# open-exchange

**Repository Path**: pkumlg/open-exchange

## Basic Information

- **Project Name**: open-exchange
- **Description**: open-exchange is a model conversion and visualization tool to help users inter-operate among different deep learning frameworks. Convert models between PyTorch and Tensorflow.
- **Primary Language**: Python
- **License**: MIT
- **Default Branch**: new
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-05-12
- **Last Updated**: 2021-05-12

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# open-exchange
open-exchange is a model conversion and visualization tool to help users inter-operate among different deep learning frameworks. Convert models between PyTorch and Tensorflow.
IR is based on the National Information Technology Standardization ```Neural Network Representation and Model Compression Part 1: Convolution Neural Network``.

## Requirments
- tensorflow==1.8.0
- pytorch==0.4.0
- torchvision==0.2.0
- protobuf>=3.6.1
- python>=3.6
- flask

## How to deploy visualization on Web
If you want to access the deployed web page from an external network, you should first modify the *host* and *port* for the *./visualization/app.py* file.

You can deploy it on Web by running:
```shell
python app.py
```

You can also choose to visit the <http://dgx.pkuml.org:6269/> website directly. After uploading the JSON file of IR, you can see the basic information about the model.

## Model
Framework | ResNet50 | Inception V3 | ShuffleNet | FCN | LSTM |
:----------:|:----------:|:----------:|:----------:|:----------:|:----------:|
TensorFlow | √ | √ | √ | √ | √ |
PyTorch | √ | √ | √ | √ | √ | 

## Test for Tensorflow and PyTorch
```shell
CUDA_VSIBLE_DEVICES=0 python test.py
```