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

scatter_nd

def scatter_nd[output_type: DType, indices_type: DType, //, target: StringSlice[ImmStaticOrigin] = StringSlice("cpu")](data: TileTensor[output_type, Storage=data.Storage, linear_idx_type=data.linear_idx_type], indices: TileTensor[indices_type, Storage=indices.Storage, linear_idx_type=indices.linear_idx_type], updates: TileTensor[output_type, Storage=updates.Storage, linear_idx_type=updates.linear_idx_type], output: TileTensor[output_type, Storage=output.Storage, linear_idx_type=output.linear_idx_type], context: DeviceContext)

Scatter_nd operation without any reduction.

Parameters:

output_type (DType): Element type of data, updates, and output.
indices_type (DType): Element type of the indices tensor.
target (StringSlice[ImmStaticOrigin]): Target backend to execute on, such as "cpu" or "cuda".

Args:

data (TileTensor[output_type, Storage=data.Storage, linear_idx_type=data.linear_idx_type]): Source tensor of rank at least 1 copied into output before scattering.
indices (TileTensor[indices_type, Storage=indices.Storage, linear_idx_type=indices.linear_idx_type]): Tensor of indices addressing where to write updates into output.
updates (TileTensor[output_type, Storage=updates.Storage, linear_idx_type=updates.linear_idx_type]): Tensor of values to scatter into output at the positions given by indices.
output (TileTensor[output_type, Storage=output.Storage, linear_idx_type=output.linear_idx_type]): Output tensor, shaped the same as data, holding the copied and scattered result.
context (DeviceContext): Device context for execution.