← All skills
cuml
85 / 100 · Low risk
Fine to try.
By NVIDIA/skills (GitHub) · coding · 3,933 stars · scanned at commit 2630133
Use when writing, running, debugging or upgrading Python code that uses NVIDIA cuML (GPU machine learning), cuml.accel (zero-code-change GPU acceleration of scikit-learn, umap-learn and hdbscan), or GPU UMAP/HDBSCAN/RandomForest/KMeans/SVC, including through libraries that wrap cuML (BERTopic, rapids-singlecell). Covers APIs removed or renamed in 2025-2026 releases, verifying code really ran on the GPU, large-data UMAP/HDBSCAN memory recipes, and exporting trained models. Not for cuDF-only data…
Works with any agent that supports SKILL.md skills.
What this skill can do
Downloads and installs other software
The author does not mention this.
Found in 13 places.
Show 10 more
- skills/cuml/evals/evals.json, line 120
"expected_output": "Lays out the integration options with tradeoffs: (1) zero-code path: document that users can run under cuml.accel / CUML_ACCEL_ENABLED=1 and - skills/cuml/evals/evals.json, line 170
"expected_output": "Gives `pip install cuml-cu13` (CUDA 13 wheels on pypi.org match driver 580), with no extra index URL required, and a quick import check. Doe - skills/cuml/evals/evals.json, line 173
"Does not recommend `pip install cuml` without a CUDA suffix", - skills/cuml/evals/evals.json, line 192
"Does not suggest `cuml-cu11` or `pip install cuml`" - skills/cuml/evals/evals.json, line 205
"expected_output": "Recommends the official `rapidsai/base` image with a CUDA-major tag such as `rapidsai/base:26.08-cuda12-py3.12` (or `-cuda13-`), or a slim P - skills/cuml/evals/evals.json, line 207
"Uses a valid base image approach: rapidsai/base with a CUDA-major tag (cudaNN, e.g. cuda12 or cuda13), or a Python image with pip install cuml-cu12/cuml-cu13", - skills/cuml/evals/evals.json, line 331
"expected_output": "Uses `dask_cuda.LocalCUDACluster` (one worker per GPU) and a `dask.distributed.Client`, reads the parquet files with `dask_cudf.read_parquet - skills/cuml/evals/evals.json, line 435
"prompt": "Can I install cuML on Windows 11 with my RTX 3080? I'd like to just pip install it into my normal Windows Python. If not, what's the simplest way to - skills/cuml/evals/evals.json, line 436
"expected_output": "Native Windows is not supported; use WSL2 (Ubuntu) with the Windows NVIDIA driver, then `pip install cuml-cu12` or `cuml-cu13` inside WSL2 ( - skills/cuml/evals/evals.json, line 438
"States that native Windows pip install is not supported",
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'cuml' ] || [ -L ~/.agents/skills/'cuml' ]; then echo 'A skill folder named cuml already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/NVIDIA/skills' "$d" && git -C "$d" checkout --quiet '2630133f22e357f2005886567a39eddcc1b520e2' -- 'skills/cuml' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/cuml' ~/.agents/skills/'cuml' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'cuml' ~/.claude/skills/'cuml' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for NVIDIA/skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/NVIDIA/skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-11 | 2630133 | 85 |
| 2026-11-08 | fce3cca | 85 |
| 2026-11-07 | fce3cca | 85 |