meyelens 0.2.0

Gaze-tracking and pupillometry with MEYE

MEYElens

Quick checklist – fresh install of WSL + GPU-enabled TensorFlow (Windows 11 / Windows 10 21H2+)

All commands are meant for PowerShell (admin) or Ubuntu 22.04 inside WSL as noted.

1. Install WSL and Ubuntu

   Open PowerShell as Administrator

   wsl --install

    Reboot when prompted.
   

   Launch Ubuntu from the Start menu and create a Linux username and password.

2. Install NVIDIA WSL GPU driver on Windows

   Download “CUDA enabled driver for WSL” from NVIDIA’s site, run the installer, then reboot Windows.

3. Prepare Ubuntu repositories (inside WSL)

import NVIDIA key

sudo mkdir -p /usr/share/keyrings curl -fsSL https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-archive-keyring.gpg
| sudo tee /usr/share/keyrings/cuda-archive-keyring.gpg >/dev/null

CUDA repo

echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg]
https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /"
| sudo tee /etc/apt/sources.list.d/cuda.list

cuDNN repo

echo "deb [signed-by=/usr/share/keyrings/cuda-archive-keyring.gpg]
https://developer.download.nvidia.com/compute/cudnn/repos/ubuntu2204/x86_64/ /"
| sudo tee /etc/apt/sources.list.d/cudnn.list

sudo apt update

4. Install minimal CUDA 12.4 runtime and cuDNN 8.9

sudo apt install -y cuda-compat-12-4 libcudnn8 libcudnn8-dev

Add runtime path:

echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/lib/wsl/lib:$LD_LIBRARY_PATH' >> ~/.bashrc source ~/.bashrc sudo ldconfig

Verify cuDNN appears:

ldconfig -p | grep cudnn | head

5. Install Miniconda and TensorFlow

Miniconda (run once)

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh bash Miniconda3-latest-Linux-x86_64.sh source ~/.bashrc

create env and install TF built for CUDA 12 + cuDNN 8.9

conda create -n tf python=3.10 -y conda activate tf pip install --upgrade pip pip install --no-deps tensorflow==2.16.2

6. Confirm everything

driver visible?

nvidia-smi

TensorFlow sees GPU?

python - <<'PY' import tensorflow as tf print("TF:", tf.version) print("GPUs:", tf.config.list_physical_devices('GPU')) PY