https://github.com/tjyuyao/bevutils

Bird's Eye View Utils
https://github.com/tjyuyao/bevutils

Last synced: 15 days ago
JSON representation

Bird's Eye View Utils

• Host: GitHub
• URL: https://github.com/tjyuyao/bevutils
• Owner: tjyuyao
• License: mit
• Created: 2019-08-03T06:47:13.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2022-06-21T22:28:14.000Z (over 2 years ago)
• Last Synced: 2024-12-16T05:20:55.638Z (about 1 month ago)
• Language: Python
• Size: 1.68 MB
• Stars: 1
• Watchers: 0
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Bird's-Eye View Utilities

by Huang Yuyao, Wen YongKun

## Features

1. A differentiable Perspective Transformer Layer inspired by Spatial Transformer Networks. See `bevutils/layers/perspective_transformer.py` and `samples/basic/sample.py`.

2. A simulated dataset. See `bevutils/datasets/simulated.py` and `samples/simulated_dataset/demo_dataset.py`.

3. A training system inspired by [victoresque's template](https://github.com/victoresque/pytorch-template) and [mmdetection](https://github.com/open-mmlab/mmdetection/)

## Installation

`pip install git+https://github.com/huangyuyao/bevutils.git@master`

## Development

```

git clone https://github.com/huangyuyao/bevutils.git

cd bevutils

pip install -e .

```

## Usage

You can make use of functions, datasets and layers of this project by import the functionalities from anywhere.

If you want to use the capability of training a new network, you can do three things:

    - registry your custom network modules(TBD),

    - write a config file,

    - and call `bevtrain -c config.json` or `bevtest -c config.json` tools to reuse the training process. (TBD)

You can refer to `samples/perspective_transformer_network` as a quick tutorial. (TBD)

## License

This project is licensed under the MIT License. See LICENSE for more details