OpenVINO 部署 YOLOv5 转换IR文件

Posted 2021-09-08 ʚVVcatɞ

环境：

• Windows：10
• YOLOv5：3.1
• Python 3.7.10
  • torch：1.7.0+cu101
  • torchvision：0.8.1+cu101
• OpenVINO：openvino_2021.2.185
• Anaconda：2.0.4

运行以下脚本临时设置 OpenVINO 环境和变量：

cd C:\\Program Files (x86)\\Intel\\openvino_2021.2.185\\bin

运行以下命令来生成YOLOv5模型的IR：

python mo.py  --input_model M:\\yolov5-3.1\\yolov5-v3\\best.onnx --model_name M:\\yolov5-3.1\\best -s 255 --reverse_input_channels --output Conv_487,Conv_471,Conv_455

注：mo.py文件在 \\openvino_[版本号]\\deployment_tools\\model_optimizer 路径下

• –input_model 定义了预训练的模型
• –model_name 是生成的 IR 和输出 .xml/.bin 文件中的网络名称
• -s 表示来自原始网络输入的所有输入值将除以 this value，
• –reverse_input_channels 用于将输入通道顺序从RGB 切换到BGR（反之亦然）
• –output 表示模型的输出操作的名称。

(pytorch) C:\\Program Files (x86)\\Intel\\openvino_2021.2.185\\deployment_tools\\model_optimizer>python mo.py --input_model M:\\yolov5-3.1\\yolov5-v3\\best.onnx --model_name M:\\yolov5-3.1\\best -s 255 --reverse_input_channels --output Conv_487,Conv_471,Conv_455
Model Optimizer arguments:
Common parameters:
- Path to the Input Model: M:\\yolov5-3.1\\yolov5-v3\\best.onnx
- Path for generated IR: C:\\Program Files (x86)\\Intel\\openvino_2021.2.185\\deployment_tools\\model_optimizer.
- IR output name: M:\\yolov5-3.1\\best
- Log level: ERROR
- Batch: Not specified, inherited from the model
- Input layers: Not specified, inherited from the model
- Output layers: Conv_487,Conv_471,Conv_455
- Input shapes: Not specified, inherited from the model
- Mean values: Not specified
- Scale values: Not specified
- Scale factor: 255.0
- Precision of IR: FP32
- Enable fusing: True
- Enable grouped convolutions fusing: True
- Move mean values to preprocess section: None
- Reverse input channels: True
ONNX specific parameters:
Model Optimizer version: 2021.2.0-1877-176bdf51370-releases/2021/2
[ SUCCESS ] Generated IR version 10 model.
[ SUCCESS ] XML file: M:\\yolov5-3.1\\best.xml
[ SUCCESS ] BIN file: M:\\yolov5-3.1\\best.bin
[ SUCCESS ] Total execution time: 37.53 seconds.
It’s been a while, check for a new version of Intel® Distribution of OpenVINO™ toolkit here https://software.intel.com/content/www/us/en/develop/tools/openvino-toolkit/choose-download.html?cid=other&source=Prod&campid=ww_2021_bu_IOTG&content=upg_pro&medium=organic_uid_agjj or on the GitHub*

对转换的IR文件使用OpenVINO 2021.2 进行测试

python yolov5_OV2021.2.py -i img.jpg -m best.xml

yolov5_OV2021.2.py 文件 部分代码

#!/usr/bin/env python
from __future__ import print_function, division

import logging
import os
import sys
from argparse import ArgumentParser, SUPPRESS
from math import exp as exp
from time import time
import numpy as np

import cv2
from openvino.inference_engine import IENetwork, IECore

logging.basicConfig(format="[ %(levelname)s ] %(message)s", level=logging.INFO, stream=sys.stdout)
log = logging.getLogger()


def build_argparser():
    parser = ArgumentParser(add_help=False)
    args = parser.add_argument_group('Options')
    args.add_argument('-h', '--help', action='help', default=SUPPRESS, help='Show this help message and exit.')
    args.add_argument("-m", "--model", help="Required. Path to an .xml file with a trained model.",
                      required=True, type=str)
    args.add_argument("-i", "--input", help="Required. Path to an image/video file. (Specify 'cam' to work with "
                                            "camera)", required=True, type=str)
    args.add_argument("-l", "--cpu_extension",
                      help="Optional. Required for CPU custom layers. Absolute path to a shared library with "
                           "the kernels implementations.", type=str, default=None)
    args.add_argument("-d", "--device",
                      help="Optional. Specify the target device to infer on; CPU, GPU, FPGA, HDDL or MYRIAD is"
                           " acceptable. The sample will look for a suitable plugin for device specified. "
                           "Default value is CPU", default="CPU", type=str)
    args.add_argument("--labels", help="Optional. Labels mapping file", default=None, type=str)
    args.add_argument("-t", "--prob_threshold", help="Optional. Probability threshold for detections filtering",
                      default=0.5, type=float)
    args.add_argument("-iout", "--iou_threshold", help="Optional. Intersection over union threshold for overlapping "
                                                       "detections filtering", default=0.4, type=float)
    args.add_argument("-ni", "--number_iter", help="Optional. Number of inference iterations", default=1, type=int)
    args.add_argument("-pc", "--perf_counts", help="Optional. Report performance counters", default=False,
                      action="store_true")
    args.add_argument("-r", "--raw_output_message", help="Optional. Output inference results raw values showing",
                      default=False, action="store_true")
    args.add_argument("--no_show", help="Optional. Don't show output", action='store_true')
    return parser

如果需要将yolov5s.pt转换为IR文件使用OpenVINO 部署以下是yolov5s.pt的下载链接：
https://github.com/ultralytics/yolov5/releases/download/v3.0/yolov5s.pt

自定义训练生成权重文件：

python train.py --img 640 --batch-size 2 --epoch 5 --data ./hand_data/hand.yaml --cfg ./models/yolov5s.yaml --weights '' --workers 0 --nosave --cache --device cpu

• –img：输入图片分辨率大小，nargs=’+'表示参数可设置一个或多个
• –batch-size：批次大小，一次训练所选取的样本数，显卡不行的话，就调小点
• –epoch：训练总轮次，1个epoch等于使用训练集中的全部样本训练一次，值越大模型越精确，训练时间也越长。
• –data：数据集配置文件，数据集路径，类名等，使用数据集方面的coco.yaml文件
• –cfg：模型配置文件，网络结构，使用修改好的yolov5m.yaml文件
• –weights：选用训练的权重，可用根目录下的yolov5s.pt，也可用runs/train/exp/weights/best.pt，不选择的话，从头开始训练。
• –workers：dataloader的最大worker数量，建议为0
• –nosave：不保存模型，默认False
• –cache： 是否提前缓存图片到内存，以加快训练速度，默认False
• –device：训练的设备，cpu；0(表示一个gpu设备cuda:0)；0,1,2,3(多个gpu设备)。值为空时，训练时默认使用 计算机自带的显卡或CPU

以下为测试自定义训练生成权重文件使用

hand.yaml 文件

train: ../hand_data/images/
val: ../hand_data/images/

nc: 1
names: ['1']

数据集标定工具：LabelImg

下载：https://github.com/tzutalin/labelImg