# caffe2onnx

**Repository Path**: beiming52/caffe2onnx

## Basic Information

- **Project Name**: caffe2onnx
- **Description**: caffe2onnx
- **Primary Language**: Unknown
- **License**: BSD-3-Clause
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2023-08-25
- **Last Updated**: 2023-08-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# caffe2onnx
This tool converts Caffe models to ONNX via command line (without Caffe environment).

## Introduction
**caffe2onnx** is based on previous work by [htshinichi](https://github.com/htshinichi/caffe-onnx).  
  
This is the second version of converting caffe model to onnx model. In this version, all the parameters will be transformed to tensor and tensor value info when reading **.caffemodel** file and each operator node is constructed directly into the type of NodeProto in **onnx**.

## Installation

Install latest from github  
```
pip install git+https://github.com/kumardesappan/caffe2onnx
```

Build and install latest from source  
```
git clone https://github.com/kumardesappan/caffe2onnx
python setup.py install
``` 

## Usage
To get started with **caffe2onnx**, run the *caffe2onnx.convert* command, providing:
* the path to your caffe prototxt,
* the path to your caffe model (*not required*),
* the output path of the onnx model (*not required*),
* frozen graph or not (*not required*).

```
python -m caffe2onnx.convert
    --prototxt          caffe's prototxt file path
    [--caffemodel       caffe's caffemodel file path]
    [--onnx             output onnx file path]
    [--frozen           frozen graph or not]
```

## Operators
* BatchNorm
* Convolution
* Deconvolution
* Concat
* Dropout
* InnerProduct (Reshape + Gemm)
* LRN
* Pooling
* Unpooling
* ReLU
* Softmax
* Eltwise
* Upsample
* Scale

## Models
* ResNet
* AlexNet
* AgeNet
* Yolo V3
* VGG
* GoogleNet
* CaffeNet
* Sphereface

## License
[BSD-3](https://github.com/asiryan/caffe2onnx/blob/main/LICENSE)