ONNX 开始

环境

基础

• Anaconda

conda create -n onnx python=3.8 -y
conda activate onnx

# ONNX
#  https://github.com/onnx/onnx
conda install -c conda-forge onnx -y
python -c "import onnx; print(onnx.__version__)"

import onnx
model = onnx.load("model.onnx")

简化

# ONNX Simplifier
#  https://github.com/daquexian/onnx-simplifier
pip install onnx-simplifier
python -m onnxsim -h

import onnxsim
model_simp, check = onnxsim.simplify(model, perform_optimization=False)
assert check, "Simplified ONNX model could not be validated"

使用

给出些 ONNX 模型使用的示例方法。

提取子模型

import onnx

input_path = "path/to/the/original/model.onnx"
output_path = "path/to/save/the/extracted/model.onnx"
input_names = ["input_0", "input_1", "input_2"]
output_names = ["output_0", "output_1"]

onnx.utils.extract_model(input_path, output_path, input_names, output_names)

修改输入输出名称

def _onnx_rename(model, names, names_new):
  for node in model.graph.node:
    for i, n in enumerate(node.input):
      if n in names:
        node.input[i] = names_new[names.index(n