Akshay Agrawal
Akshay Agrawal
Perhaps the docs can be clearer. The scalar product is a quasiconcave atom, provided that its arguments are nonnegative. This does not mean that the product of two arbitrary nonnegative...
Unfortunately, no, there is no more user-friendly way to override or extend CVXPY's analyses of expressions. The error with the sum expression has to do with the fact (I'm guessing)...
@dotPiano, thanks for the kind & helpful feedback!
This is a know issue. DPP compilation will likely never be as fast as non-DPP compilation -- the first compilation of a parametrized program performs significantly more work, to dramatically...
@kmonson , @sjschneider : this is a reasonable request. We have plans to decrease the overhead of DPP compilation, but these plans are not on our near-term roadmap. Having a...
3 is super cool! I endorse 1 and 2 as things we can do immediately.
> Another option would be to check all numeric constants, so in constant.py. That won't catch parameters but might be the best place to start, since it will be easier...
Any piecewise-linear function can be written as the [maximum of linear functions](https://web.stanford.edu/class/ee364a/lectures/functions.pdf). So in CVXPY, you might write something like ``` pwl = cp.maximum(a_1.T @ x + b_1, a_2.T @...
> I would expose the backend similarly to how we expose solvers, as an argument to solve. If we do this, we should have a plan to deprecate the `CPP`...
Looks great! What a Herculean PR. Congrats Philipp for getting this through.