mobilenetv2 310infer amend

model_zoo/official/cv/mobilenetv2/README.md

@@ -252,9 +252,10 @@ Current batch_size can only be set to 1.
 
 ```shell
 # Ascend310 inference
-bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DVPP] [DEVICE_ID]
+bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [LABEL_PATH] [DVPP] [DEVICE_ID]
 ```
 
+- `LABEL_PATH` label.txt path. Write a py script to sort the category under the dataset, map the file names under the categories and category sort values,Such as[file name : sort value], and write the mapping results to the labe.txt file.
 - `DVPP` is mandatory, and must choose from ["DVPP", "CPU"], it's case-insensitive.The size of the picture that MobilenetV2 performs inference is [224, 224], the DVPP hardware limits the width of divisible by 16, and the height is divisible by 2. The network conforms to the standard, and the network can pre-process the image through DVPP.
 - `DEVICE_ID` is optional, default value is 0.
 

model_zoo/official/cv/mobilenetv2/README_CN.md

@@ -259,9 +259,10 @@ python export.py --platform [PLATFORM] --ckpt_file [CKPT_PATH] --file_format [EX
 
 ```shell
 # Ascend310 inference
-bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DVPP] [DEVICE_ID]
+bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [LABEL_PATH] [DVPP] [DEVICE_ID]
 ```
 
+- `LABEL_PATH` label.txt存放的路径，写一个py脚本对数据集下的类别名进行排序，对类别下的文件名和类别排序值做映射，例如[文件名:排序值]，将映射结果写到labe.txt文件中。
 - `DVPP` 为必填项，需要在["DVPP", "CPU"]选择，大小写均可。Mobilenetv2执行推理的图片尺寸为[224, 224]，DVPP硬件限制宽为16整除，高为2整除，网络符合标准，网络可以通过DVPP对图像进行前处理。
 - `DEVICE_ID` 可选，默认值为0。
 

model_zoo/official/cv/mobilenetv2/ascend310_infer/src/main.cc

@@ -117,11 +117,11 @@ int main(int argc, char **argv) {
       auto resizeShape = {FLAGS_image_height, FLAGS_image_width};
       std::shared_ptr<TensorTransform> resize(new Resize(resizeShape));
       auto crop_size = {224, 224};
-      std::shared_ptr<TensorTransform> center_crop(new CenterCrop(center_crop));
+      std::shared_ptr<TensorTransform> center_crop(new CenterCrop(crop_size));
       Execute transform({decode, resize, center_crop, normalize, hwc2chw});
       auto img = MSTensor();
       auto image = ReadFileToTensor(all_files[i]);
-      composeDecode(image, &img);
+      transform(image, &img);
       std::vector<MSTensor> model_inputs = model.GetInputs();
       inputs.emplace_back(model_inputs[0].Name(), model_inputs[0].DataType(), model_inputs[0].Shape(),
                        img.Data().get(), img.DataSize());

model_zoo/official/cv/mobilenetv2/postprocess.py

@@ -28,24 +28,30 @@ def calcul_acc(labels, preds):
     return sum(1 for x, y in zip(labels, preds) if x == y) / len(labels)
 
 
+def read_label(label_path):
+    label_dict = {}
+    with open(label_path, 'r') as f:
+        lines = f.readlines()
+    for line in lines:
+        file_name = line.split(':')[0]
+        label = line.split(':')[1]
+        label_dict[file_name] = label
+    return label_dict
+
+
 def get_result(result_path, label_path):
     files = os.listdir(result_path)
     preds = []
     labels = []
-    label_dict = {}
-    with open(label_path, 'w') as f:
-        lines = f.readlines()
-        for line in lines:
-            label_dict[line.split(',')[0]] = line.split(',')[1]
+    label_dict = read_label(label_path)
     for file in files:
         file_name = file.split('.')[0]
-        label = int(label_dict[file_name + '.JEPG'])
+        label = int(label_dict[file_name])
         labels.append(label)
-        resultPath = os.path.join(result_path, file)
-        output = np.fromfile(resultPath, dtype=np.float32)
+        output = np.fromfile(os.path.join(result_path, file), dtype=np.float32)
         preds.append(np.argmax(output, axis=0))
     acc = calcul_acc(labels, preds)
-    print("accuracy: {}".format(acc))
+    print("total{}, accuracy: {}".format(len(labels), acc))
 
 
 if __name__ == '__main__':

model_zoo/official/cv/mobilenetv2/scripts/run_infer_310.sh

@@ -14,8 +14,8 @@
 # limitations under the License.
 # ============================================================================
 
-if [[ $# -lt 3 || $# -gt 4 ]]; then
-    echo "Usage: bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DVPP] [DEVICE_ID]
+if [[ $# -lt 4 || $# -gt 5 ]]; then
+    echo "Usage: bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [LABEL_PATH] [DVPP] [DEVICE_ID]
     DVPP is mandatory, and must choose from [DVPP|CPU], it's case-insensitive
     DEVICE_ID is optional, it can be set by environment variable device_id, otherwise the value is zero"
 exit 1
@@ -30,15 +30,17 @@ get_real_path(){
 }
 model=$(get_real_path $1)
 data_path=$(get_real_path $2)
-DVPP=${3^^}
+label_path=$(get_real_path $3)
+DVPP=${4^^}
 
 device_id=0
-if [ $# == 4 ]; then
-    device_id=$4
+if [ $# == 5 ]; then
+    device_id=$5
 fi
 
 echo "mindir name: "$model
 echo "dataset path: "$data_path
+echo "label path: "$label_path
 echo "image process mode: "$DVPP
 echo "device id: "$device_id
 
@@ -85,7 +87,7 @@ function infer()
 
 function cal_acc()
 {
-    python3.7 ../postprocess.py --result_path=./result_Files --label_path=../label.txt &> acc.log &
+    python3.7 ../postprocess.py --result_path=./result_Files --label_path=$label_path &> acc.log &
 }
 
 compile_app