An open API service indexing awesome lists of open source software.

https://github.com/xlite-dev/netron-vscode-extension

☕️ A vscode extension for netron, support *.pdmodel, *.nb, *.onnx, *.pb, *.h5, *.tflite, *.pth, *.pt, *.mnn, *.param, etc.
https://github.com/xlite-dev/netron-vscode-extension

deeplearning mnn ncnn netron netron-vscode netron-vscode-extension onnx onnxruntime paddle tnn vscode-netron yolov5 yolov8

Last synced: 3 months ago
JSON representation

☕️ A vscode extension for netron, support *.pdmodel, *.nb, *.onnx, *.pb, *.h5, *.tflite, *.pth, *.pt, *.mnn, *.param, etc.

Awesome Lists containing this project

README

        

![](https://github.com/DefTruth/netron-vscode-extension/assets/31974251/8da27592-bafb-40fb-a4f3-d3ae0cda01b2)

**netron-vscode-extension**: A vscode extension for netron, support *.pdmodel, *.nb, *.onnx, *.pb, *.h5, *.tflite, *.pth, *.pt, *.mnn, *.tnnproto, *.param, etc. This code is based on [NetronInVSCode](https://github.com/chingweihsu0809/NetronInVSCode) but adds more model format support (such as *.pdmodel, *.nb, *.mnn, *.param, *.tnnproto, etc.) and some new [Features](#Features). **NOTE: This extension is mainly for personal use, so the development plans are based on individual needs and may not have intensive plans in the future.**

## Features

- [x] add *.pdmodel, *.nb, *.mnn, *.param, *.tnnproto format support
- [x] remove manually `pip install` requirements
- [ ] [WIP] onnx-simplifier support
- [ ] [WIP] paddle2onnx converter support
- [ ] [WIP] x2paddle converter support
- [ ] [WIP] MNN converter support
- [ ] [WIP] Optimize some UI components

## Install

- Install `netron-vscode-extension` from [vscode marketplace](https://marketplace.visualstudio.com/items?itemName=DefTruth.netron-vscode-extension).
![](https://github.com/DefTruth/netron-vscode-extension/assets/31974251/3139a461-42ff-48f7-9c8b-5920f25b6da5)

- Install `Python 3.x` on your computer. The step about netron installation follows is optional. This extension will call netron from python via [node-python-bride](https://github.com/Submersible/node-python-bridge) to form an url which contains the visualized contents, then, pass it to this extension. `netron-vscode-extension` will check whether the netron is avaliable on your python and auto install it if necessary.
```bash
python -m pip install netron # optional
```

## Usage

- Click on a saved model to see the following, such as [picodet](models/picodet_l_320_coco_lcnet/model.pdmodel), [yolov8n](models/yolov8n.onnx), etc.
![](https://github.com/DefTruth/netron-vscode-extension/assets/31974251/6acdfaa9-f56a-4f83-aa78-b85248a1e44b)
![](https://github.com/DefTruth/netron-vscode-extension/assets/31974251/9bbcbdb1-bfc7-4094-8f83-626f750f843d)

## Build

You can build the extension from sources:
- Open this repo in VS Code
- `npm install`
- `F5` to start debugging

## LICENSE

The codes is released under the MIT License.

## Reference

- [netron](https://github.com/lutzroeder/netron)
- [node-python-bride](https://github.com/Submersible/node-python-bridge)
- [vscode-extension-samples](https://github.com/microsoft/vscode-extension-samples)
- [NetronInVSCode](https://github.com/chingweihsu0809/NetronInVSCode)