geektechnophile/open3d-pytorch-cuda repository overview

⁠geektechnophile/open3d-pytorch-cuda:1.0

⁠CI

Conda/Micromamba-based build image for Debian with Python 3.11, Open3D, and PyTorch CUDA 13.0. This image is used to power GPU-based ML and 3D application workflows.

⁠Description

This image provides a ready-to-use environment with:

• Python 3.11 via Micromamba
• Open3D 0.18.0
• PyTorch 2.9.0 + CUDA 13.0 and related NVIDIA CUDA libraries
• Torchvision 0.24.0, Torchaudio 2.9.0, Triton 3.5.0
• Common Python utilities: filelock, fsspec, networkx, sympy, mpmath, jinja2
• Minimal Debian packages for GPU and OpenGL support: libgomp1, libgl1

⁠License

This image is licensed under BSD-3 Clause⁠.

⁠Documentation & Contributing

• Documentation for usage is available via the upstream Dockerfile.
• To report issues or request features, please open an issue on the repository repo⁠.
• Contributions are welcome via pull requests, following the code of conduct.

⁠Important Image Tags

• 1.0 — Base image with Open3D, PyTorch, CUDA 13.0, ready for GPU workloads

⁠Layer Overview

⁠Getting Started & Usage

Run the image directly with GPU support:

docker run --rm -it --gpus all geektechnophile/open3d-pytorch-cuda:1.0 bash

Use it as a base image in your Dockerfile:

FROM geektechnophile/open3d-pytorch-cuda:1.0
WORKDIR /workspace

This image is suitable for CI pipelines, GPU experiments, or as a base for building ML and 3D applications.