pytaco.as_tensor

pytaco.as_tensor(obj, copy=True)

Converts array_like or scipy csr and csr to tensors.

Converts an array_like object (list of lists, etc..) or scipy csr and csc matrices to a taco tensor.

Parameters

obj: array_like, scipy.sparse.csr_matrix, scipy.sparse.csc_matrix, tensor

The object to convert to a taco tensor. If the object is a tensor, it will be copied depending on the copy flag.

copy: boolean. optional

If true, taco will attempt to take a reference to the input if possible. If false, taco will always copy the input.

Returns

t: tensor

A tensor initialized with the data from the object passed in.

Notes

This method internally uses from_array(), from_sp_csr() and from_sp_csc(). As a result the restrictions to those methods and their copy parameters apply here. For instance, non-contiguous arrays will always be copied regardless of the copy flag.

Python objects will also always be copied since this internally uses np.array to create an array_like object.