For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /get-started.md).

Mojo function

group_norm_reshape

def group_norm_reshape[dtype: DType, rank: Int](shape: Coord, buf: TileTensor[dtype, Engine=buf.Engine, address_space=buf.address_space, linear_idx_type=buf.linear_idx_type], channels_per_group: Int, spatial: Int) -> TileTensor[dtype, Layout[TypeList[Int64, Int64](), TypeList[Int64, Int64]()], buf.origin, address_space=buf.address_space]

Reshapes an input buffer for group normalization by flattening all dimensions except the group dimension. Returns a 2D buffer of shape (num_groups * N, group_size), where group_size is the product of channels_per_group and spatial.

Returns:

TileTensor[dtype, Layout[TypeList[Int64, Int64](), TypeList[Int64, Int64]()], buf.origin, address_space=buf.address_space]