Akshay Agrawal
Akshay Agrawal
> Thanks for the response, and I tried using ECOS but it gives an implementation error. 'NotImplementedError: Exponential cones not supported yet.' @sinemuysal Can you give a complete code snippet...
Great question. We have an open issue to add warm-start support, but have not yet gotten around to it. (https://github.com/cvxgrp/cvxpylayers/issues/35) We'll keep this issue open until we do add it.
Try upgrading your version of CVXPY
Sorry for the late response ...! Looks mostly good! Just a few comments. 1. I was confused by the wording in the first sentence, since the support of a probability...
LGTM modulo the comments I made previously.
@haowwang, thanks for trying out CVXPY Layers! PyTorch is most efficient at computing gradients of scalar valued functions, so by default it disallows computing a full Jacobian. To compute a...
For something like that, you might need to write some custom code. It will be equivalent to solving the problem, getting the optimal value of the variable you are interested...
Not quite. It computes the gradient times the all ones vector. PyTorch never materializes full derivative matrices, it only implements the (transpose) of the matrix vector product for derivatives. This...
Hey Brandon! I suppose code like this could eventually live in CVXPY. Your instinct is right, in that CVXPY Layers shouldn't rely on the internals of the `compiler` object, since...
Unassigned myself, because I'm not working on this currently. If anyone wants to attempt this, that would be great!