Zj; SSKrSSKrSSKrSSKJrJrJrJr SSKrSSK J r \"S5r Sr SS\RS\S \S \S \R4 S jjr"S S5rS\ S \ 4Sjr"SS5rg)N)AnyOptionalTypeVarUnion)tqdmTcBSSKnU$![a [S5ef=f)NrzVideo decoding capabilities were removed from torchvision and migrated to TorchCodec. Please install TorchCodec following instructions at https://github.com/pytorch/torchcodec#installing-torchcodec) torchcodec ImportError)r s q/root/GenerationalWealth/GenerationalWealth/venv/lib/python3.13/site-packages/torchvision/datasets/video_utils.py_get_torchcodecr s6     J   stensorsizestepdilationreturnc&UR5S:wa[SUR535eURS5nUR5nX$-X4-4nXSUS- -S-- U-S-U4nUSS:aSU4n[R "XU5$)z similar to tensor.unfold, but with the dilation and specialized for 1d tensors Returns all consecutive windows of `size` elements, with `step` between windows. The distance between each element in a window is given by `dilation`. rz*tensor should have 1 dimension instead of r)dim ValueErrorstridenumeltorch as_strided)rrrro_strider new_stridenew_sizes r unfoldrszz|qEfjjl^TUU}}QH LLNE/8#67JdQh/!34=A4HH{Qt9   Fj 99cf\rSrSrSrS\\SS4SjrS\4Sjr S\S\ \\\ \ 44S jr S rg) _VideoTimestampsDataset-z Dataset used to parallelize the reading of the timestamps of a list of videos, given their paths in the filesystem. Used in VideoClips and defined at top level, so it can be pickled when forking. video_pathsrNcXlgNr#)selfr#s r __init__ _VideoTimestampsDataset.__init__6s&rc,[UR5$r%lenr#r's r __len___VideoTimestampsDataset.__len__94##$$ridxc[5nURRURU5nURR nURR n[[U55U4$r%) rdecoders VideoDecoderr#metadata num_frames average_fpslistrange)r'r1r decoderr6fpss r __getitem__#_VideoTimestampsDataset.__getitem__<sb$& %%2243C3CC3HI%%00 **E*%&++rr&)__name__ __module__ __qualname____firstlineno____doc__r8strr(intr.tuplerfloatr<__static_attributes__rr r!r!-sO'DI'$'%%,s,uT#Y-G'H,rr!xcU$)z@ Dummy collate function to be used with _VideoTimestampsDataset rH)rIs r _collate_fnrKDs  Hrc\rSrSrSrS,S\\S\S\S\\ S\\ \\ 4S \S \S \S \S \S\S\S\SS4Sjjr S-Sjr S\ \\ 4SS4Sjr\S\ \\ 44Sj5rS\\SS4Sjr\S.S\R(S\S\S\\ S\\ S\\R(\\\\R(444 Sjj5rS.S\S\S\\ SS4SjjrS\4SjrS\4S jrS\4S!jrS"\S\\\44S#jr\S\S$\ S%\ S\\\R(44S&j5rS"\S\\R(\R(\ \\ 4\44S'jrS\ \\ 44S(jr S)\ \\ 4SS4S*jr!S+r"g)/ VideoClipsKa Given a list of video files, computes all consecutive subvideos of size `clip_length_in_frames`, where the distance between each subvideo in the same video is defined by `frames_between_clips`. If `frame_rate` is specified, it will also resample all the videos to have the same frame rate, and the clips will refer to this frame rate. Creating this instance the first time is time-consuming, as it needs to decode all the videos in `video_paths`. It is recommended that you cache the results after instantiation of the class. Recreating the clips for different clip lengths is fast, and can be done with the `compute_clips` method. Args: video_paths (List[str]): paths to the video files clip_length_in_frames (int): size of a clip in number of frames frames_between_clips (int): step (in frames) between each clip frame_rate (float, optional): if specified, it will resample the video so that it has `frame_rate`, and then the clips will be defined on the resampled video num_workers (int): how many subprocesses to use for data loading. 0 means that the data will be loaded in the main process. (default: 0) output_format (str): The format of the output video tensors. Can be either "THWC" (default) or "TCHW". Nr#clip_length_in_framesframes_between_clips frame_rate_precomputed_metadata num_workers _video_width _video_height_video_min_dimension_video_max_dimension_audio_samples_audio_channels output_formatrc:XlX`lXplXlXlXlXlXlU R5Ul URS;a[SU S35eUcUR5 OURU5 URX#U5 g)N)THWCTCHWz5output_format should be either 'THWC' or 'TCHW', got .)r#rSrTrUrVrWrXrYupperrZr_compute_frame_pts_init_from_metadata compute_clips)r'r#rOrPrQrRrSrTrUrVrWrXrYrZs r r(VideoClips.__init__fs"'&)*$8!$8!,.*002   %5 5TUbTccdef f (  # # %  $ $%: ; 0 Src 0/Ul/UlSSKnURRR [ UR5SUR[S9n[[U5S9nUHnURS5 [[U65upVUVs/sHoqR"XqR S9PM nnURR#U5 URR#U5 M SSS5 gs snf!,(df  g=f)Nr) batch_sizerS collate_fn)totalrdtype) video_pts video_fpstorch.utils.datautilsdata DataLoaderr!r#rSrKrr,updater8zip as_tensorlongextend)r'rdlpbarbatch batch_pts batch_fpsptss r r`VideoClips._compute_frame_ptss&( */++*:*:*E*E #D$4$4 5((" +F+ B D A'+CK'8$ PYYy__S Cy Y%%i0%%i0! Z! s,.D#D=H""229q=AAH""r original_fpsnew_fpscX- nUR5(a[U5n[SSU5$[R"U[R S9U-nUR 5R[R5nU$)Nri) is_integerrDrrarangefloat32rtoint64)r6rrrrs r rVideoClips._resample_video_idxsg% ??  t9DtT* *||Jemm>:>a3q6>: 15?1TT%++& ?  ),I")I MM   . "+ gJ   !* 3; @s B=(CrcSU;aXlg[R"US[RS9n[R"X!SSS9nUS X!S'XlUR UR URUR5 g)Nrrkrirr)r) rrrsrsplitrbr6rrQ)r'rrks r __setstate__VideoClips.__setstate__rss Q M OOAkN%++F KK ->+?QG  "+  4??DIItGr)rrYrXrUrWrVrTrrrQr6rSrZrrrlr#rk) rerNNrrrrrrrr\)rNr%)#r>r?r@rArBr8rCrDrrFdictrr(r`rapropertyr5r staticmethodrrrErrrrbr.rrrrrrrrGrHrr rMrMKs:&($%&*:>$%$% ##T#Y#T ##T" #T UO #T (S#X7 #T#T#T#T"#T"#T#T#T#T #TJ14/DcN/t/$sCx. d3i L 2qu<<-08;BJ5/_ghm_n u||U4; #<== >2@@3@HUO@_c@.  %C%)3) #S #U38_ #  5 5 UZ[`bgbnbn[nUo  1-C1-E%,, d3PS8nVY*Y$Z1-fd38n8Hd38nHHrrM)r)rrrtypingrrrrrrnrr rrrDrr!rKrMrHrr rs 00  CL :5<<:s:#::U\\:(,,. 1  uHuHr