automatic differentiation made easier for C++
Full Changelog: https://github.com/autodiff/autodiff/compare/v1.1.1...v1.1.2
Full Changelog: https://github.com/autodiff/autodiff/compare/v1.1.0...v1.1.1
Full Changelog: https://github.com/autodiff/autodiff/compare/v1.0.3...v1.1.0
real
, dual
and var
.AUTODIFF_ENABLE_IMPLICIT_CONVERSION_VAR
in addition to AUTODIFF_ENABLE_IMPLICIT_CONVERSION_REAL
, AUTODIFF_ENABLE_IMPLICIT_CONVERSION_DUAL
and AUTODIFF_ENABLE_IMPLICIT_CONVERSION
.Full Changelog: https://github.com/autodiff/autodiff/compare/v1.0.2...v1.0.3
__rpow
Python binding by @allanleal in https://github.com/autodiff/autodiff/pull/290
Full Changelog: https://github.com/autodiff/autodiff/compare/v1.0.1...v1.0.2
autodiff::real
types by @allanleal in https://github.com/autodiff/autodiff/pull/282
Full Changelog: https://github.com/autodiff/autodiff/compare/v1.0.0...v1.0.1
autodiff
has now been sufficiently battle tested and its API has remained quite stable over the past years. It's now been promoted to v1.0.0! 🚀
Full Changelog: https://github.com/autodiff/autodiff/compare/v0.6.12...v1.0.0
This release introduces support to the use of Eigen::Map
when performing gradient
, jacobian
and hessian
operations with number types autodiff::real
and autodiff::dual
.
This release also introduces static assertions that ensure the provided vector in a wrt
list is non-const, because the numbers in these vectors need to be seeded before evaluating the function being derived, and later unseeded; they must then be altered.
This release fixed the computation of higher-order derivatives when hypot
is used with var
(see #243).
It also implements a single template constructor for Dual<T, G>
so that Intellisense errors are prevented.
cmake -E rm
command in autodiff/python/package/CMakeLists.txt
by cmake -E remove_directory
so that CMake v3.16 can be used. cmake -E rm
is only available from CMake 3.17