Numpy's eigh(a, 'L') (eigh(a, 'U')) is a function of tril(a) (triu(a))
only.  This means that partial derivative of eigh(a, 'L') (eigh(a,
'U')) with respect to a[i,j] is zero for i < j (i > j).  At the same
time, non-zero components of the gradient must account for the fact
that variation of the opposite triangle contributes to variation of
two elements of Hermitian (symmetric) matrix.

Issue #1080: Make TensorVariable interface more similar to that of numpy.ndarray

Fixed a typo.

Ready to merge: Cuda allocation code

Fixed a typo.

make compiled function's fgraphs have same name as function

Make work the new extra debug info just merged and add extra debug about...

This time I ran the code and saw the error to validate it. It is working correctly.

to make it more clear how this function avoids unnecessary reallocation

Make opt work when shape has been specified.