rapids — NVIDIA RAPIDS
NVIDIA’s RAPIDS suite — cuDF, cuML, cuGraph, and friends. GPU-accelerated drop-in replacements for pandas, scikit-learn, NetworkX.
Image tag
docker.io/manvarharsh/rapids:cuda12
What’s in this image
- Base:
nvidia/cuda:12.4.1-cudnn-devel-ubuntu22.04 - Python 3.10 (conda)
- cuDF, cuML, cuGraph, cuPy, cuSpatial
- Dask, Dask-CUDA for multi-GPU workloads
- JupyterHub with Podstack authenticator
- OpenSSH server
Default ports
| Port | Service |
|---|---|
| 22 | SSH |
| 8000 | JupyterHub |
Use cases
- GPU-accelerated ETL on large dataframes
- Multi-GPU data processing with Dask
- Fast classical ML (Random Forest, XGBoost) on big tabular data
- Graph analytics on GPU (PageRank, community detection)
Environment variables
| Variable | Description |
|---|---|
ENABLE_SSH | Enable SSH server |
ENABLE_JUPYTERHUB | Enable JupyterHub on port 8000 |
PODSTACK_API_URL | Backend URL for JupyterHub token validation |
SSH_PUBLIC_KEY | Public key for SSH |
Persistence
Mount at /data. Datasets and notebooks under /data/.