nanodet-deploy

This project is for deployment of nanodet-m onnx model, including onnxruntime, ncnn...

How to Use

Requirements

• CUDA = 11.3
• Python = 3.8
• Pytorch = 1.10.1
• torchvision==0.11.2

Env

1. Create a conda virtual environment and then activate it.

 conda create -n nanodet python=3.8 -y
 conda activate nanodet

2. Install pytorch

conda install pytorch==1.10.1 torchvision==0.11.2 cudatoolkit=11.3 -c pytorch -y

3. Clone this repository

git clone https://github.com/aaalvinnn/nanodet-deploy.git
cd nanodet-deploy

4. Install requirements

pip install -r requirements.txt

5. Setup NanoDet

python setup.py develop

Infer

onnxruntime

It is provided with 2 versions of inferring:

exam.

1. infer.py

python .\user\onnxruntime\infer.py image --config .\config\nanodet_custom_xml_dataset.yml --model .\models\nanodet.onnx --path .\user\images\image1.jpg --save_result

while --save_result is optional.

2. infer_yaml.py

   1. change config nanodet_custom_xml_dataset.yml;

   2. python .\user\onnxruntime\infer_yaml.py --config .\config\nanodet_custom_xml_dataset.yml
      

ncnn

Thanks

• https://github.com/RangiLyu/nanodet.git

Test

this is a test to test if available when a user in server trying to push and pull respository.