NVIDIA introduces pip-installable cuML wheels on PyPI, simplifying installation and broadening accessibility by reducing CUDA binary sizes.