gador/nixpkgs
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

python3Packages.torch: 2.10.0 -> 2.11.0 (#502810)

Browse Source
This commit is contained in:
Gaétan Lepage
2026-03-27 23:29:43 +00:00
committed by GitHub
parent de13f4ba6c 1c51f99cb0
commit 4fc8064217
5 changed files with 211 additions and 222 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pkgs/development/python-modules/torch/source/default.nix
+5 -18
View File
@@ -5,7 +5,6 @@
fetchFromGitLab,
git-unroll,
buildPythonPackage,
fetchpatch,
python,
runCommand,
writeShellScript,
@@ -118,7 +117,7 @@ let
setBool = v: if v then "1" else "0";
# https://github.com/pytorch/pytorch/blob/v2.10.0/torch/utils/cpp_extension.py#L2411-L2414
# https://github.com/pytorch/pytorch/blob/v2.11.0/torch/utils/cpp_extension.py#L2569-L2572
supportedTorchCudaCapabilities =
let
real = [
@@ -141,10 +140,10 @@ let
"9.0a"
"10.0"
"10.0a"
"11.0"
"11.0a"
"10.3"
"10.3a"
"11.0"
"11.0a"
"12.0"
"12.0a"
"12.1"
@@ -281,7 +280,7 @@ in
buildPythonPackage.override { inherit stdenv; } (finalAttrs: {
pname = "torch";
# Don't forget to update torch-bin to the same version.
version = "2.10.0";
version = "2.11.0";
pyproject = true;
outputs = [
@@ -305,14 +304,6 @@ buildPythonPackage.override { inherit stdenv; } (finalAttrs: {
patches = [
./clang19-template-warning.patch
# [CPUBLAS] Fix UB: use vector::resize() instead of reserve() before operator[] access
# Merged in https://github.com/pytorch/pytorch/pull/175315
# TODO: drop at the next release
(fetchpatch {
name = "fix-ub-in-cpublas";
url = "https://github.com/pytorch/pytorch/commit/f08aafa9e82c5ae142b97dbfcac1ebd5d9ca7fde.patch";
hash = "sha256-J9QNKDWytA0nBpKr5q4kVnufyMEJHev0mfmyQCxog/w=";
})
]
++ lib.optionals cudaSupport [
./fix-cmake-cuda-toolkit.patch
@@ -368,12 +359,8 @@ buildPythonPackage.override { inherit stdenv; } (finalAttrs: {
--replace-fail '"clang++" if sys.platform == "darwin" else "g++"' \
'"${lib.getExe' targetPackages.stdenv.cc "${targetPackages.stdenv.cc.targetPrefix}c++"}"'
''
# Doesn't pick up the environment variable?
+ lib.optionalString rocmSupport ''
# https://github.com/facebookincubator/gloo/pull/297
substituteInPlace third_party/gloo/cmake/Hipify.cmake \
--replace-fail "\''${HIPIFY_COMMAND}" "python \''${HIPIFY_COMMAND}"
# Doesn't pick up the environment variable?
substituteInPlace third_party/kineto/libkineto/CMakeLists.txt \
--replace-fail "\''$ENV{ROCM_SOURCE_DIR}" "${rocmtoolkit_joined}"
''
pkgs/development/python-modules/torch/source/src.nix
+71 -142
View File
@@ -4,13 +4,13 @@
fetchFromGitHub,
runCommand,
}:
assert version == "2.10.0";
assert version == "2.11.0";
rec {
src_aiter = fetchFromGitHub {
owner = "ROCm";
repo = "aiter";
rev = "01aae101b9e5e94d6c16a9514c9fb8df99c93150";
hash = "sha256-rJwKTzUi066ZRroK7eFsOPmNRRRq7VMQz0Xw7qcgtNo=";
rev = "9a469a608b2c10b7157df573a38d31e5bf4038b4";
hash = "sha256-kaX3uAkgE99puYu+ODdKjvsN+LLl1Jt95vtd5Xh0Mg8=";
};
src_asmjit = fetchFromGitHub {
owner = "asmjit";
@@ -24,12 +24,6 @@ rec {
rev = "299e5928955cc62af9968370293b916f5130916f";
hash = "sha256-iPK3qLrZL2L08XW1a7SGl7GAt5InQ5nY+Dn8hBuxSOg=";
};
src_benchmark_opentelemetry-cpp = fetchFromGitHub {
owner = "google";
repo = "benchmark";
rev = "d572f4777349d43653b21d6c2fc63020ab326db2";
hash = "sha256-gg3g/0Ki29FnGqKv9lDTs5oA9NjH23qQ+hTdVtSU+zo=";
};
src_benchmark_protobuf = fetchFromGitHub {
owner = "google";
repo = "benchmark";
@@ -37,12 +31,6 @@ rec {
hash = "sha256-iFRgjLkftuszAqBnmS9GXU8BwYnabmwMAQyw19sfjb4=";
};
src_civetweb = fetchFromGitHub {
owner = "civetweb";
repo = "civetweb";
rev = "eefb26f82b233268fc98577d265352720d477ba4";
hash = "sha256-Qh6BGPk7a01YzCeX42+Og9M+fjXRs7kzNUCyT4mYab4=";
};
src_civetweb_prometheus-cpp_dynolog = fetchFromGitHub {
owner = "civetweb";
repo = "civetweb";
rev = "d7ba35bbb649209c66e582d5a0244ba988a15159";
@@ -57,26 +45,26 @@ rec {
src_composable_kernel = fetchFromGitHub {
owner = "ROCm";
repo = "composable_kernel";
rev = "7fe50dc3da2069d6645d9deb8c017a876472a977";
hash = "sha256-OxA0ekcaRxAmBFlXkvS7XAX40kcWCwyytHWV6vROWjo=";
rev = "fcc9372c009c8e0a23fece77b582da83b04a654f";
hash = "sha256-Xwj48Ftwqlea5ZIP7q7cRh2U2tlHTd1cdW4TYf5J0Dg=";
};
src_composable_kernel_aiter = fetchFromGitHub {
src_composable_kernel_fbgemm_MSLK = fetchFromGitHub {
owner = "ROCm";
repo = "composable_kernel";
rev = "cffe8fa2a442ac8e80dd236a1a5d24fe3d7e0cbf";
hash = "sha256-KDmSs9NDFYbyE4wzBedVDBZGhI1BAoJHWOStFkuEk9U=";
rev = "7fe50dc3da2069d6645d9deb8c017a876472a977";
hash = "sha256-OxA0ekcaRxAmBFlXkvS7XAX40kcWCwyytHWV6vROWjo=";
};
src_composable_kernel_flash-attention = fetchFromGitHub {
owner = "ROCm";
repo = "composable_kernel";
rev = "888317e698e9803c62bd38568abc9e05d7709f33";
hash = "sha256-0FAiGf7AErBzGxhOFcLQ4ceshqp0Esqo4ee7NsjfJUo=";
rev = "13f6d635653bd5ffbfcac8577f1ef09590c23d78";
hash = "sha256-nS1Apx4kLTIz7U2/X1BVQHiBwa5j59VboaibOhH9ADM=";
};
src_cpp-httplib = fetchFromGitHub {
owner = "yhirose";
repo = "cpp-httplib";
rev = "89c932f313c6437c38f2982869beacc89c2f2246";
hash = "sha256-+VPebnFMGNyChM20q4Z+kVOyI/qDLQjRsaGS0vo8kDM=";
rev = "bd95e67c234930cd6d6bb11309588c5462c63cec";
hash = "sha256-5q77ersAJnPPpVChvntnqEly1/ek2KfX2iukTPUbKHc=";
};
src_cpr = fetchFromGitHub {
owner = "libcpr";
@@ -93,20 +81,20 @@ rec {
src_cpuinfo_fbgemm = fetchFromGitHub {
owner = "pytorch";
repo = "cpuinfo";
rev = "6543fec09b2f04ac4a666882998b534afc9c1349";
hash = "sha256-M33853TIMNTJwBjxIG8qeTPwXlX8b8uKiCrdnlAtAA8=";
rev = "161a9ec374884f4b3e85725cb22e05f9458fdc93";
hash = "sha256-uzo6QpNfzTcqOpDse14e2OoxNyKDU8jSx+/wPLxmpJg=";
};
src_cudnn-frontend = fetchFromGitHub {
owner = "NVIDIA";
repo = "cudnn-frontend";
rev = "0258951d4d512f4714eb1574496f4d57669b1b93";
hash = "sha256-SxdxVpUNTfjptcBcezeu4VglWFSvLfAqL492ETJzdwU=";
rev = "b8c0656e6f6c84fc194f4d57329b55d609eff596";
hash = "sha256-G1WYxRCsg67umzOZ9W+JwXV6hfl5n2wtsH9KxVUccTU=";
};
src_cutlass = fetchFromGitHub {
owner = "NVIDIA";
repo = "cutlass";
rev = "f88806b1e31dfa579842638740216dd41fc6c588";
hash = "sha256-2+jUA2M6MmUAAPUYNxdgI464Z48/cIgTi4fOBvf4RZA=";
rev = "0d2b201e8c1c4a03efa6e9c468161916e2334725";
hash = "sha256-2xgdS2P2tEI/4qcZv9qjCYHFbAcVayMdCDeJfIiQN4U=";
};
src_cutlass_fbgemm = fetchFromGitHub {
owner = "jwfromm";
@@ -117,8 +105,14 @@ rec {
src_cutlass_flash-attention = fetchFromGitHub {
owner = "NVIDIA";
repo = "cutlass";
rev = "c506e16788cb08416a4a57e11a9067beeee29420";
hash = "sha256-P/BgJ7MdNTzNgsqoTjxrOQwvKfhAS+rmBw3sC/SdO1I=";
rev = "7127592069c2fe01b041e174ba4345ef9b279671";
hash = "sha256-/fEfuriQbrjjLP+yRjeo88SgW3IurdlU+6rR9+w5woQ=";
};
src_cutlass_MSLK = fetchFromGitHub {
owner = "jwfromm";
repo = "cutlass";
rev = "571edeb2d0ac872a8392fc49285b156b07884b4e";
hash = "sha256-EnEtWPJqJBLGOk93HdUL45NFqIVG5qetJX6vnc7K6pE=";
};
src_DCGM = fetchFromGitHub {
owner = "NVIDIA";
@@ -135,8 +129,8 @@ rec {
src_fbgemm = fetchFromGitHub {
owner = "pytorch";
repo = "fbgemm";
rev = "a4a112c26ac01c78a86317cf54c69481478d95d2";
hash = "sha256-Iy06EIWQ0ThLfO1xmkn3xaO3qWPhhbthff3JxL6P748=";
rev = "c246916f9e3804eacc3c95058e51cce02ae00fff";
hash = "sha256-jNc9Z3fe4pUTP5FY3sV1WINoEEd8te6tTyjNsFWZFxY=";
};
src_fbjni = fetchFromGitHub {
owner = "facebookincubator";
@@ -147,8 +141,8 @@ rec {
src_flash-attention = fetchFromGitHub {
owner = "Dao-AILab";
repo = "flash-attention";
rev = "979702c87a8713a8e0a5e9fee122b90d2ef13be5";
hash = "sha256-7jk8TD+EkkOPrVT5Nzoc+zoZ4jH/C1AtjLcP12hBrp0=";
rev = "e2743ab5b3803bb672b16437ba98a3b1d4576c50";
hash = "sha256-ft3jPiKZDHzZvkGPI34l8/Hq9Rf2f6UjDPKGU2IYz+E=";
};
src_flatbuffers = fetchFromGitHub {
owner = "google";
@@ -213,8 +207,8 @@ rec {
src_gloo = fetchFromGitHub {
owner = "pytorch";
repo = "gloo";
rev = "54cbae0d3a67fa890b4c3d9ee162b7860315e341";
hash = "sha256-4g/AffVyU7iEHJI0KRkvJqeHSBMRB89V4/cfVf3yPf4=";
rev = "3135b0b41b67dde590eef0938a0bf3d6238df5f7";
hash = "sha256-lCKZyH8p54UaSDGRTxM6ZMorc+japeJulv7FQn0GnHc=";
};
src_googletest = fetchFromGitHub {
owner = "google";
@@ -222,13 +216,7 @@ rec {
rev = "52eb8108c5bdec04579160ae17225d66034bd723";
hash = "sha256-HIHMxAUR4bjmFLoltJeIAVSulVQ6kVuIT2Ku+lwAx/4=";
};
src_googletest_opentelemetry-cpp = fetchFromGitHub {
owner = "google";
repo = "googletest";
rev = "b796f7d44681514f58a683a3a71ff17c94edb0c1";
hash = "sha256-LVLEn+e7c8013pwiLzJiiIObyrlbBHYaioO/SWbItPQ=";
};
src_googletest_prometheus-cpp_prometheus-cpp_dynolog = fetchFromGitHub {
src_googletest_prometheus-cpp = fetchFromGitHub {
owner = "google";
repo = "googletest";
rev = "e2239ee6043f73722e7aa812a459f54a28552929";
@@ -246,12 +234,6 @@ rec {
rev = "aee0f9d9b5b87796ee8a0ab26b7587ec30e8858e";
hash = "sha256-L2HR+QTQmagk92JiuW3TRx47so33xQvewdeYL1ipUPs=";
};
src_GSL = fetchFromGitHub {
owner = "microsoft";
repo = "GSL";
rev = "6f4529395c5b7c2d661812257cd6780c67e54afa";
hash = "sha256-sNTDH1ohz+rcnBvA5KkarHKdRMQPW0c2LeSVPdEYx6Q=";
};
src_hipify_torch = fetchFromGitHub {
owner = "ROCmSoftwarePlatform";
repo = "hipify_torch";
@@ -261,14 +243,14 @@ rec {
src_ideep = fetchFromGitHub {
owner = "intel";
repo = "ideep";
rev = "719d8e6cd7f7a0e01b155657526d693acf97c2b3";
hash = "sha256-WX1lluqYeoB2IuwJX1vqZ6nr8W8KsWxUDIvKNA3CAxo=";
rev = "8e7ddd65df95f13e41f0a40c820c5f35ae4a0ea3";
hash = "sha256-AVSsugGYiQ4QOWMVaHj1hzlPTZmg65yrGMmrWytvUuM=";
};
src_ittapi = fetchFromGitHub {
owner = "intel";
repo = "ittapi";
rev = "dec1d23ca65ab069d225dfe40dea14f455170959";
hash = "sha256-aP95qtYi1s8go6HbXW6hDd0XzdWQhOwXLJBNPCIkSCI=";
rev = "0c57540822deb5dae43bef6c1cc9b3be4772a033";
hash = "sha256-v6efQEMW1r5fsjOIpJQQPoau6sina/iKxAY1cfEUZQc=";
};
src_json = fetchFromGitHub {
owner = "nlohmann";
@@ -288,17 +270,11 @@ rec {
rev = "9cca280a4d0ccf0c08f47a99aa71d1b0e52f8d03";
hash = "sha256-7F0Jon+1oWL7uqet5i1IgHX0fUw/+z0QwEcA3zs5xHg=";
};
src_json_opentelemetry-cpp = fetchFromGitHub {
owner = "nlohmann";
repo = "json";
rev = "bc889afb4c5bf1c0d8ee29ef35eaaf4c8bef8a5d";
hash = "sha256-SUdhIV7tjtacf5DkoWk9cnkfyMlrkg8ZU7XnPZd22Tw=";
};
src_kineto = fetchFromGitHub {
owner = "pytorch";
repo = "kineto";
rev = "31f85df8fbd89c188f14ef10f1ec65379786b943";
hash = "sha256-4H0T8le73QX17kThG9gJEIvCSbRg/WQOF2TDoaXstEQ=";
rev = "7a731b6ae01cfc2b1fc75d83a91f84e682e43fd7";
hash = "sha256-kKASzILEvFhXDWDqBiNh21VxhtdT506NYoEDVrtVcGU=";
};
src_kleidiai = fetchFromGitHub {
owner = "ARM-software";
@@ -321,14 +297,20 @@ rec {
src_mimalloc = fetchFromGitHub {
owner = "microsoft";
repo = "mimalloc";
rev = "fbd8b99c2b828428947d70fdc046bb55609be93e";
hash = "sha256-+8xZT+mVEqlqabQc+1buVH/X6FZxvCd0rWMyjPu9i4o=";
rev = "048d969a1c5ee2fb89c226298f41ea38445546ef";
hash = "sha256-ZxIkzMIqwYZpiqbPAHDrO26xaiaF77Dlosbw43VkOpc=";
};
src_mkl-dnn = fetchFromGitHub {
owner = "intel";
repo = "mkl-dnn";
rev = "8d263e693366ef8db40acc569cc7d8edf644556d";
hash = "sha256-+4z5l0mJsw0SOW245GfZh41mdHGZ8u+xED7afm6pQjs=";
rev = "f1d471933dc852f956fd05389f9313c7148783d5";
hash = "sha256-/e57voLBNun/2koTF3sEb0Z/nDjCwq9NJVk7TaTSvMY=";
};
src_MSLK = fetchFromGitHub {
owner = "meta-pytorch";
repo = "MSLK";
rev = "3d332d1c0c0ac7765852c97b3979c9ef913e037f";
hash = "sha256-iuwAI8ko4yzifjoqKLxtz6UFOAOoWhsw4+3Unkiv6aE=";
};
src_NNPACK = fetchFromGitHub {
owner = "Maratyszcza";
@@ -348,24 +330,6 @@ rec {
rev = "e709452ef2bbc1d113faf678c24e6d3467696e83";
hash = "sha256-UhtF+CWuyv5/Pq/5agLL4Y95YNP63W2BraprhRqJOag=";
};
src_opentelemetry-cpp = fetchFromGitHub {
owner = "open-telemetry";
repo = "opentelemetry-cpp";
rev = "a799f4aed9c94b765dcdaabaeab7d5e7e2310878";
hash = "sha256-jLRUpB9aDvxsc7B42b08vN2rygN/ycgOyt78i2Hms0Q=";
};
src_opentelemetry-proto = fetchFromGitHub {
owner = "open-telemetry";
repo = "opentelemetry-proto";
rev = "4ca4f0335c63cda7ab31ea7ed70d6553aee14dce";
hash = "sha256-A14YrqvBAEBBPzvxcNVY2sJok+54/mHKNQPRaf9QLzs=";
};
src_opentracing-cpp = fetchFromGitHub {
owner = "opentracing";
repo = "opentracing-cpp";
rev = "06b57f48ded1fa3bdd3d4346f6ef29e40e08eaf5";
hash = "sha256-XlQi26ynXKDwA86DwsDw+hhKR8bcdnrtFH1CpAzVlLs=";
};
src_PeachPy = fetchFromGitHub {
owner = "malfet";
repo = "PeachPy";
@@ -385,12 +349,6 @@ rec {
hash = "sha256-Fu786IHiU6Bl66gZ/UJmqOROjlya3viLyzOxwdZVi9c=";
};
src_prometheus-cpp = fetchFromGitHub {
owner = "jupp0r";
repo = "prometheus-cpp";
rev = "c9ffcdda9086ffd9e1283ea7a0276d831f3c8a8d";
hash = "sha256-qx6oBxd0YrUyFq+7ArnKBqOwrl5X8RS9nErhRDUJ7+8=";
};
src_prometheus-cpp_dynolog = fetchFromGitHub {
owner = "jupp0r";
repo = "prometheus-cpp";
rev = "b1234816facfdda29845c46696a02998a4af115a";
@@ -435,8 +393,8 @@ rec {
src_pytorch = fetchFromGitHub {
owner = "pytorch";
repo = "pytorch";
rev = "v2.10.0";
hash = "sha256-3c/1lrKgAVD9lCFGYv8tilg1cL7Z64lkfiOG6kIs/BY=";
rev = "v2.11.0";
hash = "sha256-eA/pRQiibLJCKDUkMvmGq4suhrz37i0x0Lc/sH3Ag8E=";
};
src_sleef = fetchFromGitHub {
owner = "shibatch";
@@ -450,12 +408,6 @@ rec {
rev = "2b4cd91092d335a697416b2a3cb398283246849d";
hash = "sha256-ZidonG6q621rbdRrlW6ad7WdH0os81GNBBuPE5kQEsU=";
};
src_vcpkg = fetchFromGitHub {
owner = "Microsoft";
repo = "vcpkg";
rev = "8eb57355a4ffb410a2e94c07b4dca2dffbee8e50";
hash = "sha256-u+4vyOphnowoaZgfkCbzF7Q4tuz2GN1bHylaKw352Lc=";
};
src_VulkanMemoryAllocator = fetchFromGitHub {
owner = "GPUOpen-LibrariesAndSDKs";
repo = "VulkanMemoryAllocator";
@@ -471,17 +423,15 @@ rec {
src_aiter_recursive = runCommand "aiter" { } ''
cp -r ${src_aiter} $out
chmod u+w $out/3rdparty/composable_kernel
cp -r ${src_composable_kernel_aiter_recursive}/* $out/3rdparty/composable_kernel
cp -r ${src_composable_kernel_recursive}/* $out/3rdparty/composable_kernel
'';
src_asmjit_recursive = src_asmjit;
src_benchmark_recursive = src_benchmark;
src_benchmark_opentelemetry-cpp_recursive = src_benchmark_opentelemetry-cpp;
src_benchmark_protobuf_recursive = src_benchmark_protobuf;
src_civetweb_recursive = src_civetweb;
src_civetweb_prometheus-cpp_dynolog_recursive = src_civetweb_prometheus-cpp_dynolog;
src_clang-cindex-python3_recursive = src_clang-cindex-python3;
src_composable_kernel_recursive = src_composable_kernel;
src_composable_kernel_aiter_recursive = src_composable_kernel_aiter;
src_composable_kernel_fbgemm_MSLK_recursive = src_composable_kernel_fbgemm_MSLK;
src_composable_kernel_flash-attention_recursive = src_composable_kernel_flash-attention;
src_cpp-httplib_recursive = src_cpp-httplib;
src_cpr_recursive = src_cpr;
@@ -491,6 +441,7 @@ rec {
src_cutlass_recursive = src_cutlass;
src_cutlass_fbgemm_recursive = src_cutlass_fbgemm;
src_cutlass_flash-attention_recursive = src_cutlass_flash-attention;
src_cutlass_MSLK_recursive = src_cutlass_MSLK;
src_DCGM_recursive = src_DCGM;
src_dynolog_recursive = runCommand "dynolog" { } ''
cp -r ${src_dynolog} $out
@@ -511,14 +462,14 @@ rec {
chmod u+w $out/third_party/pfs
cp -r ${src_pfs_recursive}/* $out/third_party/pfs
chmod u+w $out/third_party/prometheus-cpp
cp -r ${src_prometheus-cpp_dynolog_recursive}/* $out/third_party/prometheus-cpp
cp -r ${src_prometheus-cpp_recursive}/* $out/third_party/prometheus-cpp
'';
src_fbgemm_recursive = runCommand "fbgemm" { } ''
cp -r ${src_fbgemm} $out
chmod u+w $out/external/asmjit
cp -r ${src_asmjit_recursive}/* $out/external/asmjit
chmod u+w $out/external/composable_kernel
cp -r ${src_composable_kernel_recursive}/* $out/external/composable_kernel
cp -r ${src_composable_kernel_fbgemm_MSLK_recursive}/* $out/external/composable_kernel
chmod u+w $out/external/cpuinfo
cp -r ${src_cpuinfo_fbgemm_recursive}/* $out/external/cpuinfo
chmod u+w $out/external/cutlass
@@ -554,12 +505,9 @@ rec {
src_glog_recursive = src_glog;
src_gloo_recursive = src_gloo;
src_googletest_recursive = src_googletest;
src_googletest_opentelemetry-cpp_recursive = src_googletest_opentelemetry-cpp;
src_googletest_prometheus-cpp_prometheus-cpp_dynolog_recursive =
src_googletest_prometheus-cpp_prometheus-cpp_dynolog;
src_googletest_prometheus-cpp_recursive = src_googletest_prometheus-cpp;
src_googletest_protobuf_recursive = src_googletest_protobuf;
src_googletest_tensorpipe_recursive = src_googletest_tensorpipe;
src_GSL_recursive = src_GSL;
src_hipify_torch_recursive = src_hipify_torch;
src_ideep_recursive = runCommand "ideep" { } ''
cp -r ${src_ideep} $out
@@ -570,7 +518,6 @@ rec {
src_json_recursive = src_json;
src_json_dynolog_recursive = src_json_dynolog;
src_json_fbgemm_recursive = src_json_fbgemm;
src_json_opentelemetry-cpp_recursive = src_json_opentelemetry-cpp;
src_kineto_recursive = runCommand "kineto" { } ''
cp -r ${src_kineto} $out
chmod u+w $out/libkineto/third_party/dynolog
@@ -585,6 +532,17 @@ rec {
src_libuv_recursive = src_libuv;
src_mimalloc_recursive = src_mimalloc;
src_mkl-dnn_recursive = src_mkl-dnn;
src_MSLK_recursive = runCommand "MSLK" { } ''
cp -r ${src_MSLK} $out
chmod u+w $out/external/composable_kernel
cp -r ${src_composable_kernel_fbgemm_MSLK_recursive}/* $out/external/composable_kernel
chmod u+w $out/external/cutlass
cp -r ${src_cutlass_MSLK_recursive}/* $out/external/cutlass
chmod u+w $out/external/googletest
cp -r ${src_googletest_recursive}/* $out/external/googletest
chmod u+w $out/external/hipify_torch
cp -r ${src_hipify_torch_recursive}/* $out/external/hipify_torch
'';
src_NNPACK_recursive = src_NNPACK;
src_NVTX_recursive = src_NVTX;
src_onnx_recursive = runCommand "onnx" { } ''
@@ -592,27 +550,6 @@ rec {
chmod u+w $out/third_party/pybind11
cp -r ${src_pybind11_onnx_recursive}/* $out/third_party/pybind11
'';
src_opentelemetry-cpp_recursive = runCommand "opentelemetry-cpp" { } ''
cp -r ${src_opentelemetry-cpp} $out
chmod u+w $out/third_party/benchmark
cp -r ${src_benchmark_opentelemetry-cpp_recursive}/* $out/third_party/benchmark
chmod u+w $out/third_party/googletest
cp -r ${src_googletest_opentelemetry-cpp_recursive}/* $out/third_party/googletest
chmod u+w $out/third_party/ms-gsl
cp -r ${src_GSL_recursive}/* $out/third_party/ms-gsl
chmod u+w $out/third_party/nlohmann-json
cp -r ${src_json_opentelemetry-cpp_recursive}/* $out/third_party/nlohmann-json
chmod u+w $out/third_party/opentelemetry-proto
cp -r ${src_opentelemetry-proto_recursive}/* $out/third_party/opentelemetry-proto
chmod u+w $out/third_party/opentracing-cpp
cp -r ${src_opentracing-cpp_recursive}/* $out/third_party/opentracing-cpp
chmod u+w $out/third_party/prometheus-cpp
cp -r ${src_prometheus-cpp_recursive}/* $out/third_party/prometheus-cpp
chmod u+w $out/tools/vcpkg
cp -r ${src_vcpkg_recursive}/* $out/tools/vcpkg
'';
src_opentelemetry-proto_recursive = src_opentelemetry-proto;
src_opentracing-cpp_recursive = src_opentracing-cpp;
src_PeachPy_recursive = src_PeachPy;
src_pfs_recursive = src_pfs;
src_pocketfft_recursive = src_pocketfft;
@@ -621,14 +558,7 @@ rec {
chmod u+w $out/3rdparty/civetweb
cp -r ${src_civetweb_recursive}/* $out/3rdparty/civetweb
chmod u+w $out/3rdparty/googletest
cp -r ${src_googletest_prometheus-cpp_prometheus-cpp_dynolog_recursive}/* $out/3rdparty/googletest
'';
src_prometheus-cpp_dynolog_recursive = runCommand "prometheus-cpp_dynolog" { } ''
cp -r ${src_prometheus-cpp_dynolog} $out
chmod u+w $out/3rdparty/civetweb
cp -r ${src_civetweb_prometheus-cpp_dynolog_recursive}/* $out/3rdparty/civetweb
chmod u+w $out/3rdparty/googletest
cp -r ${src_googletest_prometheus-cpp_prometheus-cpp_dynolog_recursive}/* $out/3rdparty/googletest
cp -r ${src_googletest_prometheus-cpp_recursive}/* $out/3rdparty/googletest
'';
src_protobuf_recursive = runCommand "protobuf" { } ''
cp -r ${src_protobuf} $out
@@ -692,6 +622,8 @@ rec {
cp -r ${src_kleidiai_recursive}/* $out/third_party/kleidiai
chmod u+w $out/third_party/mimalloc
cp -r ${src_mimalloc_recursive}/* $out/third_party/mimalloc
chmod u+w $out/third_party/mslk
cp -r ${src_MSLK_recursive}/* $out/third_party/mslk
chmod u+w $out/third_party/nlohmann
cp -r ${src_json_recursive}/* $out/third_party/nlohmann
chmod u+w $out/third_party/NNPACK
@@ -700,8 +632,6 @@ rec {
cp -r ${src_NVTX_recursive}/* $out/third_party/NVTX
chmod u+w $out/third_party/onnx
cp -r ${src_onnx_recursive}/* $out/third_party/onnx
chmod u+w $out/third_party/opentelemetry-cpp
cp -r ${src_opentelemetry-cpp_recursive}/* $out/third_party/opentelemetry-cpp
chmod u+w $out/third_party/pocketfft
cp -r ${src_pocketfft_recursive}/* $out/third_party/pocketfft
chmod u+w $out/third_party/protobuf
@@ -735,7 +665,6 @@ rec {
chmod u+w $out/third_party/pybind11
cp -r ${src_pybind11_tensorpipe_recursive}/* $out/third_party/pybind11
'';
src_vcpkg_recursive = src_vcpkg;
src_VulkanMemoryAllocator_recursive = src_VulkanMemoryAllocator;
src_XNNPACK_recursive = src_XNNPACK;
}
pkgs/development/python-modules/torchaudio/default.nix
+110 -56
View File
@@ -3,88 +3,142 @@
buildPythonPackage,
fetchFromGitHub,
# nativeBuildInputs
cmake,
pkg-config,
ninja,
# buildInputs
pybind11,
sox,
torch,
llvmPackages,
# build-system
setuptools,
# dependencies
torch,
torchcodec,
# tests
inflect,
parameterized,
pytestCheckHook,
pytorch-lightning,
scipy,
sentencepiece,
unidecode,
# passthru
torchaudio,
cudaSupport ? torch.cudaSupport,
cudaPackages,
rocmSupport ? torch.rocmSupport,
rocmPackages,
}:
buildPythonPackage.override { stdenv = torch.stdenv; } (finalAttrs: {
buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
pname = "torchaudio";
version = "2.10.0";
version = "2.11.0";
pyproject = true;
src = fetchFromGitHub {
owner = "pytorch";
repo = "audio";
tag = "v${finalAttrs.version}";
hash = "sha256-b1sjHVFXdNFDbdtXWSM2KisSRE/8IbzJI4rvzYQ4UMg=";
hash = "sha256-TncROn9wfn5HOaIvupS2/KD9JCgwfHyfnbZRc+SiqJ0=";
};
patches = [
./0001-setup.py-propagate-cmakeFlags.patch
];
postPatch = lib.optionalString rocmSupport ''
# There is no .info/version-dev, only .info/version
substituteInPlace cmake/LoadHIP.cmake \
--replace-fail "/.info/version-dev" "/.info/version"
'';
env = {
# CTC seems to be missing no matter what. No obvious flag to force its compilation.
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_CTC_DECODER = 1;
# CUDA
USE_CUDA = cudaSupport;
TORCH_CUDA_ARCH_LIST = "${lib.concatStringsSep ";" torch.cudaCapabilities}";
BUILD_SOX = 0;
BUILD_KALDI = 0;
BUILD_RNNT = 0;
BUILD_CTC_DECODER = 0;
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_CUCTC_DECODER = 1;
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_CUDA = 1;
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_MULTIGPU_CUDA = 1;
# At of 2026-03-27, the default `cudaPackages` version is 12.9.1
# According to Nvidia, it should support GCC versions up to 14.x:
# -> https://docs.nvidia.com/cuda/archive/12.9.1/cuda-installation-guide-linux/index.html#host-compiler-support-policy
# However, PyTorch's *strict* upper bound is 14.0:
# -> https://github.com/pytorch/pytorch/blob/v2.11.0/torch/utils/cpp_extension.py#L75
# Hence, the build fails with:
# RuntimeError: The current installed version of g++ (14.3.0) is greater than the maximum
# required version by CUDA 12.9. Please make sure to use an adequate version of g++
# (>=6.0.0, <14.0).
# Hence, we disable the version check to silence the error:
TORCH_DONT_CHECK_COMPILER_ABI = cudaSupport;
# ROCM
USE_ROCM = rocmSupport;
PYTORCH_ROCM_ARCH = lib.optionalString rocmSupport torch.gpuTargetString;
ROCM_PATH = lib.optionalString rocmSupport torch.rocmtoolkit_joined;
# demucs is not packaged in nixpkgs and is archived anyway
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_MOD_demucs = true;
# fairseq is unmaintained and broken in nixpkgs
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_MOD_fairseq = true;
# Fails even on python>3.10 with:
# RuntimeError: Test is known to fail for Python 3.10, disabling for now
# See: https://github.com/pytorch/audio/pull/2224#issuecomment-1048329450
TORCHAUDIO_TEST_ALLOW_SKIP_IF_ON_PYTHON_310 = true;
};
nativeBuildInputs = [
cmake
pkg-config
ninja
]
++ lib.optionals cudaSupport [ cudaPackages.cuda_nvcc ]
++ lib.optionals rocmSupport (
with rocmPackages;
[
clr
rocblas
hipblas
build-system = [
setuptools
];
nativeBuildInputs = lib.optionals cudaSupport [
cudaPackages.cuda_nvcc
];
buildInputs =
lib.optionals cudaSupport [
cudaPackages.cuda_cudart
]
);
++ lib.optionals torch.stdenv.cc.isClang [
llvmPackages.openmp
];
buildInputs = [
pybind11
sox
torch.cxxdev
]
++ lib.optionals torch.stdenv.cc.isClang [ llvmPackages.openmp ];
dependencies = [ torch ];
preConfigure = lib.optionalString rocmSupport ''
export ROCM_PATH=${torch.rocmtoolkit_joined}
export PYTORCH_ROCM_ARCH="${torch.gpuTargetString}"
'';
dontUseCmakeConfigure = true;
doCheck = false; # requires sox backend
dependencies = [
torch
torchcodec
];
pythonImportsCheck = [ "torchaudio" ];
nativeCheckInputs = [
inflect
parameterized
pytestCheckHook
pytorch-lightning
scipy
sentencepiece
unidecode
];
disabledTestPaths = [
# Require internet access
"test/integration_tests"
];
disabledTests = [
# AssertionError: Tensor-likes are not close!
# Mismatched elements: 1070 / 32800 (3.3%)
"test_amplitude_to_DB"
"test_amplitude_to_DB_power"
# Very long to run
"AutogradCPUTest"
];
passthru.gpuCheck = torchaudio.overridePythonAttrs (old: {
requiredSystemFeatures = [ "cuda" ];
env = (old.env or { }) // {
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_CUCTC_DECODER = 0;
TORCHAUDIO_TEST_ALLOW_SKIP_IF_NO_CUDA = 0;
};
});
meta = {
description = "PyTorch audio library";
homepage = "https://pytorch.org/";
pkgs/development/python-modules/torchcodec/default.nix
+10 -3
View File
@@ -21,18 +21,19 @@
cudaSupport ? torch.cudaSupport,
cudaPackages,
rocmSupport ? torch.rocmSupport,
}:
buildPythonPackage (finalAttrs: {
buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
pname = "torchcodec";
version = "0.10.0";
version = "0.11.0";
pyproject = true;
src = fetchFromGitHub {
owner = "meta-pytorch";
repo = "torchcodec";
tag = "v${finalAttrs.version}";
hash = "sha256-2uo2HtA5hh8s7F1Z8SS78bMwyOvw2H2j4p0Znan552I=";
hash = "sha256-VKpwKV+B2e5z1V1XAdvk6K4/4C3ISPsSiwZ05w5XTBU=";
};
postPatch = ''
@@ -93,6 +94,12 @@ buildPythonPackage (finalAttrs: {
I_CONFIRM_THIS_IS_NOT_A_LICENSE_VIOLATION = true;
ENABLE_CUDA = cudaSupport;
}
// lib.optionalAttrs rocmSupport {
ROCM_PATH = torch.rocmtoolkit_joined;
ROCM_SOURCE_DIR = torch.rocmtoolkit_joined;
PYTORCH_ROCM_ARCH = torch.gpuTargetString;
CMAKE_CXX_FLAGS = "-I${torch.rocmtoolkit_joined}/include";
};
pythonImportsCheck = [ "torchcodec" ];
pkgs/development/python-modules/torchvision/default.nix
+15 -3
View File
@@ -26,16 +26,16 @@ let
inherit (torch) cudaCapabilities cudaPackages cudaSupport;
in
buildPythonPackage.override { stdenv = torch.stdenv; } (finalAttrs: {
buildPythonPackage.override { inherit (torch) stdenv; } (finalAttrs: {
pname = "torchvision";
version = "0.25.0";
version = "0.26.0";
pyproject = true;
src = fetchFromGitHub {
owner = "pytorch";
repo = "vision";
tag = "v${finalAttrs.version}";
hash = "sha256-oktJHcT6T4f58pUO+HSBpbyS1ISH3zDlTsXQh6PcMy4=";
hash = "sha256-FOdDGY3v8yWBhtNo9tZP79/xwrc7AoIY5Y1ZABzWe6g=";
};
nativeBuildInputs = [
@@ -63,6 +63,18 @@ buildPythonPackage.override { stdenv = torch.stdenv; } (finalAttrs: {
TORCHVISION_LIBRARY = "${libjpeg_turbo}/lib/";
}
// lib.optionalAttrs cudaSupport {
# At of 2026-03-27, the default `cudaPackages` version is 12.9.1
# According to Nvidia, it should support GCC versions up to 14.x:
# -> https://docs.nvidia.com/cuda/archive/12.9.1/cuda-installation-guide-linux/index.html#host-compiler-support-policy
# However, PyTorch's *strict* upper bound is 14.0:
# -> https://github.com/pytorch/pytorch/blob/v2.11.0/torch/utils/cpp_extension.py#L75
# Hence, the build fails with:
# RuntimeError: The current installed version of g++ (14.3.0) is greater than the maximum
# required version by CUDA 12.9. Please make sure to use an adequate version of g++
# (>=6.0.0, <14.0).
# Hence, we disable the version check to silence the error:
TORCH_DONT_CHECK_COMPILER_ABI = 1;
TORCH_CUDA_ARCH_LIST = "${lib.concatStringsSep ";" cudaCapabilities}";
FORCE_CUDA = 1;
};