f8bbcb4f57f6d
--- /dev/null
+++ b/research/cv/sknet/preprocess.py
@@ -0,0 +1,36 @@
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+"""preprocess"""
+import os
+import argparse
+import numpy as np
+from src.dataset import create_dataset_cifar10
+parser = argparse.ArgumentParser('preprocess')
+parser.add_argument('--data_path', type=str, default='', help='eval data dir')
+
+args = parser.parse_args()
+if __name__ == "__main__":
+    dataset = create_dataset_cifar10(args.data_path, False)
+    img_path = os.path.join('./preprocess_Result/', "00_data")
+    os.makedirs(img_path)
+    label_list = []
+    batch_size = 32
+    for idx, data in enumerate(dataset.create_dict_iterator(output_numpy=True)):
+        file_name = "sknet_data_bs" + str(batch_size) + "_" + str(idx) + ".bin"
+        file_path = os.path.join(img_path, file_name)
+        data["image"].tofile(file_path)
+        label_list.append(data["label"])
+    np.save(os.path.join('./preprocess_Result/', "cifar10_label_ids.npy"), label_list)
+    print("=" * 20, "export bin files finished", "=" * 20)
diff --git a/research/cv/sknet/scripts/run_infer_310.sh b/research/cv/sknet/scripts/run_infer_310.sh
new file mode 100644
index 0000000000000000000000000000000000000000..8c710986ac045cd3323cf4aab7c5332136b731a5
--- /dev/null
+++ b/research/cv/sknet/scripts/run_infer_310.sh
@@ -0,0 +1,145 @@
+#!/bin/bash
+# Copyright 2021 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+#
+if [[ $# -lt 4 || $# -gt 5 ]]; then
+    echo "Usage: bash run_infer_310.sh [MINDIR_PATH] [DATASET_NAME] [DATASET_PATH] [NEED_PREPROCESS] [DEVICE_ID]
+    DATASET_NAME can choose from ['cifar10', 'imagenet2012'].
+    NEED_PREPROCESS means weather need preprocess or not, it's value is 'y' or 'n'.
+    DEVICE_ID is optional, it can be set by environment variable device_id, otherwise the value is zero"
+exit 1
+fi
+
+get_real_path(){
+    if [ "${1:0:1}" == "/" ]; then
+        echo "$1"
+    else
+        echo "$(realpath -m $PWD/$1)"
+    fi
+}
+model=$(get_real_path $1)
+if [ $2 == 'cifar10' ] || [ $2 == 'imagenet2012' ]; then
+  dataset_name=$2
+else
+  echo "DATASET_NAME can choose from ['cifar10', 'imagenet2012']"
+  exit 1
+fi
+
+dataset_path=$(get_real_path $3)
+
+if [ "$4" == "y" ] || [ "$4" == "n" ];then
+    need_preprocess=$4
+else
+  echo "weather need preprocess or not, it's value must be in [y, n]"
+  exit 1
+fi
+
+device_id=0
+if [ $# == 5 ]; then
+    device_id=$5
+fi
+
+echo "mindir name: "$model
+echo "dataset name: "$dataset_name
+echo "dataset path: "$dataset_path
+echo "need preprocess: "$need_preprocess
+echo "device id: "$device_id
+
+export ASCEND_HOME=/usr/local/Ascend/
+if [ -d ${ASCEND_HOME}/ascend-toolkit ]; then
+    export PATH=$ASCEND_HOME/fwkacllib/bin:$ASCEND_HOME/fwkacllib/ccec_compiler/bin:$ASCEND_HOME/ascend-toolkit/latest/fwkacllib/ccec_compiler/bin:$ASCEND_HOME/ascend-toolkit/latest/atc/bin:$PATH
+    export LD_LIBRARY_PATH=$ASCEND_HOME/fwkacllib/lib64:/usr/local/lib:$ASCEND_HOME/ascend-toolkit/latest/atc/lib64:$ASCEND_HOME/ascend-toolkit/latest/fwkacllib/lib64:$ASCEND_HOME/driver/lib64:$ASCEND_HOME/add-ons:$LD_LIBRARY_PATH
+    export TBE_IMPL_PATH=$ASCEND_HOME/ascend-toolkit/latest/opp/op_impl/built-in/ai_core/tbe
+    export PYTHONPATH=$ASCEND_HOME/fwkacllib/python/site-packages:${TBE_IMPL_PATH}:$ASCEND_HOME/ascend-toolkit/latest/fwkacllib/python/site-packages:$PYTHONPATH
+    export ASCEND_OPP_PATH=$ASCEND_HOME/ascend-toolkit/latest/opp
+else
+    export PATH=$ASCEND_HOME/fwkacllib/bin:$ASCEND_HOME/fwkacllib/ccec_compiler/bin:$ASCEND_HOME/atc/ccec_compiler/bin:$ASCEND_HOME/atc/bin:$PATH
+    export LD_LIBRARY_PATH=$ASCEND_HOME/fwkacllib/lib64:/usr/local/lib:$ASCEND_HOME/atc/lib64:$ASCEND_HOME/acllib/lib64:$ASCEND_HOME/driver/lib64:$ASCEND_HOME/add-ons:$LD_LIBRARY_PATH
+    export PYTHONPATH=$ASCEND_HOME/fwkacllib/python/site-packages:$ASCEND_HOME/atc/python/site-packages:$PYTHONPATH
+    export ASCEND_OPP_PATH=$ASCEND_HOME/opp
+fi
+export ASCEND_HOME=/usr/local/Ascend
+
+export PATH=$ASCEND_HOME/fwkacllib/ccec_compiler/bin:$ASCEND_HOME/fwkacllib/bin:$ASCEND_HOME/toolkit/bin:$PATH
+
+export LD_LIBRARY_PATH=/usr/local/lib/:/usr/local/fwkacllib/lib64:$ASCEND_HOME/driver/lib64:$ASCEND_HOME/add-ons:/usr/local/Ascend/toolkit/lib64:$LD_LIBRARY_PATH
+
+export PYTHONPATH=$ASCEND_HOME/fwkacllib/python/site-packages
+
+export PATH=/usr/local/python375/bin:$PATH
+export NPU_HOST_LIB=/usr/local/Ascend/acllib/lib64/stub
+export ASCEND_OPP_PATH=/usr/local/Ascend/opp
+export ASCEND_AICPU_PATH=/usr/local/Ascend
+export LD_LIBRARY_PATH=/usr/local/lib64/:$LD_LIBRARY_PATH
+function preprocess_data()
+{
+    if [ -d preprocess_Result ]; then
+        rm -rf ./preprocess_Result
+    fi
+    mkdir preprocess_Result
+    python3.7 ../preprocess.py --data_path=$dataset_path #--result_path=./preprocess_Result/
+}
+
+function compile_app()
+{
+    cd ../ascend310_infer/ || exit
+    bash build.sh &> build.log
+}
+
+function infer()
+{
+    cd - || exit
+    if [ -d result_Files ]; then
+        rm -rf ./result_Files
+    fi
+    if [ -d time_Result ]; then
+        rm -rf ./time_Result
+    fi
+    mkdir result_Files
+    mkdir time_Result
+
+        ../ascend310_infer/out/main --mindir_path=$model --dataset_name=$dataset_name --input0_path=../scripts/preprocess_Result/00_data --device_id=$device_id  &> infer.log
+}
+
+function cal_acc()
+{
+    python3.7 ../postprocess.py  &> acc.log
+}
+
+if [ $need_preprocess == "y" ]; then
+    preprocess_data
+    if [ $? -ne 0 ]; then
+        echo "preprocess dataset failed"
+        exit 1
+    fi
+fi
+compile_app
+if [ $? -ne 0 ]; then
+    echo "compile app code failed"
+    exit 1
+fi
+echo "compile success"
+infer
+if [ $? -ne 0 ]; then
+    echo " execute inference failed"
+    exit 1
+fi
+echo "infer success"
+cal_acc
+if [ $? -ne 0 ]; then
+    echo "calculate accuracy failed"
+    exit 1
+fi
+echo "cal_acc success"
diff --git a/research/cv/sknet/src/dataset.py b/research/cv/sknet/src/dataset.py
index a76103daf2bbb58dade9542c9b8b81e3e987fc34..ee19694bc88afd274ff264bd0e99157f6ed4f436 100644
--- a/research/cv/sknet/src/dataset.py
+++ b/research/cv/sknet/src/dataset.py
@@ -22,8 +22,8 @@ import mindspore.dataset.vision.c_transforms as C
 import mindspore.dataset.transforms.c_transforms as C2
 from mindspore.communication.management import init, get_rank, get_group_size
 
-def create_dataset1(dataset_path, do_train, repeat_num=1, batch_size=32, target="Ascend", distribute=False,
-                    enable_cache=False, cache_session_id=None):
+def create_dataset_cifar10(dataset_path, do_train=False, repeat_num=1, batch_size=32, target="Ascend", distribute=False,
+                           enable_cache=False, cache_session_id=None):
     """
     create a train or evaluate cifar10 dataset for sknet50
     Args: