Skip to content

Latest commit

 

History

History
97 lines (74 loc) · 3.88 KB

README.md

File metadata and controls

97 lines (74 loc) · 3.88 KB
Raw

OpenLenda

License

OpenLenda is a Japanese traffic light detection and recognition model.

sample

🚥Introduction

OpenLenda is a Japanese traffic light recognition model based on YOLOX. This model not only detects traffic lights but can also recognize their colors and specific arrow indications in an end-to-end manner. For more information about the development, please visit https://zenn.dev/turing_motors/articles/traffic-light (Japanese article). This repository is for model publication.

🚙Getting Started

Installation

  1. Clone this repository.

    git clone https://github.com/turingmotors/openlenda.git
cd openlenda

  2. Install python dependencies. We strongly recommend using pyenv and poetry to manage python versions and dependencies.

    pyenv install 3.8.10
pyenv local 3.8.10
poetry install

  3. Download the pretrained weights.

    wget https://github.com/turingmotors/openlenda/releases/download/v0.1.0/openlenda_s.pth -P models

Demo

If you don't have a GPU, please remove --device gpu from the following commands.

python tools/demo.py video -f exps/openlenda_s.py\
    -c models/openlenda_s.pth\
    --path assets/tokyo_day.mp4\
    --conf 0.5 --nms 0.01 --tsize 640 --save_result --device gpu

Export ONNX

python tools/export_onnx.py --output-name openlenda_s.onnx\
    -f exps/openlenda_s.py\
    -c models/openlenda_s.pth\

📈Performance

AP and AR are evaluated on the private validation set. The results are shown below.

Model size AP50 AR50 Download
OpenLenda-nano 416x416 0.941 0.830 link
OpenLenda-tiny 416x416 0.944 0.807 link
OpenLenda-s 640x640 0.964 0.946 link
OpenLenda-x 640x640 0.950 0.961 link

📚Documentation

Please see YOLOX's README or docs (almost same as YOLOX's).

🤔Q & A

Check Q and A.

About Datasets

Is the dataset publicly available?

Sorry, but the dataset is private.

How many images are there in the dataset?

About 44,000 images.

About Training and Evaluation

Can I use this repository for learning and evaluation?

This is possible under the Apache 2.0 license. However, at this time, it is not possible to use the code in this repository as-is for training and evaluation, so you will need to modify the code yourself.

Orginization

Turing Inc.

📝License

This project is licensed under the Apache License 2.0. See LICENSE for more details.

🙏Acknowledgement

This project is based on YOLOX.