, t2, out_format=mode_format(compressed), dtype=None)

The dot product of two tensors.

This implements the same spec as NumPy but allows for sparse taco tensors as operands.

  • If both t1 and t2 are 1-D then this is an inner product.

  • If either operand is a scalar, this is equivalent to element-wise multiply.

  • Equivalent to matrix multiplication if both tensors are 2-D

  • If t1 is N-D and t2 is 1-D,t is a sum product over the last axis of t1 and t2

  • if t1 is an N-D array and t2 is an M-D array (where M>=2), it is a sum product over the last axis of t1 and the second-to-last axis of t2:

We could write this using index expressions for two 3-D tensors as: T[i, j, k, m] = t1[i, j, l] * t2[k, l, m]

In the above, T = dot(t1, t2)

t1, t2: tensors, array_like

The arguments to dot. The side of the last dimension of t1 must be equal to the size of the second to last dimension of t2.

out_format: format, mode_format, optional
  • If a format is specified, the result tensor is stored in the format out_format.

  • If a mode_format is specified, the result the result tensor has a with all of the dimensions stored in the mode_format passed in.

dtype: Datatype

The datatype of the output tensor.

res: tensor

The dot product of the input tensors.


>>> import pytaco as pt
>>>, 4)[0]
>>> a = [[1, 0], [0, 1]]
>>> b = [[4, 1], [2, 2]]
>>>, b).to_array()
array([[4, 1],
       [2, 2]], dtype=int64)