nixpkgs/pkgs/development/python-modules/tensorflow/default.nix
Dmitry Vyal 6abb29f460
tensorflow-1.1.0 for python3.6
Google publishes prebuilt tensorflow whl for python 3.4, 3.5, 3.6,
but nix expression for tensorflow only supported 3.5.
This change adds support for python-3.6.
2017-09-12 12:52:35 +02:00

140 lines
4.0 KiB
Nix

{ stdenv
, fetchurl
, buildPythonPackage
, isPy36, isPy35, isPy27
, cudaSupport ? false
, cudatoolkit ? null
, cudnn ? null
, linuxPackages ? null
, numpy
, six
, protobuf
, swig
, werkzeug
, mock
, zlib
}:
assert cudaSupport -> cudatoolkit != null
&& cudnn != null
&& linuxPackages != null;
# unsupported combination
assert ! (stdenv.isDarwin && cudaSupport);
# tensorflow is built from a downloaded wheel, because the upstream
# project's build system is an arcane beast based on
# bazel. Untangling it and building the wheel from source is an open
# problem.
buildPythonPackage rec {
pname = "tensorflow";
version = "1.1.0";
name = "${pname}-${version}";
format = "wheel";
disabled = ! (isPy36 || isPy35 || isPy27);
src = let
tfurl = sys: proc: pykind:
let
tfpref = if proc == "gpu"
then "gpu/tensorflow_gpu"
else "cpu/tensorflow";
in
"https://storage.googleapis.com/tensorflow/${sys}/${tfpref}-${version}-${pykind}.whl";
dls =
{
darwin.cpu = {
py2 = {
url = tfurl "mac" "cpu" "py2-none-any" ;
sha256 = "1fgf26lw0liqxc9pywc8y2mj8l1mv48nhkav0pag9vavdacb9mqr";
};
py3 = {
url = tfurl "mac" "cpu" "py3-none-any" ;
sha256 = "0z5p1fra7bih0vqn618i2w3vyy8d1rkc72k7bmjq0rw8msl717ia";
};
};
linux-x86_64.cpu = {
py2 = {
url = tfurl "linux" "cpu" "cp27-none-linux_x86_64";
sha256 = "0ld3hqx3idxk0zcrvn3p9yqnmx09zsj3mw66jlfw6fkv5hznx8j2";
};
py35 = {
url = tfurl "linux" "cpu" "cp35-cp35m-linux_x86_64";
sha256 = "0ahz9222rzqrk43lb9w4m351klkm6mlnnvw8xfqip28vbmymw90b";
};
py36 = {
url = tfurl "linux" "cpu" "cp36-cp36m-linux_x86_64";
sha256 = "1a2cc8ihl94iqff76nxg6bq85vfb7sj5cvvi9sxy2f43k32fi4lv";
};
};
linux-x86_64.cuda = {
py2 = {
url = tfurl "linux" "gpu" "cp27-none-linux_x86_64";
sha256 = "1baa9jwr6f8f62dyx6isbw8yyrd0pi1dz1srjblfqsyk1x3pnfvh";
};
py35 = {
url = tfurl "linux" "gpu" "cp35-cp35m-linux_x86_64";
sha256 = "0606m2awy0ifhniy8lsyhd0xc388dgrwksn87989xlgy90wpxi92";
};
py36 = {
url = tfurl "linux" "gpu" "cp36-cp36m-linux_x86_64";
sha256 = "0lvbmfa87qzrajadpsf13gi3l71vryzkryzqfvkykivqrdjsvj8q";
};
};
};
in
fetchurl (
if stdenv.isDarwin then
if isPy27 then
dls.darwin.cpu.py2
else
dls.darwin.cpu.py3
else if isPy36 then
if cudaSupport then
dls.linux-x86_64.cuda.py36
else dls.linux-x86_64.cpu.py36
else if isPy35 then
if cudaSupport then
dls.linux-x86_64.cuda.py35
else dls.linux-x86_64.cpu.py35
else
if cudaSupport then
dls.linux-x86_64.cuda.py2
else
dls.linux-x86_64.cpu.py2
);
propagatedBuildInputs = with stdenv.lib;
[ numpy six protobuf swig werkzeug mock ]
++ optionals cudaSupport [ cudatoolkit cudnn stdenv.cc ];
# Note that we need to run *after* the fixup phase because the
# libraries are loaded at runtime. If we run in preFixup then
# patchelf --shrink-rpath will remove the cuda libraries.
postFixup = let
rpath = stdenv.lib.makeLibraryPath
(if cudaSupport then
[ stdenv.cc.cc.lib zlib cudatoolkit cudnn
linuxPackages.nvidia_x11 ]
else
[ stdenv.cc.cc.lib zlib ]
);
in
''
find $out -name '*.so' -exec patchelf --set-rpath "${rpath}" {} \;
'';
doCheck = false;
meta = with stdenv.lib; {
description = "TensorFlow helps the tensors flow";
homepage = http://tensorflow.org;
license = licenses.asl20;
maintainers = with maintainers; [ jpbernardy ];
platforms = with platforms; if cudaSupport then linux else linux ++ darwin;
};
}