Autograd can automatically differentiate native Python and Numpy code…

…. It can

handle a large subset of Python's features, including loops, ifs, recursion and
closures, and it can even take derivatives of derivatives of derivatives. It
supports reverse-mode differentiation (a.k.a. backpropagation), which means it
can efficiently take gradients of scalar-valued functions with respect to
array-valued arguments, as well as forward-mode differentiation, and the two
can be composed arbitrarily. The main intended application of Autograd is
gradient-based optimization.

WWW: https://github.com/HIPS/autograd

math/Makefile

@@ -690,6 +690,7 @@
     SUBDIR += py-algopy
     SUBDIR += py-altgraph
     SUBDIR += py-apgl
+    SUBDIR += py-autograd
     SUBDIR += py-basemap
     SUBDIR += py-basemap-data
     SUBDIR += py-bayesian-optimization

math/py-autograd/Makefile

@@ -0,0 +1,22 @@
+# $FreeBSD$
+
+PORTNAME=	autograd
+DISTVERSION=	1.2
+CATEGORIES=	math python
+MASTER_SITES=	CHEESESHOP
+PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}
+
+MAINTAINER=	[email protected]
+COMMENT=	Efficiently computes derivatives of numpy code
+
+LICENSE=	MIT
+
+RUN_DEPENDS=	${PYNUMPY} \
+		${PYTHON_PKGNAMEPREFIX}future>=0.15.2:devel/py-future@${PY_FLAVOR}
+
+USES=		python
+USE_PYTHON=	autoplist distutils
+
+NO_ARCH=	yes
+
+.include <bsd.port.mk>

math/py-autograd/distinfo

@@ -0,0 +1,3 @@
+TIMESTAMP = 1551302910
+SHA256 (autograd-1.2.tar.gz) = a08bfa6d539b7a56e7c9f4d0881044afbef5e75f324a394c2494de963ea4a47d
+SIZE (autograd-1.2.tar.gz) = 32540

math/py-autograd/pkg-descr

@@ -0,0 +1,10 @@
+Autograd can automatically differentiate native Python and Numpy code. It can
+handle a large subset of Python's features, including loops, ifs, recursion and
+closures, and it can even take derivatives of derivatives of derivatives. It
+supports reverse-mode differentiation (a.k.a. backpropagation), which means it
+can efficiently take gradients of scalar-valued functions with respect to
+array-valued arguments, as well as forward-mode differentiation, and the two
+can be composed arbitrarily. The main intended application of Autograd is
+gradient-based optimization.
+
+WWW: https://github.com/HIPS/autograd