Error on export to executorch

[eikaramba]

No matter the arch/encoder combination i seem to always get errors on trying to export the model to executorch:

Windows 10
pytorch 3.9.0 (with or without CUDA 13.8)
python 3.12
i tried FPN/unet and efficient/mobilenet among other combinations
Also running inside Visual Studio Developer Powershell as descriped

import torch
from executorch.backends.xnnpack.partition.xnnpack_partitioner import XnnpackPartitioner
from executorch.exir import to_edge_transform_and_lower
import segmentation_models_pytorch as smp


# first train model and do save_pretrained('./model')
model = smp.from_pretrained('./model').eval()
sample_inputs = (torch.randn(1, 3, 256, 256), )

et_program = to_edge_transform_and_lower(
    torch.export.export(model, sample_inputs),
    partitioner=[XnnpackPartitioner()]
).to_executorch()

with open("xnnpack.pte", "wb") as f:
    f.write(et_program.buffer)

Loading weights from local directory
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0] Error while creating guard:
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0] Name: ''
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Source: shape_env
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Create Function: SHAPE_ENV
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Guard Types: ['SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV', 'SHAPE_ENV']
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Code List: ["L['x'].size()[0] == 1", "L['x'].size()[1] == 3", "L['x'].size()[2] == 4", "L['x'].size()[3] == 4", "L['x'].stride()[0] == 48", "L['x'].stride()[1] == 16", "L['x'].stride()[2] == 4", "L['x'].stride()[3] == 1", "L['x'].storage_offset() == 0"]
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Object Weakref: None
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     Guarded Class Weakref: None
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0] Traceback (most recent call last):
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_guards.py", line 366, in create
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     return self.create_fn(builder, self)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\guards.py", line 2671, in SHAPE_ENV
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     clib = CppCodeCache.load(func_str)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\codecache.py", line 2839, in load
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     return cls.load_async(*args, **kwargs)()
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\codecache.py", line 2705, in load_async
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     "vec_isa": pick_vec_isa(),
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]                ^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 497, in pick_vec_isa
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     _valid_vec_isa_list: list[VecISA] = valid_vec_isa_list()
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]                                         ^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 484, in valid_vec_isa_list
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     isa_list.extend(
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 487, in <genexpr>
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     if all(flag in _cpu_supported_x86_isa for flag in str(isa).split()) and isa
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]                                                                             ^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 143, in __bool__
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     return self.__bool__impl(config.cpp.vec_isa_ok)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 153, in __bool__impl
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     return self.check_build(VecISA._avx_code)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 103, in check_build
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     extra=_get_isa_dry_compile_fingerprint(self._arch_flags),
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 29, in _get_isa_dry_compile_fingerprint
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     compiler_info = get_compiler_version_info(get_cpp_compiler())
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]                                               ^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 339, in get_cpp_compiler
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     check_msvc_cl_language_id(compiler)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 313, in check_msvc_cl_language_id
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     if not _is_msvc_cl(compiler):
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]            ^^^^^^^^^^^^^^^^^^^^^
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 431, in _is_msvc_cl
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0]     subprocess.check_output([cpp_compiler, "/help"], stderr=subprocess.STDOUT)
E1027 20:17:58.074000 12988 .venv\Lib\site-packages\torch\_guards.py:368] [0/0] UnicodeDecodeError: 'utf-8' codec can't decode byte 0x81 in position 62: invalid start byte
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0] Created at:
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 773, in trace_frame
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]     tracer = InstructionTranslator(
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\symbolic_convert.py", line 3847, in __init__
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]     output=OutputGraph(
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\output_graph.py", line 508, in __init__
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]     self.init_ambient_guards()
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]   File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\output_graph.py", line 668, in init_ambient_guards
E1027 20:17:58.078000 12988 .venv\Lib\site-packages\torch\_guards.py:370] [0/0]     self.guards.add(ShapeEnvSource().make_guard(GuardBuilder.SHAPE_ENV))
Traceback (most recent call last):
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\fx\passes\infra\pass_manager.py", line 276, in __call__
    res = fn(module)
          ^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\fx\passes\infra\pass_base.py", line 46, in __call__
    res = self.call(graph_module)
          ^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\backends\xnnpack\_passes\convert_to_upsample_bilinear2d.py", line 55, in call
    for pattern, align_corners in bilinear_2d.get_graphs_dict().items():
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\backends\xnnpack\partition\graphs\bilinear_2d.py", line 57, in get_graphs_dict
    return _get_bilinear_2d_graphs()
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\backends\xnnpack\partition\graphs\bilinear_2d.py", line 41, in _get_bilinear_2d_graphs
    edge = exir.capture(
           ^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\capture\_capture.py", line 217, in capture
    ep = export(mod, args, dynamic_shapes=dynamic_shapes, strict=True)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\__init__.py", line 311, in export
    raise e
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\__init__.py", line 277, in export
    return _export(
           ^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 1163, in wrapper
    raise e
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 1129, in wrapper
    ep = fn(*args, **kwargs)
         ^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\exported_program.py", line 124, in wrapper
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 2255, in _export
    ep = _export_for_training(
         ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 1163, in wrapper
    raise e
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 1129, in wrapper
    ep = fn(*args, **kwargs)
         ^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\exported_program.py", line 124, in wrapper
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 2071, in _export_for_training
    export_artifact = export_func(
                      ^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 1415, in _strict_export
    gm_torch_level = _export_to_torch_ir(
                     ^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\export\_trace.py", line 812, in _export_to_torch_ir
    gm_torch_level, _ = torch._dynamo.export(
                        ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\eval_frame.py", line 2002, in inner
    result_traced = opt_f(*args, **kwargs)
                    ^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\eval_frame.py", line 414, in __call__
    return super().__call__(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\nn\modules\module.py", line 1775, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\nn\modules\module.py", line 1786, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\eval_frame.py", line 832, in compile_wrapper
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\nn\modules\module.py", line 1775, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\nn\modules\module.py", line 1786, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 1874, in __call__
    result = self._torchdynamo_orig_backend(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 688, in __call__
    result = _compile(
             ^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 1494, in _compile
    raise InternalTorchDynamoError(
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 1433, in _compile
    guarded_code, tracer_output = compile_inner(code, one_graph, hooks)
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_utils_internal.py", line 92, in wrapper_function
    return function(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 1117, in compile_inner
    return _compile_inner(code, one_graph, hooks)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 1251, in _compile_inner
    check_fn = dynamo_output.build_guards(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\convert_frame.py", line 856, in build_guards
    return CheckFunctionManager(
           ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\guards.py", line 3383, in __init__
    builder, guard_manager = self.build_guards(
                             ^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\guards.py", line 3674, in build_guards
    guard.create(builder)
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_guards.py", line 366, in create
    return self.create_fn(builder, self)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_dynamo\guards.py", line 2671, in SHAPE_ENV
    clib = CppCodeCache.load(func_str)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\codecache.py", line 2839, in load
    return cls.load_async(*args, **kwargs)()
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\codecache.py", line 2705, in load_async
    "vec_isa": pick_vec_isa(),
               ^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 497, in pick_vec_isa
    _valid_vec_isa_list: list[VecISA] = valid_vec_isa_list()
                                        ^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 484, in valid_vec_isa_list
    isa_list.extend(
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 487, in <genexpr>
    if all(flag in _cpu_supported_x86_isa for flag in str(isa).split()) and isa
                                                                            ^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 143, in __bool__
    return self.__bool__impl(config.cpp.vec_isa_ok)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 153, in __bool__impl
    return self.check_build(VecISA._avx_code)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 103, in check_build
    extra=_get_isa_dry_compile_fingerprint(self._arch_flags),
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpu_vec_isa.py", line 29, in _get_isa_dry_compile_fingerprint
    compiler_info = get_compiler_version_info(get_cpp_compiler())
                                              ^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 339, in get_cpp_compiler
    check_msvc_cl_language_id(compiler)
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 313, in check_msvc_cl_language_id
    if not _is_msvc_cl(compiler):
           ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\_inductor\cpp_builder.py", line 431, in _is_msvc_cl
    subprocess.check_output([cpp_compiler, "/help"], stderr=subprocess.STDOUT)
torch._dynamo.exc.InternalTorchDynamoError: UnicodeDecodeError: 'utf-8' codec can't decode byte 0x81 in position 62: invalid start byte

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "D:\cloud\machine learning\pytorch\edge-detection\export-execu.py", line 17, in <module>
    et_program = to_edge_transform_and_lower(
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 114, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 1371, in to_edge_transform_and_lower
    edge_manager = edge_manager.to_backend(method_to_partitioner)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 114, in wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 1672, in to_backend
    new_edge_programs = to_backend(method_to_programs_and_partitioners)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\eike\AppData\Roaming\uv\python\cpython-3.12.8-windows-x86_64-none\Lib\functools.py", line 909, in wrapper
    return dispatch(args[0].__class__)(*args, **kw)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\backend\backend_api.py", line 762, in _
    lower_all_submodules_to_backend(
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\backend\backend_api.py", line 591, in lower_all_submodules_to_backend
    backend_name_to_subclass[backend_id].preprocess_multimethod(
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\backend\backend_details.py", line 129, in preprocess_multimethod
    preprocess_result = cls.preprocess(program, compile_spec_for_program)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\backends\xnnpack\xnnpack_preprocess.py", line 139, in preprocess
    ep = XNNPACKPassManager(ep, passes=passes).transform()
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\backends\xnnpack\_passes\__init__.py", line 94, in transform
    ep = _transform(ep, transform_pass)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 244, in _transform
    return _transform_with_pass_manager(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\executorch\exir\program\_program.py", line 266, in _transform_with_pass_manager
    res = pass_manager(self.graph_module)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\cloud\machine learning\pytorch\edge-detection\.venv\Lib\site-packages\torch\fx\passes\infra\pass_manager.py", line 302, in __call__
    raise Exception(msg) from e  # noqa: TRY002
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^
Exception: An error occurred when running the 'ConvertToUpsampleBilinear2d' pass after the following passes: []

[eikaramba]

Interestingly if i try the same under linux i get this:

config.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 106/106 [00:00<00:00, 977kB/s]
model.safetensors: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 21.4M/21.4M [00:03<00:00, 6.84MB/s]
Loading weights from local directory
Traceback (most recent call last):
  File "/mnt/d/cloud/machine learning/pytorch/executorch/export-execu.py", line 13, in <module>
    model = smp.from_pretrained('./tu-timm-efficientnet-b0_unet', encoding='latin1').eval()
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/segmentation_models_pytorch/base/hub_mixin.py", line 142, in from_pretrained
    return model_class.from_pretrained(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/huggingface_hub/utils/_validators.py", line 89, in _inner_fn
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/huggingface_hub/hub_mixin.py", line 559, in from_pretrained
    instance = cls._from_pretrained(
               ^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/huggingface_hub/hub_mixin.py", line 777, in _from_pretrained
    return cls._load_as_safetensor(model, model_file, map_location, strict)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/huggingface_hub/hub_mixin.py", line 822, in _load_as_safetensor
    safetensors.torch.load_model(model, model_file, strict=strict, device=map_location)  # type: ignore [arg-type]
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/d/cloud/machine learning/pytorch/executorch/.venv/lib/python3.12/site-packages/safetensors/torch.py", line 271, in load_model
    raise RuntimeError(error)
RuntimeError: Error(s) in loading state_dict for Unet:
    Missing key(s) in state_dict: "encoder.bn2.bias", "encoder.bn2.num_batches_tracked", "encoder.bn2.running_mean", "encoder.bn2.running_var", "encoder.bn2.weight", "encoder.conv_head.weight"

[eikaramba]

after trying a lot i was able to make it work on linux with the following code:

import torch
from executorch.backends.xnnpack.partition.xnnpack_partitioner import XnnpackPartitioner
from executorch.exir import to_edge_transform_and_lower


model = torch.load('./mymodel.pth', weights_only=False, map_location='cpu')
sample_inputs = (torch.randn(1, 3, 256, 256), )


et_program = to_edge_transform_and_lower(
    torch.export.export(model, sample_inputs),
    partitioner=[XnnpackPartitioner()]
).to_executorch()

with open("xnnpack.pte", "wb") as f:
    f.write(et_program.buffer)

hopefully this helps someone strugling...