Table Of Contents
Table Of Contents

batch_dot

mxnet.ndarray.batch_dot(lhs=None, rhs=None, transpose_a=_Null, transpose_b=_Null, forward_stype=_Null, out=None, name=None, **kwargs)

Batchwise dot product.

batch_dot is used to compute dot product of x and y when x and y are data in batch, namely 3D arrays in shape of (batch_size, :, :).

For example, given x with shape (batch_size, n, m) and y with shape (batch_size, m, k), the result array will have shape (batch_size, n, k), which is computed by:

batch_dot(x,y)[i,:,:] = dot(x[i,:,:], y[i,:,:])

Defined in src/operator/tensor/dot.cc:L125

Parameters:
  • lhs (NDArray) – The first input
  • rhs (NDArray) – The second input
  • transpose_a (boolean, optional, default=0) – If true then transpose the first input before dot.
  • transpose_b (boolean, optional, default=0) – If true then transpose the second input before dot.
  • forward_stype ({None, 'csr', 'default', 'row_sparse'},optional, default='None') – The desired storage type of the forward output given by user, if thecombination of input storage types and this hint does not matchany implemented ones, the dot operator will perform fallback operationand still produce an output of the desired storage type.
  • out (NDArray, optional) – The output NDArray to hold the result.
Returns:

out – The output of this function.

Return type:

NDArray or list of NDArrays