Mojo function

index_tensor

index_tensor[dtype: DType, indices_type: DType, data_rank: Int, indices_rank: Int, output_rank: Int, batch_dims: Int, target: StringSlice[StaticConstantOrigin] = "cpu", single_thread_blocking_override: Bool = False](data: NDBuffer[dtype, data_rank, origin], indices: NDBuffer[indices_type, indices_rank, origin], output: NDBuffer[dtype, output_rank, origin], ctx: DeviceContextPtr)

Index_tensor operation; based on modified implementation of gather_nd.

Parameters:

dtype (DType): Type of data tensor.
indices_type (DType): Type of indices tensor.
data_rank (Int): Rank of data tensor (data_rank >= 1).
indices_rank (Int): Rank of indices tensor (indices_rank >= 1).
output_rank (Int): Rank of output tensor.
batch_dims (Int): Number of batch dimensions. The gather of indexing starts from dimension of data[batch_dims:].
target (StringSlice[StaticConstantOrigin]): The target architecture to execute on.
single_thread_blocking_override (Bool): If True, then the operation is run synchronously using a single thread.

Args:

data (NDBuffer[dtype, data_rank, origin]): Tensor of rank data_rank >= 1.
indices (NDBuffer[indices_type, indices_rank, origin]): Tensor of rank indices_rank >= 1. All index values are expected to be within bounds [-s, s-1] along axis of size s. It is an error if any of the index values are out of bounds.
output (NDBuffer[dtype, output_rank, origin]): Tensor of rank data_rank + indices_rank - indices_shape[-1] - 1 - b.
ctx (DeviceContextPtr): The DeviceContextPtr as prepared by the graph compiler.