All skills
cuopt-developer
Modify, build, test, and debug the NVIDIA cuOpt codebase, including C++/CUDA, Python bindings, and server.
- Part of NVIDIA/skills
- 2.8k installs
- Machine Learning
When to use it
You already have a cuOpt clone and need to add a solver parameter or fix a failing C++ test before opening a PR.
For example
You say
Using cuopt-developer, I want to add a new routing solver parameter to the Python bindings and C++ core, then run the relevant tests.
You get
- Parameter changes in C++ and Python
- Updated tests
- Build and test results
How it works
- Refuses sudo and system-level changes; the dev workflow stays in a user-space conda env.
- Checks CUDA driver version against the conda env file before building to avoid runtime RMM errors.
- Requires downloading test datasets and setting RAPIDS_DATASET_ROOT_DIR before running tests.
More Machine Learning skills
See all- fine-tuning-expert4.6kHelps fine-tune language models with LoRA adapters, training datasets, hyperparameters and deployment.
- h3-image-prompt-writing174Writes MiniMax-H3-Image prompts for posters, infographics, product shots, characters and scenes, and can render them.
- forecasting-reverso30Forecasts future values from time series data using the Reverso model on CPU.
- detecting-tips-zones16Finds image regions from text labels like "dog face" to produce bounding boxes automatically.
Also in this pack
- accelerated-computing-cudf3.2kGuides GPU DataFrame work with cuDF, including pandas acceleration, joins, groupby, file I/O and multi-GPU jobs.
- cuopt-install2.6kInstalls cuOpt for Python, C, or server via pip, conda, or Docker and verifies it works.
- data-designer2.6kGenerates a synthetic dataset from your description of the data you need.
- dynamo-recipe-runner2.6kSelects, validates, patches, and deploys NVIDIA Dynamo Kubernetes recipes for model serving.