Description:
I’m attempting to package a Python application into a standalone executable using PyInstaller. The application relies on several libraries, including PyTorch. Despite including typing_extensions in the hidden imports, I still encounter a ModuleNotFoundError for typing_extensions when running the executable on the client system.

Environment:
OS: Windows 11
Python Version: 3.12.3
PyInstaller Version: 6.6.0
Virtual Environment: Yes (using .venv)

Directory Structure:
project-root/
├── GUI4.py
├── hook-torch.py
├── gui4.spec
└── .venv/
└── Lib/
└── site-packages/
└── (all installed packages including torch and typing_extensions)

Issue:
When I run the generated executable on a client system, I get the following error:

Error: Traceback (most recent call last):

type here

File “C:PathtoExecutableNewExecutableNameNewExecutableName_internaldetect.py”, line 38, in
import torch
File “C:PathtoExecutableNewExecutableNameNewExecutableName_internaltorch_init_.py”, line 1403, in
from .serialization import save, load
File “C:PathtoExecutableNewExecutableNameNewExecutableName_internaltorchserialization.py”, line 18, in
from typing_extensions import TypeAlias # Python 3.10+
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ModuleNotFoundError: No module named ‘typing_extensions’

Spec File:
Here is my gui4.spec file:

# -*- mode: python ; coding: utf-8 -*-

block_cipher = None

a = Analysis(
    ['GUI4.py'],
    pathex=['.'],
    binaries=[],
    datas=[
        ('weights/best.pt', 'weights'),
        ('detect.py', '.'),
        ('models', 'models'),
        ('utils', 'utils'),
        ('data', 'data'),
        ('export.py', '.'),
        ('requirements.txt', '.'),
    ],
    hiddenimports=[
        'models.common',
        'utils.general',
        'utils.plots',
        'export',
        'typing_extensions',
        'torch',
        'torch.utils.tensorboard',
    ],
    hookspath=['.', 'Path\to\your\project\.venv\Lib\site-packages'],
    runtime_hooks=[],
    excludes=[
        'torch.distributed._shard.checkpoint._dedup_tensors',
        'torch.distributed._shard.checkpoint._fsspec_filesystem',
        'torch.distributed._shard.checkpoint._nested_dict',
        'torch.distributed._shard.checkpoint._sharded_tensor_utils',
        'torch.distributed._shard.checkpoint._state_dict_utils',
        'torch.distributed._shard.checkpoint._traverse',
        'torch.distributed._shard.checkpoint.api',
        'torch.distributed._shard.checkpoint.default_planner',
        'torch.distributed._shard.checkpoint.filesystem',
        'torch.distributed._shard.checkpoint.metadata',
        'torch.distributed._shard.checkpoint.optimizer',
        'torch.distributed._shard.checkpoint.planner',
        'torch.distributed._shard.checkpoint.planner_helpers',
        'torch.distributed._shard.checkpoint.resharding',
        'torch.distributed._shard.checkpoint.state_dict',
        'torch.distributed._shard.checkpoint.state_dict_loader',
        'torch.distributed._shard.checkpoint.state_dict_saver',
        'torch.distributed._shard.checkpoint.stateful',
        'torch.distributed._shard.checkpoint.storage',
        'torch.distributed._shard.checkpoint.utils',
        'torch.distributed._sharded_tensor._ops',
        'torch.distributed._sharded_tensor._ops._common',
        'torch.distributed._sharded_tensor._ops.binary_cmp',
        'torch.distributed._sharded_tensor._ops.init',
        'torch.distributed._sharded_tensor._ops.misc_ops',
        'torch.distributed._sharded_tensor._ops.tensor_ops',
        'torch.distributed._sharded_tensor.api',
        'torch.distributed._sharded_tensor.logger',
        'torch.distributed._sharded_tensor.logging_handlers',
        'torch.distributed._sharded_tensor.metadata',
        'torch.distributed._sharded_tensor.reshard',
        'torch.distributed._sharded_tensor.shard',
        'torch.distributed._sharded_tensor.utils',
        'torch.distributed._sharding_spec._internals',
        'torch.distributed._sharding_spec.api',
        'torch.distributed._sharding_spec.chunk_sharding_spec',
        'torch.distributed._sharding_spec.chunk_sharding_spec_ops',
        'torch.distributed._sharding_spec.chunk_sharding_spec_ops._common',
        'torch.distributed._sharding_spec.chunk_sharding_spec_ops.embedding',
        'torch.distributed._sharding_spec.chunk_sharding_spec_ops.embedding_bag',
    ],
    noarchive=False,
    optimize=0,
    cipher=block_cipher,
)

pyz = PYZ(a.pure, a.zipped_data, cipher=block_cipher)

exe = EXE(
    pyz,
    a.scripts,
    [],
    exclude_binaries=True,
    name='GUI4',
    debug=False,
    bootloader_ignore_signals=False,
    strip=False,
    upx=True,
    upx_exclude=[],
    runtime_tmpdir=None,
    console=True,
    disable_windowed_traceback=False,
    argv_emulation=False,
    target_arch=None,
    codesign_identity=None,
    entitlements_file=None,
)

coll = COLLECT(
    exe,
    a.binaries,
    a.zipfiles,
    a.datas,
    strip=False,
    upx=True,
    upx_exclude=[],
    name='GUI4',
)

Custom Hook:
I also created a custom hook for PyTorch to ensure all dependencies are included:

hook-torch.py:

from PyInstaller.utils.hooks import collect_all

datas, binaries, hiddenimports = collect_all('torch')


typing_extensions_datas, typing_extensions_binaries, typing_extensions_hiddenimports = collect_all('typing_extensions')

datas += typing_extensions_datas
binaries += typing_extensions_binaries
hiddenimports += typing_extensions_hiddenimports

binaries = binaries
datas = datas
hiddenimports = hiddenimports

Steps Taken:
Installed all dependencies in a virtual environment.
Created a spec file and included necessary data files and hidden imports.
Added the site-packages directory to the hookspath in the spec file.
Created a custom hook to ensure PyTorch and typing_extensions are included.
Ran PyInstaller with the spec file to generate the executable.
Issue:
Despite these efforts, the executable still fails with a ModuleNotFoundError for typing_extensions on the client system.

Request:
I need help identifying why typing_extensions is not being included in the standalone executable and how to resolve this issue.

Any insights or suggestions would be greatly appreciated!