# Install Pytorch on Win11 WSL2

## [0-Install the GPU driver]
https://developer.nvidia.com/cuda/wsl
Download and install NVIDIA GeForce Game Ready or NVIDIA RTX Quadro Windows 11 display driver on your system with a compatible GeForce or NVIDIA RTX/Quadro card from
https://developer.nvidia.com/cuda/wsl

## [1-Install WSL2 & ...]

$ wsl --install -d Ubuntu-18.04
$ sudo add-apt-repository universe
$ sudo apt-get update
$ sudo apt-get install python3-pip
$ sudo pip3 install virtualenv virtualenvwrapper
$ sudo rm -rf ~/get-pip.py ~/.cache/pip

$ vi ~/.bashrc
add 4 lines at the bottom of file

# virtualenv and virtualenvwrapper
export WORKON_HOME=$HOME/.virtualenvs
export VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3
source /usr/local/bin/virtualenvwrapper.sh

$ source ~/.bashrc

## [2-Install Cuda]

wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/11.4.0/local_installers/cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-wsl-ubuntu-11-4-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

### nVidia Version

$ nvtop
$ nvidia-smi
$ nvidia-smi -q
$ nvidia-smi -l 1

$ cd /mnt/c/Users/'Nico Huang'/Desktop

## [3-Install Jetson inference files only]

$ sudo apt-get update
$ sudo apt-get install git cmake libpython3-dev python3-numpy
$ git clone --recursive https://github.com/dusty-nv/jetson-inference

## [4-Create pytorch virtual environment named pytorch]

$ mkvirtualenv pytorch -p python3
$ workon pytorch    // do everytime before training
# $ deactivate

## [5-Install PyTorch]
https://pytorch.org/get-started/locally/

### CUDA 11.0
https://pytorch.org/get-started/previous-versions/

pip3  install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html

Successfully installed:
- dataclasses-0.8
- numpy-1.19.5
- pillow-8.4.0
- torch-1.7.1+cu110
- torchaudio-0.7.2
- torchvision-0.8.2+cu110
- typing-extensions-4.0.0

### Verifying PyTorch

    (pytorch) nico@Nico-ZenBook15:~$ python3
Python 3.6.9 (default, Jan 26 2021, 15:33:00)
[GCC 8.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> print(torch.__version__)
1.7.1+cu110
>>> print('CUDA available: ' + str(torch.cuda.is_available()))
CUDA available: True
>>> a = torch.cuda.FloatTensor(2).zero_()
>>> print('Tensor a = ' + str(a))
Tensor a = tensor([0., 0.], device='cuda:0')
>>> b = torch.randn(2).cuda()
>>> print('Tensor b = ' + str(b))
Tensor b = tensor([-0.7677,  0.7887], device='cuda:0')
>>> c = a + b
>>> print('Tensor c = ' + str(c))
Tensor c = tensor([-0.7677,  0.7887], device='cuda:0')
>>>
>>> import torchvision
>>> print(torchvision.__version__)
0.8.2+cu110

### install image processing libraries including OpenCV (optional)

$ pip install opencv-contrib-python
$ pip install scikit-image
$ pip install pillow
$ pip install imutils

## [6-Jetson Re-training on the Cat/Dog Dataset]

$ cd jetson-inference/python/training/classification/data
$ wget https://nvidia.box.com/shared/static/o577zd8yp3lmxf5zhm38svrbrv45am3y.gz -O cat_dog.tar.gz
$ tar xvzf cat_dog.tar.gz

$ cd jetson-inference/python/training/classification
$ python3 train.py --model-dir=models/cat_dog data/cat_dog

$ python3 onnx_export.py --model-dir=models/cat_dog