Open SuperSashka opened 1 year ago
Some computation speedup probably can be achieved with pytorch 2.0 using the compiler.
At the time being it works only in Linux or WSL. We'll add functionality as soon as possible. The test version is in https://github.com/ITMO-NSS-team/torch_DE_solver/tree/compiler branch.
We need to make sure that we can use LazyTensor mode https://arxiv.org/pdf/2102.13267.pdf and implement it for faster computation.