~ruther/guix-local

a0a9044d356295be45841add1c239d377ebb3ddf — David Elsing 1 year, 4 months ago 9a6221e 
gnu: python-pytorch: Update to 2.5.1.

This also updates the qnnpack-pytorch package.

* gnu/packages/machine-learning.scm (%python-pytorch-version): Set to 2.5.1.
(%python-pytorch-src): Adjust hash accordingly.
(python-pytorch)[arguments]<#:phases>: Add 'remove-fr-trace-script
phase. Adjust 'use-system-libraries phase.
[native-inputs]: Add nlohmann-json. Replace cmake with cmake-minimal.
[inputs]: Remove foxi.
* gnu/packages/patches/python-pytorch-fix-codegen.patch: Adjust patch.
* gnu/packages/patches/python-pytorch-system-libraries.patch: Adjust patch.

Signed-off-by: Ludovic Courtès <ludo@gnu.org>
patch browse .tar.gz 
3 files changed, 90 insertions(+), 73 deletions(-)

M gnu/packages/machine-learning.scm
M gnu/packages/patches/python-pytorch-fix-codegen.patch
M gnu/packages/patches/python-pytorch-system-libraries.patch

M gnu/packages/machine-learning.scm => gnu/packages/machine-learning.scm +16 -5
@@ 4670,7 4670,7 @@ PyTorch.")
        (base32
         "0hdpkhcjry22fjx2zg2r48v7f4ljrclzj0li2pgk76kvyblfbyvm"))))))

(define %python-pytorch-version "2.4.0")
(define %python-pytorch-version "2.5.1")

(define %python-pytorch-src
  (origin


@@ 4681,7 4681,7 @@ PyTorch.")
    (file-name (git-file-name "python-pytorch" %python-pytorch-version))
    (sha256
     (base32
      "18hdhzr12brj0b7ppyiscax0dbra30207qx0cckw78midfkcn7cn"))
      "052cvagpmm9y7jspjpcyysx8yc5fhxnjl8rcz6nndis06v8dcj8s"))
    (patches (search-patches "python-pytorch-system-libraries.patch"
                             "python-pytorch-runpath.patch"
                             "python-pytorch-without-kineto.patch"


@@ 4827,6 4827,14 @@ PyTorch.")
                 (string-append #$output "/lib/python"
                                #$(version-major+minor (package-version python))
                                "/site-packages")))))
          ;; This entry point is broken, because it refers to a module that is
          ;; (intentionally) not installed
          ;; (https://github.com/pytorch/pytorch/pull/134729), which causes
          ;; the 'sanity-check phase to fail.
          (add-after 'unpack 'remove-fr-trace-script
            (lambda _
              (substitute* "setup.py"
                (("entry_points\\[\"console_scripts\"\\]\\.append\\(") "("))))
          (add-before 'build 'use-system-libraries
            (lambda _
              (substitute* '("caffe2/serialize/crc.cc"


@@ 4864,7 4872,10 @@ PyTorch.")
                          (or (%current-target-system)
                              (%current-system))
                          (package-transitive-supported-systems qnnpack)))
                  (setenv "USE_QNNPACK" "0"))))
                  (setenv "USE_QNNPACK" "0"))
              (substitute* '("requirements.txt" "setup.py")
                (("sympy==1\\.13\\.1")
                 "sympy>=1.13.1"))))
          ;; PyTorch is still built with AVX2 and AVX-512 support selected at
          ;; runtime, but these dependencies require it (nnpack only for
          ;; x86_64).


@@ 4965,10 4976,11 @@ PyTorch.")
      ;; Even only the core tests take a very long time to run.
      #:tests? #f))
    (native-inputs
     (list cmake
     (list cmake-minimal
           doxygen
           ideep-pytorch
           ninja
           nlohmann-json
           pocketfft-cpp
           python-expecttest
           python-pytest-flakefinder


@@ 4988,7 5000,6 @@ PyTorch.")
            eigen
            flatbuffers-next
            fmt
            foxi
            fp16
            fxdiv
            gemmlowp


M gnu/packages/patches/python-pytorch-fix-codegen.patch => gnu/packages/patches/python-pytorch-fix-codegen.patch +26 -16
@@ 6,7 6,7 @@ is later corrected.  codegen_external.py is patched to avoid duplicate
functions and add the static keyword as in the existing generated file.

diff --git a/tools/gen_flatbuffers.sh b/tools/gen_flatbuffers.sh
index cc0263d..ac34e84 100644
index cc0263dbbf..ac34e84b82 100644
--- a/tools/gen_flatbuffers.sh
+++ b/tools/gen_flatbuffers.sh
@@ -1,13 +1,13 @@


@@ 32,7 32,7 @@ index cc0263d..ac34e84 100644
      -c "$ROOT/torch/csrc/jit/serialization/mobile_bytecode.fbs"
 echo '// @generated' >> "$ROOT/torch/csrc/jit/serialization/mobile_bytecode_generated.h"
diff --git a/torch/csrc/jit/tensorexpr/codegen_external.py b/torch/csrc/jit/tensorexpr/codegen_external.py
index 5dcf1b2..0e20b0c 100644
index 5dcf1b2840..0e20b0c102 100644
--- a/torch/csrc/jit/tensorexpr/codegen_external.py
+++ b/torch/csrc/jit/tensorexpr/codegen_external.py
@@ -21,9 +21,14 @@ def gen_external(native_functions_path, tags_path, external_path):


@@ 61,7 61,7 @@ index 5dcf1b2..0e20b0c 100644
     void** buf_data,
     int64_t* buf_ranks,
diff --git a/torchgen/decompositions/gen_jit_decompositions.py b/torchgen/decompositions/gen_jit_decompositions.py
index 7a0024f..6b2445f 100644
index b42948045c..e1cfc73a5e 100644
--- a/torchgen/decompositions/gen_jit_decompositions.py
+++ b/torchgen/decompositions/gen_jit_decompositions.py
@@ -1,8 +1,12 @@


@@ 76,9 76,9 @@ index 7a0024f..6b2445f 100644
+else:
+    decomposition_table = {}
 
 # from torchgen.code_template import CodeTemplate
 
@@ -85,7 +89,7 @@ def write_decomposition_util_file(path: str) -> None:
 # from torchgen.code_template import CodeTemplate
@@ -86,7 +90,7 @@ def write_decomposition_util_file(path: str) -> None:
 
 
 def main() -> None:


@@ 88,27 88,28 @@ index 7a0024f..6b2445f 100644
     write_decomposition_util_file(str(upgrader_path))
 
diff --git a/torchgen/operator_versions/gen_mobile_upgraders.py b/torchgen/operator_versions/gen_mobile_upgraders.py
index 2907076..6866332 100644
index 362ce427d5..245056f815 100644
--- a/torchgen/operator_versions/gen_mobile_upgraders.py
+++ b/torchgen/operator_versions/gen_mobile_upgraders.py
@@ -3,10 +3,12 @@ import os
@@ -6,10 +6,13 @@ import os
 from enum import Enum
 from operator import itemgetter
 from pathlib import Path
+import sys
 from typing import Any, Dict, List
 from typing import Any
 
-import torch
-from torch.jit.generate_bytecode import generate_upgraders_bytecode
+if len(sys.argv) < 2 or sys.argv[1] != "dummy":
+    import torch
+    from torch.jit.generate_bytecode import generate_upgraders_bytecode
 
+
 from torchgen.code_template import CodeTemplate
 from torchgen.operator_versions.gen_mobile_upgraders_constant import (
@@ -263,7 +265,10 @@ def construct_register_size(register_size_from_yaml: int) -> str:
     MOBILE_UPGRADERS_HEADER_DESCRIPTION,
@@ -265,7 +268,10 @@ def construct_register_size(register_size_from_yaml: int) -> str:
 def construct_version_maps(
     upgrader_bytecode_function_to_index_map: Dict[str, Any]
     upgrader_bytecode_function_to_index_map: dict[str, Any]
 ) -> str:
-    version_map = torch._C._get_operator_version_map()
+    if len(sys.argv) < 2 or sys.argv[1] != "dummy":


@@ 118,7 119,7 @@ index 2907076..6866332 100644
     sorted_version_map_ = sorted(version_map.items(), key=itemgetter(0))  # type: ignore[no-any-return]
     sorted_version_map = dict(sorted_version_map_)
 
@@ -379,7 +384,10 @@ def sort_upgrader(upgrader_list: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
@@ -381,7 +387,10 @@ def sort_upgrader(upgrader_list: list[dict[str, Any]]) -> list[dict[str, Any]]:
 
 
 def main() -> None:


@@ 131,16 132,24 @@ index 2907076..6866332 100644
     for up in sorted_upgrader_list:
         print("after sort upgrader : ", next(iter(up)))
diff --git a/torchgen/shape_functions/gen_jit_shape_functions.py b/torchgen/shape_functions/gen_jit_shape_functions.py
index bdfd5c7..72b237a 100644
index 56a3d8bf0d..490a3ea2e7 100644
--- a/torchgen/shape_functions/gen_jit_shape_functions.py
+++ b/torchgen/shape_functions/gen_jit_shape_functions.py
@@ -18,16 +18,20 @@ you are in the root directory of the Pytorch git repo"""
@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 import os
 import sys
+import importlib
 from importlib.util import module_from_spec, spec_from_file_location
 from itertools import chain
 from pathlib import Path
@@ -18,16 +19,21 @@ you are in the root directory of the Pytorch git repo"""
 if not file_path.exists():
     raise Exception(err_msg)  # noqa: TRY002
 
-spec = importlib.util.spec_from_file_location(module_name, file_path)
-spec = spec_from_file_location(module_name, file_path)
-assert spec is not None
-module = importlib.util.module_from_spec(spec)
-module = module_from_spec(spec)
-sys.modules[module_name] = module
-assert spec.loader is not None
-assert module is not None


@@ 159,6 168,7 @@ index bdfd5c7..72b237a 100644
+
+    bounded_compute_graph_mapping = module.bounded_compute_graph_mapping
+    shape_compute_graph_mapping = module.shape_compute_graph_mapping
+
+else:
+    bounded_compute_graph_mapping = {}
+    shape_compute_graph_mapping = {}


M gnu/packages/patches/python-pytorch-system-libraries.patch => gnu/packages/patches/python-pytorch-system-libraries.patch +48 -52
@@ 5,10 5,10 @@ httlib, ideep, miniz, nnpack, oneapi-dnnl, pocketfft, pthreadpool,
qnnpack-pytorch, tensorpipe, valgrind and xnnpack packages.

diff --git a/aten/src/ATen/CMakeLists.txt b/aten/src/ATen/CMakeLists.txt
index 0087dd9..0235fa1 100644
index 6d9152a4d0..97cb945722 100644
--- a/aten/src/ATen/CMakeLists.txt
+++ b/aten/src/ATen/CMakeLists.txt
@@ -419,9 +419,9 @@ if(AT_NNPACK_ENABLED)
@@ -414,9 +414,9 @@ if(AT_NNPACK_ENABLED)
   list(APPEND ATen_CPU_DEPENDENCY_LIBS nnpack) # cpuinfo is added below
 endif()
 


@@ 22,7 22,7 @@ index 0087dd9..0235fa1 100644
 if(NOT CMAKE_SYSTEM_PROCESSOR MATCHES "^(s390x|ppc64le)$")
   list(APPEND ATen_CPU_DEPENDENCY_LIBS cpuinfo)
diff --git a/caffe2/CMakeLists.txt b/caffe2/CMakeLists.txt
index 89c31fa..e6d9ef1 100644
index 9be7f3732f..111215dacc 100644
--- a/caffe2/CMakeLists.txt
+++ b/caffe2/CMakeLists.txt
@@ -91,9 +91,6 @@ if(NOT MSVC AND USE_XNNPACK)


@@ 35,7 35,7 @@ index 89c31fa..e6d9ef1 100644
   endif()
 endif()
 
@@ -1075,7 +1072,6 @@ if(USE_XPU)
@@ -1081,7 +1078,6 @@ if(USE_XPU)
 endif()
 
 if(NOT MSVC AND USE_XNNPACK)


@@ 43,17 43,18 @@ index 89c31fa..e6d9ef1 100644
 endif()
 
 # ==========================================================
@@ -1189,6 +1185,9 @@ endif()
@@ -1195,8 +1191,8 @@ endif()
 target_include_directories(torch_cpu PRIVATE
   ${TORCH_ROOT}/third_party/cpp-httplib)
 
-target_include_directories(torch_cpu PRIVATE
-  ${TORCH_ROOT}/third_party/nlohmann/include)
+find_package(httplib REQUIRED)
+target_link_libraries(torch_cpu PUBLIC httplib::httplib)
+
 
 install(DIRECTORY "${TORCH_SRC_DIR}/csrc"
   DESTINATION ${TORCH_INSTALL_INCLUDE_DIR}/torch
   FILES_MATCHING PATTERN "*.h" PATTERN "*.hpp")
@@ -1417,6 +1416,7 @@ target_link_libraries(torch_cpu PUBLIC c10)
@@ -1425,6 +1421,7 @@ target_link_libraries(torch_cpu PUBLIC c10)
 target_link_libraries(torch_cpu PUBLIC ${Caffe2_PUBLIC_DEPENDENCY_LIBS})
 target_link_libraries(torch_cpu PRIVATE ${Caffe2_DEPENDENCY_LIBS})
 target_link_libraries(torch_cpu PRIVATE ${Caffe2_DEPENDENCY_WHOLE_LINK_LIBS})


@@ 61,7 62,7 @@ index 89c31fa..e6d9ef1 100644
 if(USE_MPI)
   target_link_libraries(torch_cpu PRIVATE MPI::MPI_CXX)
 endif()
@@ -1694,7 +1694,7 @@ if(BUILD_STATIC_RUNTIME_BENCHMARK)
@@ -1659,7 +1656,7 @@ if(BUILD_STATIC_RUNTIME_BENCHMARK)
   add_executable(static_runtime_bench "${STATIC_RUNTIME_BENCHMARK_SRCS}")
   add_executable(static_runtime_test "${STATIC_RUNTIME_TEST_SRCS}")
   target_link_libraries(static_runtime_bench torch_library benchmark)


@@ 70,7 71,7 @@ index 89c31fa..e6d9ef1 100644
 endif()
 
 if(BUILD_MOBILE_BENCHMARK)
@@ -1713,7 +1713,7 @@ if(BUILD_MOBILE_TEST)
@@ -1678,7 +1675,7 @@ if(BUILD_MOBILE_TEST)
   foreach(test_src ${ATen_MOBILE_TEST_SRCS})
     get_filename_component(test_name ${test_src} NAME_WE)
     add_executable(${test_name} "${test_src}")


@@ 79,16 80,16 @@ index 89c31fa..e6d9ef1 100644
     target_include_directories(${test_name} PRIVATE $<INSTALL_INTERFACE:include>)
     target_include_directories(${test_name} PRIVATE $<BUILD_INTERFACE:${CMAKE_BINARY_DIR}/include>)
     target_include_directories(${test_name} PRIVATE ${ATen_CPU_INCLUDE})
@@ -1734,7 +1734,7 @@ if(BUILD_TEST)
@@ -1699,7 +1696,7 @@ if(BUILD_TEST)
         if(NOT MSVC)
           add_executable(${test_name}_${CPU_CAPABILITY} "${test_src}" ../aten/src/ATen/native/quantized/AffineQuantizerBase.cpp)
           # TODO: Get rid of c10 dependency (which is only needed for the implementation of AT_ERROR)
-          target_link_libraries(${test_name}_${CPU_CAPABILITY} c10 sleef gtest_main)
+          target_link_libraries(${test_name}_${CPU_CAPABILITY} c10 sleef gtest_main gtest)
-          target_link_libraries(${test_name}_${CPU_CAPABILITY} c10 sleef gtest_main nlohmann)
+          target_link_libraries(${test_name}_${CPU_CAPABILITY} c10 sleef gtest_main gtest nlohmann)
           if(USE_FBGEMM)
             target_link_libraries(${test_name}_${CPU_CAPABILITY} fbgemm)
           endif()
@@ -1748,7 +1748,7 @@ if(BUILD_TEST)
@@ -1713,7 +1710,7 @@ if(BUILD_TEST)
           endif()
         else()
           add_executable(${test_name}_${CPU_CAPABILITY} "${test_src}")


@@ 97,7 98,7 @@ index 89c31fa..e6d9ef1 100644
         endif()
         target_include_directories(${test_name}_${CPU_CAPABILITY} PRIVATE $<INSTALL_INTERFACE:include>)
         target_include_directories(${test_name}_${CPU_CAPABILITY} PRIVATE $<BUILD_INTERFACE:${CMAKE_BINARY_DIR}/include>)
@@ -1765,7 +1765,7 @@ if(BUILD_TEST)
@@ -1730,7 +1727,7 @@ if(BUILD_TEST)
   foreach(test_src ${Caffe2_CPU_TEST_SRCS})
     get_filename_component(test_name ${test_src} NAME_WE)
     add_executable(${test_name} "${test_src}")


@@ 106,7 107,7 @@ index 89c31fa..e6d9ef1 100644
     if(NOT MSVC)
       target_link_libraries(${test_name} stdc++)
     endif()
@@ -1845,7 +1845,7 @@ if(BUILD_TEST)
@@ -1810,7 +1807,7 @@ if(BUILD_TEST)
     foreach(test_src ${Caffe2_VULKAN_TEST_SRCS})
       get_filename_component(test_name ${test_src} NAME_WE)
       add_executable(${test_name} "${test_src}")


@@ 116,7 117,7 @@ index 89c31fa..e6d9ef1 100644
       target_include_directories(${test_name} PRIVATE ${Caffe2_CPU_INCLUDE})
       add_test(NAME ${test_name} COMMAND $<TARGET_FILE:${test_name}>)
diff --git a/caffe2/serialize/CMakeLists.txt b/caffe2/serialize/CMakeLists.txt
index 1552b59..67e1a9a 100644
index 1552b59d0d..67e1a9a1a3 100644
--- a/caffe2/serialize/CMakeLists.txt
+++ b/caffe2/serialize/CMakeLists.txt
@@ -2,7 +2,6 @@ file(GLOB tmp *_test.cc)


@@ 128,10 129,10 @@ index 1552b59..67e1a9a 100644
   ${CMAKE_CURRENT_SOURCE_DIR}/istream_adapter.cc
   ${CMAKE_CURRENT_SOURCE_DIR}/file_adapter.cc
diff --git a/cmake/Dependencies.cmake b/cmake/Dependencies.cmake
index f1f2eb7..b4d2033 100644
index e78305e0a8..5b3c3d7bf2 100644
--- a/cmake/Dependencies.cmake
+++ b/cmake/Dependencies.cmake
@@ -263,7 +263,7 @@ endif()
@@ -270,7 +270,7 @@ endif()
 # --- [ PocketFFT
 set(AT_POCKETFFT_ENABLED 0)
 if(NOT AT_MKL_ENABLED)


@@ 140,7 141,7 @@ index f1f2eb7..b4d2033 100644
   if(NOT EXISTS "${POCKETFFT_INCLUDE_DIR}")
     message(FATAL_ERROR "pocketfft directory not found, expected ${POCKETFFT_INCLUDE_DIR}")
   elif(NOT EXISTS "${POCKETFFT_INCLUDE_DIR}/pocketfft_hdronly.h")
@@ -458,15 +458,6 @@ if(USE_PYTORCH_QNNPACK)
@@ -465,15 +465,6 @@ if(USE_PYTORCH_QNNPACK)
       set(PYTORCH_QNNPACK_BUILD_TESTS OFF CACHE BOOL "")
       set(PYTORCH_QNNPACK_BUILD_BENCHMARKS OFF CACHE BOOL "")
       set(PYTORCH_QNNPACK_LIBRARY_TYPE "static" CACHE STRING "")


@@ 156,7 157,7 @@ index f1f2eb7..b4d2033 100644
 
       if(PYTORCH_QNNPACK_CUSTOM_THREADPOOL)
         target_compile_definitions(
@@ -653,11 +644,6 @@ if(BUILD_TEST OR BUILD_MOBILE_BENCHMARK OR BUILD_MOBILE_TEST)
@@ -660,11 +651,6 @@ if(BUILD_TEST OR BUILD_MOBILE_BENCHMARK OR BUILD_MOBILE_TEST)
   # this shouldn't be necessary anymore.
   get_property(INC_DIR_temp DIRECTORY PROPERTY INCLUDE_DIRECTORIES)
   set_property(DIRECTORY PROPERTY INCLUDE_DIRECTORIES "")


@@ 168,7 169,7 @@ index f1f2eb7..b4d2033 100644
 
   # We will not need to test benchmark lib itself.
   set(BENCHMARK_ENABLE_TESTING OFF CACHE BOOL "Disable benchmark testing as we don't need it.")
@@ -735,16 +721,6 @@ if(USE_FBGEMM)
@@ -742,16 +728,6 @@ if(USE_FBGEMM)
     if(USE_ASAN)
       set(USE_SANITIZER "address,undefined" CACHE STRING "-fsanitize options for FBGEMM")
     endif()


@@ 185,7 186,7 @@ index f1f2eb7..b4d2033 100644
   endif()
 
   if(USE_FBGEMM)
@@ -812,7 +788,7 @@ if(NOT TARGET fp16 AND NOT USE_SYSTEM_FP16)
@@ -819,7 +795,7 @@ if(NOT TARGET fp16 AND NOT USE_SYSTEM_FP16)
     "${FP16_SOURCE_DIR}"
     "${CONFU_DEPENDENCIES_BINARY_DIR}/FP16")
 elseif(NOT TARGET fp16 AND USE_SYSTEM_FP16)


@@ 194,32 195,24 @@ index f1f2eb7..b4d2033 100644
   set_target_properties(fp16 PROPERTIES LINKER_LANGUAGE C)
 endif()
 list(APPEND Caffe2_DEPENDENCY_LIBS fp16)
@@ -1159,7 +1135,6 @@ if(USE_DISTRIBUTED AND USE_TENSORPIPE)
@@ -1161,7 +1137,6 @@ if(USE_DISTRIBUTED AND USE_TENSORPIPE)
 
     # Tensorpipe uses cuda_add_library
     torch_update_find_cuda_flags()
-    add_subdirectory(${PROJECT_SOURCE_DIR}/third_party/tensorpipe)
 
     list(APPEND Caffe2_DEPENDENCY_LIBS tensorpipe)
     if(USE_CUDA)
@@ -1288,7 +1263,6 @@ if(CAFFE2_CMAKE_BUILDING_WITH_MAIN_REPO AND NOT INTERN_DISABLE_ONNX)
       set_target_properties(onnx_proto PROPERTIES CXX_STANDARD 17)
     endif()
   endif()
-  add_subdirectory(${CMAKE_CURRENT_LIST_DIR}/../third_party/foxi EXCLUDE_FROM_ALL)
 
   add_definitions(-DONNX_NAMESPACE=${ONNX_NAMESPACE})
   if(NOT USE_SYSTEM_ONNX)
@@ -1319,7 +1293,7 @@ if(CAFFE2_CMAKE_BUILDING_WITH_MAIN_REPO AND NOT INTERN_DISABLE_ONNX)
     list(APPEND Caffe2_DEPENDENCY_LIBS nlohmann)
@@ -1317,7 +1292,7 @@ if(CAFFE2_CMAKE_BUILDING_WITH_MAIN_REPO AND NOT INTERN_DISABLE_ONNX)
     endif()
     set_property(TARGET onnx_proto PROPERTY IMPORTED_LOCATION ${ONNX_PROTO_LIBRARY})
     message("-- Found onnx: ${ONNX_LIBRARY} ${ONNX_PROTO_LIBRARY}")
-    list(APPEND Caffe2_DEPENDENCY_LIBS onnx_proto onnx)
+    list(APPEND Caffe2_DEPENDENCY_LIBS onnx_proto onnx onnx_optimizer)
   endif()
   include_directories(${FOXI_INCLUDE_DIRS})
   list(APPEND Caffe2_DEPENDENCY_LIBS foxi_loader)
@@ -1476,9 +1450,8 @@ if(NOT INTERN_BUILD_MOBILE)
   # Recover the build shared libs option.
   set(BUILD_SHARED_LIBS ${TEMP_BUILD_SHARED_LIBS})
@@ -1465,9 +1440,8 @@ if(NOT INTERN_BUILD_MOBILE)
   endif()
   if(USE_MKLDNN)
     include(${CMAKE_CURRENT_LIST_DIR}/public/mkldnn.cmake)


@@ 230,7 223,7 @@ index f1f2eb7..b4d2033 100644
     else()
       message(WARNING "MKLDNN could not be found.")
       caffe2_update_option(USE_MKLDNN OFF)
@@ -1530,7 +1503,7 @@ endif()
@@ -1519,7 +1493,7 @@ endif()
 #
 set(TEMP_BUILD_SHARED_LIBS ${BUILD_SHARED_LIBS})
 set(BUILD_SHARED_LIBS OFF CACHE BOOL "Build shared libs" FORCE)


@@ 239,7 232,7 @@ index f1f2eb7..b4d2033 100644
 
 # Disable compiler feature checks for `fmt`.
 #
@@ -1539,7 +1512,6 @@ add_subdirectory(${PROJECT_SOURCE_DIR}/third_party/fmt)
@@ -1528,7 +1502,6 @@ add_subdirectory(${PROJECT_SOURCE_DIR}/third_party/fmt)
 # CMAKE_CXX_FLAGS in ways that break feature checks. Since we already know
 # `fmt` is compatible with a superset of the compilers that PyTorch is, it
 # shouldn't be too bad to just disable the checks.


@@ 248,7 241,7 @@ index f1f2eb7..b4d2033 100644
 list(APPEND Caffe2_DEPENDENCY_LIBS fmt::fmt-header-only)
 set(BUILD_SHARED_LIBS ${TEMP_BUILD_SHARED_LIBS} CACHE BOOL "Build shared libs" FORCE)
diff --git a/cmake/External/nnpack.cmake b/cmake/External/nnpack.cmake
index 9d5f064..c3624e5 100644
index 9d5f0643ec..c3624e582a 100644
--- a/cmake/External/nnpack.cmake
+++ b/cmake/External/nnpack.cmake
@@ -40,7 +40,7 @@ endif()


@@ 270,7 263,7 @@ index 9d5f064..c3624e5 100644
+set(NNPACK_FOUND TRUE)
+set(USE_NNPACK ON)
diff --git a/cmake/public/mkldnn.cmake b/cmake/public/mkldnn.cmake
index 8793562..9f8fa3d 100644
index 87935625f9..9f8fa3df81 100644
--- a/cmake/public/mkldnn.cmake
+++ b/cmake/public/mkldnn.cmake
@@ -4,7 +4,7 @@ if(CPU_AARCH64)


@@ 289,10 282,10 @@ index 8793562..9f8fa3d 100644
-  ${MKLDNN_LIBRARIES})
+  DNNL::dnnl)
diff --git a/setup.py b/setup.py
index 9ec41cd..1f505fe 100644
index 2b0cfa99d7..2d76b0d035 100644
--- a/setup.py
+++ b/setup.py
@@ -494,13 +494,9 @@ def build_deps():
@@ -491,13 +491,9 @@ def build_deps():
     # Windows has very poor support for them.
     sym_files = [
         "tools/shared/_utils_internal.py",


@@ 307,15 300,17 @@ index 9ec41cd..1f505fe 100644
     for sym_file, orig_file in zip(sym_files, orig_files):
         same = False
diff --git a/test/cpp/c10d/CMakeLists.txt b/test/cpp/c10d/CMakeLists.txt
index 5c89748..ef84c57 100644
index 0874852517..58ad3e9d66 100644
--- a/test/cpp/c10d/CMakeLists.txt
+++ b/test/cpp/c10d/CMakeLists.txt
@@ -16,14 +16,14 @@ function(c10d_add_test test_src)
@@ -16,15 +16,15 @@ function(c10d_add_test test_src)
   add_test(NAME ${test_name} COMMAND $<TARGET_FILE:${test_name}>)
 endfunction()
 
-c10d_add_test(BackoffTest.cpp torch_cpu gtest_main)
-c10d_add_test(FileStoreTest.cpp torch_cpu gtest_main)
-c10d_add_test(TCPStoreTest.cpp torch_cpu gtest_main)
+c10d_add_test(BackoffTest.cpp torch_cpu gtest_main gtest)
+c10d_add_test(FileStoreTest.cpp torch_cpu gtest_main gtest)
+c10d_add_test(TCPStoreTest.cpp torch_cpu gtest_main gtest)
 if(INSTALL_TEST)


@@ 328,7 323,7 @@ index 5c89748..ef84c57 100644
   if(INSTALL_TEST)
     install(TARGETS HashStoreTest DESTINATION bin)
   endif()
@@ -31,11 +31,11 @@ endif()
@@ -32,11 +32,11 @@ endif()
 
 if(USE_CUDA)
   if(USE_GLOO AND USE_C10D_GLOO)


@@ 342,7 337,7 @@ index 5c89748..ef84c57 100644
   endif()
   if(USE_NCCL AND USE_C10D_NCCL)
     # NCCL is a private dependency of libtorch, but the tests include some
@@ -44,10 +44,10 @@ if(USE_CUDA)
@@ -45,10 +45,10 @@ if(USE_CUDA)
     # a private dependency of the tests as well.
     c10d_add_test(
       ProcessGroupNCCLTest.cpp


@@ 355,7 350,7 @@ index 5c89748..ef84c57 100644
     if(INSTALL_TEST)
       install(TARGETS ProcessGroupNCCLTest DESTINATION bin)
       install(TARGETS ProcessGroupNCCLErrorsTest DESTINATION bin)
@@ -61,7 +61,7 @@ if(USE_CUDA)
@@ -62,7 +62,7 @@ if(USE_CUDA)
     # a private dependency of the tests as well.
     c10d_add_test(
       ProcessGroupUCCTest.cpp


@@ 364,7 359,7 @@ index 5c89748..ef84c57 100644
     if(INSTALL_TEST)
       install(TARGETS ProcessGroupUCCTest DESTINATION bin)
       install(TARGETS c10d_cuda_test DESTINATION lib)
@@ -69,7 +69,7 @@ if(USE_CUDA)
@@ -70,7 +70,7 @@ if(USE_CUDA)
   endif()
 else()
   if(USE_GLOO AND USE_C10D_GLOO)


@@ 374,7 369,7 @@ index 5c89748..ef84c57 100644
 endif()
 
diff --git a/test/cpp/tensorexpr/CMakeLists.txt b/test/cpp/tensorexpr/CMakeLists.txt
index 179270c..72f5582 100644
index 179270c4a4..72f5582e81 100644
--- a/test/cpp/tensorexpr/CMakeLists.txt
+++ b/test/cpp/tensorexpr/CMakeLists.txt
@@ -51,7 +51,7 @@ target_include_directories(tutorial_tensorexpr PRIVATE ${ATen_CPU_INCLUDE})


@@ 387,14 382,15 @@ index 179270c..72f5582 100644
 if(USE_CUDA)
   target_compile_definitions(test_tensorexpr PRIVATE USE_CUDA)
diff --git a/torch/CMakeLists.txt b/torch/CMakeLists.txt
index 10a44af..33e2df4 100644
index c74b45431c..5b5d0919d0 100644
--- a/torch/CMakeLists.txt
+++ b/torch/CMakeLists.txt
@@ -81,7 +81,6 @@ set(TORCH_PYTHON_LINK_LIBRARIES
@@ -82,8 +82,6 @@ set(TORCH_PYTHON_LINK_LIBRARIES
     Python::Module
     pybind::pybind11
     opentelemetry::api
-    httplib
-    nlohmann
     shm
     fmt::fmt-header-only
     ATEN_CPU_FILES_GEN_LIB)