OpenVINO模型优化器参考：https://docs.openvinotoolkit.org/2021.2/openvino_docs_MO_DG_Deep_Learning_Model_Optimizer_DevGuide.html
配置环境
- Windows 10
- Anaconda 2.0.4
- Pycharm 2021.1
- OpenVINO 工具包 2021.2
- Python 环境 3.6.13
- torch 1.9.0
- onnx 1.10.1
准备好torch模型
自定义神经网络模型pth转ONNX代码
import torch
from torch import nn
# 搭建神经网络
class VVcatModel(nn.Module):
def __init__(self):
super().__init__()
self.model = nn.Sequential(
nn.Conv2d(3, 32, 5, 1, 2),
nn.MaxPool2d(2),
nn.Conv2d(32, 32, 5, 1, 2),
nn.MaxPool2d(2),
nn.Conv2d(32, 64, 5, 1, 2),
nn.MaxPool2d(2),
nn.Flatten(),
nn.Linear(64 * 4 * 4, 64),
nn.Linear(64, 10)
)
def forward(self, x):
x = self.model(x)
return x
model = torch.load("./vvcat_model.pth", map_location=torch.device('cuda')) # 加载模型
model.eval()
model.cuda()
dummy_input = torch.randn(1, 3, 32, 32, device='cuda') # 注意指定输入尺寸
torch.onnx.export(model, dummy_input, "vvcat_model.onnx", output_names={"output"}, verbose=True)
通过执行代码生成了 .onnx 文件。
转IR格式
使用 OpenVINO 工具包提供的 mo_onnx.py文件，对模型进行转换。
管理员模式打开Anaconda，启动虚拟环境的终端，cd进openVINO转换工具目录，执行转换代码：
cd C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizerpython mo_onnx.py --input_model M:\python\深度学习\OpenVINO\15.模型转换器\Main\vvcat_model.onnx转换成功后，在转换工具 model_optimizer 目录下生成了bin和xml文件，然后就可以用 OpenVINO部署了。
- .xml - 描述网络拓扑
- .bin - 包含权重和偏差二进制数据。
(pytorch) C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizer>python mo_onnx.py --input_model M:\python\深度学习\OpenVINO\15.模型转换器\Main\vvcat_model.onnx
Model Optimizer arguments:
Common parameters:
- Path to the Input Model: M:\python\深度学习\OpenVINO\15.模型转换器\Main\vvcat_model.onnx
- Path for generated IR: C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizer\.
- IR output name: vvcat_model
- Log level: ERROR
- Batch: Not specified, inherited from the model
- Input layers: Not specified, inherited from the model
- Output layers: Not specified, inherited from the model
- Input shapes: Not specified, inherited from the model
- Mean values: Not specified
- Scale values: Not specified
- Scale factor: Not specified
- Precision of IR: FP32
- Enable fusing: True
- Enable grouped convolutions fusing: True
- Move mean values to preprocess section: None
- Reverse input channels: False
ONNX specific parameters:
Model Optimizer version: 2021.2.0-1877-176bdf51370-releases/2021/2
[ SUCCESS ] Generated IR version 10 model.
[ SUCCESS ] XML file: C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizer\.\vvcat_model.xml
[ SUCCESS ] BIN file: C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizer\.\vvcat_model.bin
[ SUCCESS ] Total execution time: 2.72 seconds.
It's been a while, check for a new version of Intel(R) Distribution of OpenVINO(TM) 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*
以下为生成的为我们需要的文件。
踩坑一：
没有安装 onnx、onnxruntime库导致的错误
[ ERROR ]
Detected not satisfied dependencies:
onnx: not installed, required: >= 1.1.2
Please install required versions of components or use install_prerequisites script
C:\Program Files (x86)\Intel\openvino_2021.2.185\deployment_tools\model_optimizer\install_prerequisites\install_prerequisites_onnx.bat
Note that install_prerequisites scripts may install additional components.
[ ERROR ] check_requirements exit with return code 1
安装 onnx、onnxruntime库即可解决
pip install onnx
pip install onnxruntime
踩坑二：
因为权限问题，导致无法转换IR文件。
解决方法：管理员模式打开Anaconda，启动虚拟环境的终端，cd进openVINO转换工具目录，再执行转换代码
