[fx2trt ]add temporary setup.py for fx2trt

README.md

@@ -217,6 +217,14 @@ bazel build //:libtorchtrt --compilation_mode opt
 ``` shell
 bazel build //:libtorchtrt --compilation_mode=dbg
 ```
+### Install only FX module
+
+FX2TRT is python based module in torch_tensorrt. If the users would like to install only the FX module of torch_tensorrt, please run
+``` shell
+pushd py
+python3 setup.py install --fx2trt-only
+popd
+```
 
 ### Native compilation on NVIDIA Jetson AGX
 We performed end to end testing on Jetson platform using Jetpack SDK 4.6.
@@ -316,4 +324,4 @@ Take a look at the [CONTRIBUTING.md](CONTRIBUTING.md)
 
 ## License
 
-The Torch-TensorRT license can be found in the LICENSE file. It is licensed with a BSD Style licence
+The Torch-TensorRT license can be found in the LICENSE file. It is licensed with a BSD Style licence

py/setup.py

@@ -23,11 +23,14 @@
 JETPACK_VERSION = None
 
 __version__ = '1.2.0a0'
-
+FX2TRT_ONLY = False
 
 def get_git_revision_short_hash() -> str:
     return subprocess.check_output(['git', 'rev-parse', '--short', 'HEAD']).decode('ascii').strip()
 
+if "--fx2trt-only" in sys.argv:
+    FX2TRT_ONLY = True
+    sys.argv.remove("--fx2trt-only")
 
 if "--release" not in sys.argv:
     __version__ = __version__ + "+" + get_git_revision_short_hash()
@@ -138,11 +141,14 @@ def finalize_options(self):
         develop.finalize_options(self)
 
     def run(self):
-        global CXX11_ABI
-        build_libtorchtrt_pre_cxx11_abi(develop=True, cxx11_abi=CXX11_ABI)
-        gen_version_file()
-        copy_libtorchtrt()
-        develop.run(self)
+        if FX2TRT_ONLY:
+            develop.run(self)
+        else:
+            global CXX11_ABI
+            build_libtorchtrt_pre_cxx11_abi(develop=True, cxx11_abi=CXX11_ABI)
+            gen_version_file()
+            copy_libtorchtrt()
+            develop.run(self)
 
 
 class InstallCommand(install):
@@ -155,11 +161,14 @@ def finalize_options(self):
         install.finalize_options(self)
 
     def run(self):
-        global CXX11_ABI
-        build_libtorchtrt_pre_cxx11_abi(develop=False, cxx11_abi=CXX11_ABI)
-        gen_version_file()
-        copy_libtorchtrt()
-        install.run(self)
+        if FX2TRT_ONLY:
+            install.run(self)
+        else:
+            global CXX11_ABI
+            build_libtorchtrt_pre_cxx11_abi(develop=False, cxx11_abi=CXX11_ABI)
+            gen_version_file()
+            copy_libtorchtrt()
+            install.run(self)
 
 
 class BdistCommand(bdist_wheel):
@@ -254,6 +263,23 @@ def run(self):
         ] + (["-D_GLIBCXX_USE_CXX11_ABI=1"] if CXX11_ABI else ["-D_GLIBCXX_USE_CXX11_ABI=0"]),
         undef_macros=["NDEBUG"])
 ]
+if FX2TRT_ONLY:
+    ext_modules=None
+    packages=[
+        "torch_tensorrt.fx",
+        "torch_tensorrt.fx.converters",
+        "torch_tensorrt.fx.passes",
+        "torch_tensorrt.fx.tools",
+        "torch_tensorrt.fx.tracer.acc_tracer",
+    ]
+    package_dir={
+        "torch_tensorrt.fx": "torch_tensorrt/fx",
+        "torch_tensorrt.fx.converters": "torch_tensorrt/fx/converters",
+        "torch_tensorrt.fx.passes": "torch_tensorrt/fx/passes",
+        "torch_tensorrt.fx.tools": "torch_tensorrt/fx/tools",
+        "torch_tensorrt.fx.tracer.acc_tracer": "torch_tensorrt/fx/tracer/acc_tracer",
+    }
+
 
 with open("README.md", "r", encoding="utf-8") as fh:
     long_description = fh.read()
@@ -282,7 +308,8 @@ def run(self):
     },
     zip_safe=False,
     license="BSD",
-    packages=find_packages(),
+    packages=packages if FX2TRT_ONLY else find_packages(),
+    package_dir=package_dir if FX2TRT_ONLY else {},
     classifiers=[
         "Development Status :: 5 - Stable", "Environment :: GPU :: NVIDIA CUDA",
         "License :: OSI Approved :: BSD License", "Intended Audience :: Developers",
@@ -311,4 +338,4 @@ def run(self):
     exclude_package_data={
         '': ['*.cpp'],
         'torch_tensorrt': ['csrc/*.cpp'],
-    })
+    }),