dot-abstract_operator: Abstract function for tensor-numeric operations Backward...

Description Usage

Description

Abstract function for tensor-numeric operations Backward function rather than partial derivative? Problem: newly created tensor has by default Maybe we don't need to track some 'abstract tensors' Memoisation can handle this case instead

Usage

1

krzjoa/aria documentation built on Oct. 1, 2020, 12:48 p.m.