Route Python GPU compute to CuPy or NumPy with CPU fallbacks.
npx skills add https://github.com/VKirill/antigravity-for-claude-code --skill cuda-python
Please help me install this Agent Skill. Skill: cuda-python Source: https://github.com/VKirill/antigravity-for-claude-code/tree/main/skills/cuda-python Command: npx skills add https://github.com/VKirill/antigravity-for-claude-code --skill cuda-python
GPU-accelerated Python code often breaks on machines without NVIDIA GPUs or with mismatched CUDA driver/toolkit/library versions, forcing developers to add brittle imports and scattered conditionals.
xp namespace (CuPy if usable, NumPy otherwise), and guarantees CPU-safe imports so the same package works in CI and production.Ask: "Plan and implement a CPU/GPU optional module using cuda-python optional dependency detection, then show how to write one function that uses xp on GPU when available and falls back to CPU when not."
High-intent search queries and answers about installing and using this skill.
To make Python CUDA code fall back to CPU, implement an optional-dependency pattern that detects the GPU at runtime and routes array workloads to NumPy if CuPy is unavailable, ensuring imports remain CPU-safe in CI and production.
Choosing between CuPy, Numba, and cuda-python depends on task level: use CuPy for array workloads, Numba @cuda.jit for custom kernel development, and cuda-python for driver-level diagnostics and low-level bindings across multi-GPU environments.
Write a GPU-optional Python function by exposing an xp namespace that maps to CuPy on GPU and NumPy on CPU, allowing the same code path to execute on either architecture based on runtime hardware detection.
Python GPU code fails in CI from CUDA driver and toolkit mismatches because runtime environments lack compatible NVIDIA hardware or libraries, requiring guardrails to detect versions and enforce safe CPU fallbacks.
You can run CUDA-accelerated Python packages in Docker containers without NVIDIA GPUs by enforcing an optional-dependency pattern with runtime GPU detection that gracefully degrades to CPU-safe NumPy operations.
jayll1303Build GPU-enabled Docker containers with NGC base images and uv.
K-Dense-AITransform CPU-bound Python numerical code into CUDA-accelerated versions.
guqiong96Add custom CUDA kernels to the SGLang JIT kernel module.
mkurmanEnable GPU acceleration for Python workloads across CUDA-enabled libraries.