annotations/Azaion.Inference/README.md

Azaion AI

Azaion AI is a worker written on cython (c-compilable python) which listens to socket and rabbit queue. It accepts commands om a format:

• CommandType: Inference / Load
• Filename

And correspondingly do inference or just load encrypted file from the API. Results (file or annotations) is putted to the other queue, or the same socket, depending on the command source.

Installation

Prepare correct onnx model from YOLO:

from ultralytics import YOLO
import netron

model = YOLO("azaion.pt")
model.export(format="onnx", imgsz=1280, nms=True, batch=4)
netron.start('azaion.onnx')

Read carefully about export arguments, you have to use nms=True, and batching with a proper batch size

Install libs

https://www.python.org/downloads/

Windows

• Install CUDA

Linux

    sudo apt install nvidia-driver-535
    
    wget https://developer.download.nvidia.com/compute/cudnn/9.2.0/local_installers/cudnn-local-repo-ubuntu2204-9.2.0_1.0-1_amd64.deb
    sudo dpkg -i cudnn-local-repo-ubuntu2204-9.2.0_1.0-1_amd64.deb
    
    sudo cp /var/cudnn-local-repo-ubuntu2204-9.2.0/cudnn-*-keyring.gpg /usr/share/keyrings/
    sudo apt-get update
    sudo apt-get -y install cudnn nvidia-cuda-toolkit -y
    nvcc --version

Install dependencies

1. Install python with max version 3.11. Pytorch for now supports 3.11 max

Make sure that your virtual env is installed with links to the global python packages and headers, like this:

    python -m venv --system-site-packages venv

This is crucial for the build because build needs Python.h header and other files.

    python -m pip install --upgrade pip
    pip install requirements.txt
    pip install pyinstaller

In case of fbgemm.dll error (Windows specific):

• copypaste libomp140.x86_64.dll to C:\Windows\System32

Build

``` python setup.py build_ext --inplace ```

Build exe

``` .\build.cmd ```