installation

Installation

The AIPyS platform is engineered to streamline the integration of AI with photoswitchable genetic CRISPR screen technology. Presently, AIPyS supports installation on Windows systems and requires Python 3.8. To ensure optimal functionality, particularly for components like PyTorch and Cellpose, matching the CUDA and cuDNN versions to your system’s configuration is crucial.

Prerequisites

Before proceeding with the installation, ensure your system meets the following requirements:

  • Operating System: Windows

  • Python Version: 3.8

Additionally, for functionalities leveraging PyTorch and Cellpose, verify you have the appropriate versions of CUDA and cuDNN installed. These components are essential for computation offloading to GPUs, significantly enhancing performance for machine learning tasks.

Installation Options

AIPyS offers two primary methods for installation: via an environment.yml file using Conda and directly through pip.

Option 1: Using environment.yml with Conda

For users preferring Conda, AIPyS provides an environment.yml file that simplifies the setup process, managing both Python version constraints and necessary library dependencies, including CUDA and cuDNN configurations.

  1. Download the environment.yml file from the AIPyS repository.

  2. Open a Windows command prompt or PowerShell window.

  3. Navigate to the directory containing the downloaded environment.yml file.

  4. Execute the following command:

conda env create -f environment.yml

  1. Activate the newly created Conda environment:

conda activate aipys_env

This environment will include all the necessary dependencies pre-configured, adhering to the CUDA and cuDNN versions required for PyTorch and Cellpose.

Option 2: Installation with pip

Alternatively, AIPyS can be installed using pip, which is Python’s package manager. This method is straightforward but requires manual management of compatible CUDA and cuDNN versions.

pip install --upgrade AIPyS

Before executing the install command, ensure your Python environment is configured with Python 3.8. As AIPyS is currently optimized for Windows, users on other platforms should carefully consider compatibility, especially regarding CUDA and cuDNN dependencies.

Verifying Installation

After completing the installation using either method, you can verify the setup by checking the installed AIPyS version:

aipys --version

Alternatively, attempt running a simple command or script to confirm operational functionality.

Troubleshooting

  • CUDA/cuDNN Compatibility: If you encounter issues related to PyTorch or Cellpose, verify that you have installed the correct versions of CUDA and cuDNN that match the requirements of these libraries.

  • Python Version: Ensure that the Python environment active during installation is version 3.8, as discrepancies in Python versions can lead to compatibility problems.

  • GPU Recognition Issues: If your GPU is not recognized and you are using an NVIDIA Quadro P4000, ensure that you have the appropriate drivers installed. Additionally, install PyTorch with a CUDA version that matches your setup by executing the following command:

    pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1+cu113 -f https://download.pytorch.org/whl/torch_stable.html

    This command installs the versions of PyTorch, torchvision, and torchaudio that are compatible with CUDA 11.3, specifically optimized for the NVIDIA Quadro P4000 GPU.

Evaluate PyTorch-GPU

import torch
print(torch.version.cuda)

For further assistance, refer to the FAQs section or reach out to the AIPyS support team.