ehvr20/scvi-cuda-rapids repository overview

⁠scvi-cuda-rapids

⁠Overview:

This Docker image provides an environment based on NVIDIA CUDA with Micromamba, tailored for bioinformatics and single-cell RNA sequencing (scRNA-seq) analysis. It includes essential libraries such as scanpy, scvi-tools, and scib, with additional GPU acceleration of data frames and graphs provided by rapidsai.

⁠Features:

• NVIDIA CUDA for GPU acceleration
• micromamba for efficient package management
• scanpy: Python-based single-cell RNA-seq analysis toolkit
• scvi-tools: Deep generative models for single-cell RNA-seq
• scib: Integration benchmarking for single-cell RNA-seq
• rapidsai: python GPU-accelerated data science libraries

⁠Usage:

1. Pull the Docker image:

   docker pull ehvr20/scvi-cuda-rapids:latest
   
   

   Copy
2. Run the Docker image:

   docker run --gpus all -it ehvr20/scvi-cuda-rapids:latest python
   

   Copy

⁠Base Image:

• nvidia/cuda:12.2.2-devel-ubuntu20.04

⁠Environment:

⁠Micromamba: environment.yml

channels:
 - pytorch
 - rapidsai-nightly
 - conda-forge
 - nvidia
dependencies:
 - python=3.11
 - rapidsai-nightly::cudf=24.04
 - rapidsai-nightly::cuml=24.04
 - rapidsai-nightly::cugraph=24.04
 - pytorch::pytorch
 - pytorch::pytorch-cuda
 - pytorch::torchaudio
 - pytorch::torchvision
 - conda-forge::jax 
 - conda-forge::jaxlib
 - conda-forge::scvi-tools
 - conda-forge::cuda-version
 - nvidia::cuda-nvcc
 - scanpy=1.9.8
 - ipykernel
 - pip:
   - scib 

⁠pip:

pip install --upgrade "jax[cuda12_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html