fix unet Ascend310 inference bug

model_zoo/official/cv/unet/README.md

@@ -474,19 +474,20 @@ the steps below, this is a simple example:
 
 Export MindIR
 
+Before exporting, you need to modify the parameter in the configuration — checkpoint_file_path and batch_ Size . checkpoint_ file_ Path is the CKPT file path, batch_ Size is set to 1.
+
 ```shell
-python export.py --checkpoint_file_path [CKPT_PATH] --file_name [FILE_NAME] --file_format [FILE_FORMAT]
+python export.py --config_path=[CONFIG_PATH]
 ```
 
 The checkpoint_file_path parameter is required,
 `EXPORT_FORMAT` should be in ["AIR", "MINDIR"]
 
 Before performing inference, the MINDIR file must be exported by export script on the 910 environment.
-Current batch_size can only be set to 1.
 
 ```shell
 # Ascend310 inference
-bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DEVICE_ID]
+bash run_infer_310.sh [NETWORK] [MINDIR_PATH] [DEVICE_ID] [NEED_PREPROCESS]
 ```
 
 `DEVICE_ID` is optional, default value is 0.

model_zoo/official/cv/unet/README_CN.md

@@ -471,18 +471,17 @@ python eval.py --data_path=/path/to/data/ --checkpoint_file_path=/path/to/checkp
 
 导出mindir模型
 
+在执行导出前需要修改配置文件中的checkpoint_file_path和batch_size参数。checkpoint_file_path为ckpt文件路径，batch_size设置为1。
+
 ```shell
-python export.py --checkpoint_file_path [CKPT_PATH] --file_name [FILE_NAME] --file_format [FILE_FORMAT]
+python export.py --config_path=[CONFIG_PATH]
 ```
 
-参数`checkpoint_file_path` 是必需的，`EXPORT_FORMAT` 必须在 ["AIR", "MINDIR"]中进行选择。
-
 在执行推理前，MINDIR文件必须在910上通过export.py文件导出。
-目前仅可处理batch_Size为1。
 
 ```shell
 # Ascend310 推理
-bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DEVICE_ID]
+bash run_infer_310.sh [NETWORK] [MINDIR_PATH] [DEVICE_ID] [NEED_PREPROCESS]
 ```
 
 `DEVICE_ID` 可选，默认值为 0。

model_zoo/official/cv/unet/postprocess.py

@@ -77,7 +77,7 @@ if __name__ == '__main__':
     rst_path = config.rst_path
     metrics = dice_coeff()
 
-    if config.dataset == "Cell_nuclei":
+    if hasattr(config, "dataset") and config.dataset == "Cell_nuclei":
         img_size = tuple(config.image_size)
         for i, bin_name in enumerate(os.listdir('./preprocess_Result/')):
             f = bin_name.replace(".png", "")

model_zoo/official/cv/unet/preprocess.py

@@ -29,7 +29,7 @@ def preprocess_dataset(data_dir, result_path, cross_valid_ind=1):
     labels_list = []
     for i, data in enumerate(valid_dataset):
         file_name = "ISBI_test_bs_1_" + str(i) + ".bin"
-        file_path = result_path + file_name
+        file_path = os.path.join(result_path, file_name)
         data[0].asnumpy().tofile(file_path)
 
         labels_list.append(data[1].asnumpy())

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

@@ -15,7 +15,7 @@
 # ============================================================================
 
 if [[ $# -lt 3 || $# -gt 4 ]]; then
-    echo "Usage: bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [DEVICE_ID] [NEED_PREPROCESS]
+    echo "Usage: bash run_infer_310.sh [NETWORK] [MINDIR_PATH] [DEVICE_ID] [NEED_PREPROCESS]
     DEVICE_ID is optional, it can be set by environment variable device_id, otherwise the value is zero.
     NEED_PREPROCESS means weather need preprocess or not, it's value is 'y' or 'n'."
 exit 1
@@ -28,8 +28,9 @@ get_real_path(){
     echo "$(realpath -m $PWD/$1)"
   fi
 }
-model=$(get_real_path $1)
-data_path=$(get_real_path $2)
+
+network=$1
+model=$(get_real_path $2)
 if [ $# == 4 ]; then
     device_id=$3
     if [ -z $device_id ]; then
@@ -40,8 +41,8 @@ if [ $# == 4 ]; then
 fi
 need_preprocess=$4
 
+echo "network: " $network
 echo "mindir name: "$model
-echo "dataset path: "$data_path
 echo "device id: "$device_id
 echo "need preprocess or not: "$need_preprocess
 
@@ -65,7 +66,17 @@ function preprocess_data()
         rm -rf ./preprocess_Result
     fi
     mkdir preprocess_Result
-    python3.7 ../preprocess.py --data_url=$data_path --result_path=./preprocess_Result/
+
+    BASEPATH=$(cd "`dirname $0`" || exit; pwd)
+    if [ $network == "unet" ]; then
+        config_path="${BASEPATH}/../unet_simple_config.yaml"
+    elif [ $network == "unet++" ]; then
+        config_path="${BASEPATH}/../unet_nested_cell_config.yaml"
+    else
+        echo "unsupported network"
+        exit 1
+    fi
+    python3.7 ../preprocess.py --config_path=$config_path
 }
 
 function compile_app()
@@ -93,7 +104,13 @@ function infer()
 
 function cal_acc()
 {
-    python3.7 ../postprocess.py --data_url=$data_path --rst_path=./result_Files/ &> acc.log &
+    BASEPATH=$(cd "`dirname $0`" || exit; pwd)
+    if [ $network == "unet" ]; then
+        config_path="${BASEPATH}/../unet_simple_config.yaml"
+    elif [ $network == "unet++" ]; then
+        config_path="${BASEPATH}/../unet_nested_cell_config.yaml"
+    fi
+    python3.7 ../postprocess.py --config_path=$config_path &> acc.log &
 }
 
 preprocess_data

model_zoo/official/cv/unet/unet_nested_cell_config.yaml

@@ -45,13 +45,13 @@ eval_resize: False
 checkpoint_path: './checkpoint/'
 checkpoint_file_path: 'ckpt_unet_nested_adam-4-75.ckpt'
 rst_path: './result_Files/'
-result_path: ""
+result_path: "./preprocess_Result"
 
 # Export options
-width: 572
-height: 572
-file_name: "unet"
-file_format: "AIR"
+width: 96
+height: 96
+file_name: "unetplusplus"
+file_format: "MINDIR"
 
 ---
 # Help description for each configuration

model_zoo/official/cv/unet/unet_simple_config.yaml

@@ -41,13 +41,13 @@ eval_resize: False
 checkpoint_path: './checkpoint/'
 checkpoint_file_path: 'ckpt_unet_simple_adam-4-75.ckpt'
 rst_path: './result_Files/'
-result_path: ""
+result_path: "./preprocess_Result"
 
 # Export options
-width: 572
-height: 572
+width: 576
+height: 576
 file_name: "unet"
-file_format: "AIR"
+file_format: "MINDIR"
 
 ---
 # Help description for each configuration