lato.2_mrp_mlx/lato_mlx
John ee86bdb4e4 Pass mode='max' explicitly to downsample
LATO.2 hardcodes reduce='amax' but Pixal3D's version of the same op defaults to
'mean'. Both now share one implementation, so relying on its default would silently
change every downsampled feature if that default ever moved.
2026-08-02 10:42:18 +10:00
..
models Pass mode='max' explicitly to downsample 2026-08-02 10:42:18 +10:00
__init__.py MLX sparse core: SubMConv3d + SparseTensor + weight converter 2026-08-02 10:04:24 +10:00
convert.py MLX sparse core: SubMConv3d + SparseTensor + weight converter 2026-08-02 10:04:24 +10:00