运行记录

1.建立Mindspore运行环境

1。Mindspore的环境安装指南，按照提示进行https://www.mindspore.cn/install
2.下载Mindspore的yolov5，并下载模型和代码https://www.hiascend.com/zh/software/modelzoo/detail/C/9af06eedaa1b9577d6221a939a31747d

下载这两个文件

3。安装库，cd到模型脚本目录下，运行pip install -r requirements.txt

2.下载运行数据集MSCOCO

什么是COCO数据集？

COCO数据集是一个可用于图像检测（image detection），语义分割（semantic segmentation）和图像标题生成（image captioning）的大规模数据集。它有超过330K张图像（其中220K张是有标注的图像），包含150万个目标，80个目标类别（object categories：行人、汽车、大象等），91种材料类别（stuff categoris：草、墙、天空等），每张图像包含五句图像的语句描述，且有250,000个带关键点标注的行人。
具体的COCO数据集介绍 https://blog.csdn.net/qq_44554428/article/details/122597358

COCO数据集的下载

下载网址：官网下载链接
下载最小的val2017和annotions，放到目录下的datasets文件下

3.根据readme文件提示，运行export.py文件，导出mindir模型文件。

cd scripts
python export.py --ckpt_file [CKPT_PATH] --file_format [EXPORT_FORMAT]
#其中ckpt为ckpt模型文件地址，`EXPORT_FORMAT` 必须在 ["AIR", "MINDIR"]中选择。`BATCH_SIZE` 目前仅支持batch_size为1的推理（不用设置）。

运行后会再主目录下得到yolov5.mindir的模型文件。

4. 在执行推理前，mindir文件必须通过export.py脚本导出。以下展示了使用mindir模型执行推理的示例。

运行指令：bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [ANN_FILE] [DVPP] [DEVICE_ID]
阅读readme与sh文件提示，其中 mindir_path 为模型地址；DATApath为MS COCO val2017的images地址；ANN_FILE为标注地址；DVPP为推理设备，设置CPU，DEVICE_ID默认为0，无需设置。
如：

在cmd命令下运行export。py文件，后面加上参数

python export.py --ckpt_file D:\yanyi\project_process\ocr\code\Yolov5_for_MindSpore_1.1_code\weights\yolov5.ckpt --file_format MINDIR

5.运行推理

在执行推理前，mindir文件必须通过export.py脚本导出。以下展示了使用mindir模型执行推理的示例。

bash run_infer_310.sh [MINDIR_PATH] [DATA_PATH] [ANN_FILE] [DVPP] [DEVICE_ID]

在cmd下运行。sh文件，后面加参数

cd scripts 
run_infer_310.sh D:\yanyi\project_process\ocr\code\Yolov5_for_MindSpore_1.1_code\yolov5.mindir D:\yanyi\project_process\ocr\code\Yolov5_for_MindSpore_1.1_code\dataset\val2017 D:\yanyi\project_process\ocr\code\Yolov5_for_MindSpore_1.1_code\dataset\annotations\panoptic_val2017.json CPU

运行run_infer_310.sh 心得

不熟悉shell语言的可以加echo语句输出，查看运行到哪。

#!/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] [DATA_PATH] [ANN_FILE] [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
fi

get_real_path(){
    
    if [ "${1:0:1}" == "/" ]; then
        echo "$1"
    else
        echo "$(realpath -m $PWD/$1)"
    fi
}
model=$(get_real_path $1)
data_path=$(get_real_path $2)
ann_file=$(get_real_path $3)
DVPP=${4^^}

device_id=0
if [ $# == 5 ]; then
    device_id=$5
fi

echo "mindir name: "$model
echo "dataset path: "$data_path
echo "ann file: "$ann_file
echo "image process mode: "$DVPP
echo "device id: "$device_id

export ASCEND_HOME=/usr/local/Ascend/
if [ -d ${ASCEND_HOME}/ascend-toolkit ]; then
    echo "device id: "$device_id
    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:$PATH
    echo "device id: "$device_id
    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
    echo "device DDDDDDDDDDDDDDDDDDDid: "$device_id
    export PATH=$ASCEND_HOME/fwkacllib/bin:$ASCEND_HOME/fwkacllib/ccec_compiler/bin:$ASCEND_HOME/atc/ccec_compiler/bin:$ASCEND_HOME/atc/bin:$PATH
    echo "deviceDDDDDDDDDDDDDDDDDDDDD id: "$device_id
    export LD_LIBRARY_PATH=$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=$ASCEND_HOME/opp
fi
echo "device AAAAAAAAAAAAAAAid: "$device_id
function compile_app()
{
    
    cd ../ascend310_infer || exit
    bash build.sh &> build.log
}
echo "device BBBBBBBBBBBBBid: "$device_id
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
    if [ "$DVPP" == "DVPP" ];then
      echo "Only support CPU mode"
      exit 1
    elif [ "$DVPP" == "CPU"  ]; then
      ../ascend310_infer/out/main --mindir_path=$model --dataset_path=$data_path --device_id=$device_id --image_height=640 --image_width=640 &> infer.log
      echo "device CCCCCCCCCCCCCCid: "$device_id
    else
      echo "image process mode must be in [DVPP|CPU]"
      exit 1
    fi
}
echo "device jisuanjingduid: "$device_id
function cal_acc()
{
    
    echo "device DDDDDDDDDDDDid: "$device_id
    python3.7 ../postprocess.py --result_files=./result_Files --dataset_path=$data_path --ann_file=$ann_file &> acc.log &
    echo "AAAAAAAAAAAA"
}
echo "device compile_app id: "$device_id
compile_app
if [ $? -ne 0 ]; then
    echo "compile app code failed"
    exit 1
fi
echo "device infer id: "$device_id
infer
if [ $? -ne 0 ]; then
    echo " execute inference failed"
    exit 1
fi
echo "device cal_acc id: "$device_id
cal_acc
if [ $? -ne 0 ]; then
    echo "calculate accuracy failed"
    exit 1
fi

运行后会进入编译阶段，这里还没有结果

----------------------------------------目前只有这些-----------------------------------------------------

6运行eval文件

下载MS coco数据集val2017和instances_val2017.json，按照如下排列：
instances_val2017.json放入annotations文件中，val2017放到yolov5目录下，其他不需要设置。
运行命令：

python eval.py --data_dir=./dataset/yolov5/ --pretrained=./weights/yolov5.ckpt --testing_shape=640

运行成功后会出现

2022.7.1

尝试在win系统下运行，太难了最终还是没起来，分享一下可借鉴的推理代码

https://github.com/mindspore-ai/mindspore-21-days-tutorials/tree/main/chapter4