}j;%SSKrSSKJr SSKJr SSKJr \R"SSS9r\R\ \ S'\S\ S \S 4S j5r S \ 4S jr S \ S S4SjrS \ 4Sjrg)N) Generator)contextmanager)_core _CUDA_BACKENDdefault)rbackendreturn)NNNc## UR5nUS:XaSnUS;a[SUS35eUS:XaSn[RU5nSv [R U5 g![R U5 f=f7f)aFContext Manager to set the CUDA backend for :class:`~torchcodec.decoders.VideoDecoder`. This context manager allows you to specify which CUDA backend implementation to use when creating :class:`~torchcodec.decoders.VideoDecoder` instances with CUDA devices. The default is ``"nvdec"``. An ``"ffmpeg"`` backend is also available. Only the creation of the decoder needs to be inside the context manager, the decoding methods can be called outside of it. You still need to pass ``device="cuda"`` when creating the :class:`~torchcodec.decoders.VideoDecoder` instance. If a CUDA device isn't specified, this context manager will have no effect. See example below. This is thread-safe and async-safe. Args: backend (str): The CUDA backend to use. Can be ``"nvdec"`` (default) or ``"ffmpeg"``. Example: >>> with set_cuda_backend("ffmpeg"): ... decoder = VideoDecoder("video.mp4", device="cuda") ... ... # Only the decoder creation needs to be part of the context manager. ... # Decoder will use the FFmpeg CUDA implementation: ... decoder.get_frame_at(0) betanvdec)r ffmpegzInvalid CUDA backend (z-). Supported values are 'nvdec' and 'ffmpeg'.rN)lower ValueErrorrsetreset)rprevious_states ڰC:\Users\johnn\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.13_qbz5n2kfra8p0\LocalCache\local-packages\Python313\site-packages\torchcodec/decoders/_decoder_utils.pyset_cuda_backendrs<mmoG&))$WI-Z [  '"&&w/N, N+ N+sA BA(B(A??Bc*[R5$)N)rgetr_get_cuda_backendrFs    rcapacityc0[R"U5 g)alSet the maximum number of NVDEC decoders that can be cached (per GPU). The NVDEC decoder cache stores hardware decoders for reuse, avoiding the overhead of creating and destructing new decoders for subsequent video decoding operations on the same GPU. This function sets the capacity of the cache, i.e. the maximum number of decoders that can be cached per device. The default capacity is 20 decoders per device. If the cache contains more decoders than the target ``capacity``, excess decoders will be evicted using a least-recently-used policy. Generally, a decoder can be re-used from the cache if it matches the same codec and frame dimensions. See also :func:`~torchcodec.decoders.get_nvdec_cache_capacity`. Args: capacity (int): The maximum number of NVDEC decoders that can be cached per GPU device. Must be non-negative. Setting to 0 disables caching. N)rset_nvdec_cache_capacity)rs rrrJs( ""8,rc,[R"5$)zGet the capacity of the per-device NVDEC decoder cache. See also :func:`~torchcodec.decoders.set_nvdec_cache_capacity`. Returns: int: The maximum number of NVDEC decoders that can be cached per GPU device. )rget_nvdec_cache_capacityrrrrras  ) ) ++r) contextvarscollections.abcr contextlibr torchcodecr ContextVarrstr__annotations__rrintrrrrrr's%%.9-C-CY. {%%c* -,c-,i0@&A-,-,`3-s-t-.,#,r