Instructions to use dl1wjd2/Logistics_AX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- ultralytics
How to use dl1wjd2/Logistics_AX with ultralytics:
from huggingface_hub import hf_hub_download from ultralytics import YOLO # pick the weights file from this repo's "Files and versions" tab weights = hf_hub_download("dl1wjd2/Logistics_AX", "<weights>.pt") model = YOLO(weights) source = 'http://images.cocodataset.org/val2017/000000039769.jpg' model.predict(source=source, save=True) - Notebooks
- Google Colab
- Kaggle
Download hug/lib/python3.12/site-packages/dateutil/_tzdata_impl.py from dl1wjd2/Logistics_AX: direct link, hf CLI and curl.
- Browser
- Download file 4.84 kB
-
https://huggingface.co/dl1wjd2/Logistics_AX/resolve/main/hug/lib/python3.12/site-packages/dateutil/_tzdata_impl.py
- Command line
-
hf download hf://dl1wjd2/Logistics_AX/hug/lib/python3.12/site-packages/dateutil/_tzdata_impl.py
-
curl -L -o _tzdata_impl.py https://huggingface.co/dl1wjd2/Logistics_AX/resolve/main/hug/lib/python3.12/site-packages/dateutil/_tzdata_impl.py
4.84 kB
| """This is an implementation of functions that interact with the tzdata module. | |
| It will only live as long as :mod:`dateutil.zoneinfo` lives; when the | |
| ``dateutil.zoneinfo`` module is removed, this can all move under :mod:`tz`. | |
| """ | |
| import contextlib | |
| import importlib | |
| import io | |
| import os | |
| import pkgutil | |
| import sys | |
| import six | |
| if six.PY2: | |
| _TZDATA_LOAD_EXCEPTIONS = ( | |
| ImportError, | |
| IOError, | |
| UnicodeEncodeError, | |
| ValueError, | |
| ) | |
| else: | |
| _TZDATA_LOAD_EXCEPTIONS = ( | |
| ImportError, | |
| FileNotFoundError, | |
| UnicodeEncodeError, | |
| IsADirectoryError, | |
| ValueError, | |
| ) | |
| if sys.version_info < (3, 8): | |
| _nullcontext = getattr(contextlib, "nullcontext", None) | |
| if _nullcontext is None: | |
| def _nullcontext(v): | |
| yield v | |
| def _get_data(package, resource): | |
| # pkgutil.get_data returns None rather than raising when the package | |
| # cannot be found (and on Python 2 raises AttributeError if the | |
| # package has been blocked by setting sys.modules[name] = None), so | |
| # import the package first to get the ImportError the callers expect | |
| # from the importlib.resources versions of these functions. | |
| importlib.import_module(package) | |
| pkg_data = pkgutil.get_data(package, resource) | |
| if pkg_data is None: | |
| raise ImportError("No package named %r" % package) | |
| return pkg_data | |
| def _open_text(package, resource): | |
| str_package_data = _get_data(package, resource).decode("utf-8") | |
| return _nullcontext(io.StringIO(str_package_data)) | |
| def _open_binary(package, resource): | |
| return io.BytesIO(_get_data(package, resource)) | |
| else: | |
| import importlib.resources | |
| if sys.version_info >= (3, 9): | |
| def _open_text(package, resource): | |
| return ( | |
| importlib.resources.files(package) | |
| .joinpath(resource) | |
| .open("r", encoding="utf-8") | |
| ) | |
| def _open_binary(package, resource): | |
| return ( | |
| importlib.resources.files(package).joinpath(resource).open("rb") | |
| ) | |
| else: | |
| _open_text = importlib.resources.open_text | |
| _open_binary = importlib.resources.open_binary | |
| def _is_package_directory(package, resource): | |
| """Whether ``resource`` names a directory inside ``package``. | |
| Opening a directory raises IsADirectoryError on POSIX but PermissionError | |
| on Windows, and on Python < 3.9 there is no portable way to ask the | |
| resource API, so check the package's search path directly (the tzdata | |
| package is a regular directory-backed package). | |
| """ | |
| try: | |
| pkg = importlib.import_module(package) | |
| except ImportError: | |
| return False | |
| for base in getattr(pkg, "__path__", ()): | |
| if os.path.isdir(os.path.join(base, resource)): | |
| return True | |
| return False | |
| def _load_tzdata(key): | |
| components = key.split("/") | |
| package_name = ".".join(["tzdata.zoneinfo"] + components[:-1]) | |
| resource_name = components[-1] | |
| # Empty or relative path components cannot name a zone, and they produce | |
| # module names that the import machinery handles inconsistently across | |
| # versions, so treat them as not found up front. A key naming a directory | |
| # (a subpackage) is not a zone either, and opening it raises a | |
| # platform-dependent error, see CPython gh-85702. | |
| if any(c in ("", ".", "..") for c in components) or _is_package_directory( | |
| package_name, resource_name | |
| ): | |
| raise TZFileNotFound("Time zone not found: %s", zone_key=key) | |
| try: | |
| return _open_binary(package_name, resource_name) | |
| except _TZDATA_LOAD_EXCEPTIONS: | |
| # ImportError: package does not exist (or tzdata is not installed) | |
| # FileNotFoundError: the key is not a file in the package | |
| # UnicodeEncodeError: the key is not encodable as a filename | |
| # IsADirectoryError: the key names a package rather than a resource | |
| # ValueError: the key is not a valid path (e.g. contains a NUL byte) | |
| six.raise_from( | |
| TZFileNotFound("Time zone not found: %s", zone_key=key), None | |
| ) | |
| def _load_tzdata_keys(): | |
| with _open_text("tzdata", "zones") as f: | |
| return list(filter(None, (l.strip() for l in f))) | |
| class TZFileNotFound(ValueError): | |
| """Indicates that a time zone file isn't present in tzdata.""" | |
| def __init__(self, msg, *args, **kwargs): | |
| # TODO: Use keyword-only argument when this is Python 3-only | |
| zone_key = kwargs.pop("zone_key", None) | |
| if zone_key is None: # pragma: nocover | |
| raise TypeError("Required keyword argument missing: zone_key") | |
| super(TZFileNotFound, self).__init__(msg % zone_key, *args) | |
| self.zone_key = zone_key | |