For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).
Mojo function
group_norm_reshape
def group_norm_reshape[dtype: DType, rank: Int](shape: IndexList[rank, element_type=shape.element_type], buf: TileTensor[dtype, Storage=buf.Storage, address_space=buf.address_space, linear_idx_type=buf.linear_idx_type, element_size=buf.element_size], channels_per_group: Int, spatial: Int) -> TileTensor[dtype, Layout[*?, *?], 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[*?, *?], buf.origin, address_space=buf.address_space]
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!