kpriv/freebsd-ports
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

misc/py-xformers: New port: Transformers building blocks, supporting a composable construction

Browse source
This commit is contained in:
Yuri Victorovich 2024-03-21 21:11:05 -07:00
parent 471d29fc25
commit 60c128a82a
5 changed files with 44 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
misc/Makefile
View file

@ -509,6 +509,7 @@
SUBDIR += py-vaderSentiment
SUBDIR += py-wandb
SUBDIR += py-wurlitzer
SUBDIR += py-xformers
SUBDIR += py-xgboost
SUBDIR += pyobd
SUBDIR += pytorch

25
misc/py-xformers/Makefile Normal file
View file

@ -0,0 +1,25 @@
PORTNAME= xformers
DISTVERSION= 0.0.25
CATEGORIES= misc python # machine-learning
MASTER_SITES= PYPI
PKGNAMEPREFIX= ${PYTHON_PKGNAMEPREFIX}
MAINTAINER= yuri@FreeBSD.org
COMMENT= Transformers building blocks, supporting a composable construction
WWW= https://facebookresearch.github.io/xformers/
LICENSE= BSD3CLAUSE
LICENSE_FILE= ${WRKSRC}/LICENSE
BUILD_DEPENDS= pybind11>0:devel/pybind11 \
${PYTHON_PKGNAMEPREFIX}pytorch>=2.1:misc/py-pytorch@${PY_FLAVOR}
RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}pytorch>=2.1:misc/py-pytorch@${PY_FLAVOR} \
${PYNUMPY}
USES= compiler:c++17-lang localbase python
USE_PYTHON= distutils autoplist pytest # tests fail, see https://github.com/facebookresearch/xformers/issues/1003
post-install:
@${STRIP_CMD} ${STAGEDIR}${PYTHONPREFIX_SITELIBDIR}/xformers/_C.so
.include <bsd.port.mk>

3
misc/py-xformers/distinfo Normal file
View file

@ -0,0 +1,3 @@
TIMESTAMP = 1711070989
SHA256 (xformers-0.0.25.tar.gz) = 63f74d96c82d6b9bf3daf38f53cf173a29370ee6bd1dfde15836d0d598b6f82f
SIZE (xformers-0.0.25.tar.gz) = 4080980

11
misc/py-xformers/pkg-descr Normal file
View file

@ -0,0 +1,11 @@
xFormers is a Toolbox to Accelerate Research on Transformers.
xFormers is:
* Customizable building blocks: Independent/customizable building blocks that
can be used without boilerplate code. The components are domain-agnostic and
xFormers is used by researchers in vision, NLP and more.
* Research first: xFormers contains bleeding-edge components, that are not yet
available in mainstream libraries like PyTorch.
* Built with efficiency in mind: Because speed of iteration matters, components
are as fast and memory-efficient as possible. xFormers contains its own CUDA
kernels, but dispatches to other libraries when relevant.

4
misc/py-xformers/pkg-plist Normal file
View file

@ -0,0 +1,4 @@
%%PYTHON_SITELIBDIR%%/xformers/__pycache__/version%%PYTHON_EXT_SUFFIX%%.opt-1.pyc
%%PYTHON_SITELIBDIR%%/xformers/__pycache__/version%%PYTHON_EXT_SUFFIX%%.pyc
%%PYTHON_SITELIBDIR%%/xformers/cpp_lib.json
%%PYTHON_SITELIBDIR%%/xformers/version.py