A city recycling-sort yard already has a cleaned cuDF table of bin weights and contamination flags. Staff want scikit-learn-style fit and predict without leaving the GPU. Which official library is that?
Select an answer to reveal the explanation.
Short Explanation
A cleaned cuDF table of bin weights wants scikit-learn-style fit and predict on the GPU. That library is cuML. Do not invent a new NVIDIA product, write a kernel, or send bin weights through TensorRT-LLM.
Full Explanation
cuML is the official RAPIDS library for GPU-accelerated ML estimators on prepared frames. A cleaned cuDF table of bin weights is exactly that input. Do not invent a new NVIDIA product or write a kernel. TensorRT-LLM optimizes LLM inference, not tabular fit and predict.