mxnet.npx.reshape_like

reshape_like(lhs=None, rhs=None, lhs_begin=_Null, lhs_end=_Null, rhs_begin=_Null, rhs_end=_Null, out=None, name=None, **kwargs)

Reshape some or all dimensions of lhs to have the same shape as some or all dimensions of rhs.

Returns a view of the lhs array with a new shape without altering any data.

Example:

x = [1, 2, 3, 4, 5, 6]
y = [[0, -4], [3, 2], [2, 2]]
reshape_like(x, y) = [[1, 2], [3, 4], [5, 6]]

More precise control over how dimensions are inherited is achieved by specifying slices over the lhs and rhs array dimensions. Only the sliced lhs dimensions are reshaped to the rhs sliced dimensions, with the non-sliced lhs dimensions staying the same.

Examples:

- lhs shape = (30,7), rhs shape = (15,2,4), lhs_begin=0, lhs_end=1, rhs_begin=0, rhs_end=2, output shape = (15,2,7)
- lhs shape = (3, 5), rhs shape = (1,15,4), lhs_begin=0, lhs_end=2, rhs_begin=1, rhs_end=2, output shape = (15)

Negative indices are supported, and None can be used for either lhs_end or rhs_end to indicate the end of the range.

Example:

- lhs shape = (30, 12), rhs shape = (4, 2, 2, 3), lhs_begin=-1, lhs_end=None, rhs_begin=1, rhs_end=None, output shape = (30, 2, 2, 3)

Defined in /work/mxnet/src/operator/tensor/elemwise_unary_op_basic.cc:L629

Parameters

  • lhs (ndarray) – First input.

  • rhs (ndarray) – Second input.

  • lhs_begin (int or None, optional, default='None') – Defaults to 0. The beginning index along which the lhs dimensions are to be reshaped. Supports negative indices.

  • lhs_end (int or None, optional, default='None') – Defaults to None. The ending index along which the lhs dimensions are to be used for reshaping. Supports negative indices.

  • rhs_begin (int or None, optional, default='None') – Defaults to 0. The beginning index along which the rhs dimensions are to be used for reshaping. Supports negative indices.

  • rhs_end (int or None, optional, default='None') – Defaults to None. The ending index along which the rhs dimensions are to be used for reshaping. Supports negative indices.

  • out (ndarray, optional) – The output ndarray to hold the result.

Returns

out – The output of this function.

Return type

ndarray or list of ndarrays