,}h(ddlZddlZddlZddlZddlmZddlmZddlmZddl m Z ddl m Z m Z mZmZmZmZddlmZdd lmZmZdd lmZmZmZmZdd lmZdd lmZd e de de ddfdZ!GddeZ"y)N)partial)Pool)path)Path)AnyCallableDictOptionalTupleUnion)Tensor) find_classes make_dataset)check_integritydownload_and_extract_archive download_urlverify_str_arg) VideoClips) VisionDatasettarpath videopathlinereturnct|||yN)r)rrrs \/var/www/html/test/engine/venv/lib/python3.12/site-packages/torchvision/datasets/kinetics.py_dl_wraprs w :c)@eZdZdZddddZddddZ d'd eeefd e d ed ede e de de e de edfde de de de eeefde de de de de de dedd f(fd Zd(d Zd(d!Zd(d"Zedeeeffd#Zde fd$Zd%e de eee ffd&ZxZS))Kineticsu `Generic Kinetics `_ dataset. Kinetics-400/600/700 are action recognition video datasets. This dataset consider every video as a collection of video clips of fixed size, specified by ``frames_per_clip``, where the step in frames between each clip is given by ``step_between_clips``. To give an example, for 2 videos with 10 and 15 frames respectively, if ``frames_per_clip=5`` and ``step_between_clips=5``, the dataset size will be (2 + 3) = 5, where the first two elements will come from video 1, and the next three elements from video 2. Note that we drop clips which do not have exactly ``frames_per_clip`` elements, so not all frames in a video might be present. Args: root (str or ``pathlib.Path``): Root directory of the Kinetics Dataset. Directory should be structured as follows: .. code:: root/ ├── split │ ├── class1 │ │ ├── vid1.mp4 │ │ ├── vid2.mp4 │ │ ├── vid3.mp4 │ │ ├── ... │ ├── class2 │ │ ├── vidx.mp4 │ │ └── ... Note: split is appended automatically using the split argument. frames_per_clip (int): number of frames in a clip num_classes (int): select between Kinetics-400 (default), Kinetics-600, and Kinetics-700 split (str): split of the dataset to consider; supports ``"train"`` (default) ``"val"`` ``"test"`` frame_rate (float): If omitted, interpolate different frame rate for each clip. step_between_clips (int): number of frames between each clip transform (callable, optional): A function/transform that takes in a TxHxWxC video and returns a transformed version. download (bool): Download the official version of the dataset to root folder. num_workers (int): Use multiple workers for VideoClips creation num_download_workers (int): Use multiprocessing in order to speed up download. output_format (str, optional): The format of the output video tensors (before transforms). Can be either "THWC" or "TCHW" (default). Note that in most other utils and datasets, the default is actually "THWC". Returns: tuple: A 3-tuple with the following entries: - video (Tensor[T, C, H, W] or Tensor[T, H, W, C]): the `T` video frames in torch.uint8 tensor - audio(Tensor[K, L]): the audio frames, where `K` is the number of channels and `L` is the number of points in torch.float tensor - label (int): class of the video clip Raises: RuntimeError: If ``download is True`` and the video archives are already extracted. zChttps://s3.amazonaws.com/kinetics/400/{split}/k400_{split}_path.txtzChttps://s3.amazonaws.com/kinetics/600/{split}/k600_{split}_path.txtzMhttps://s3.amazonaws.com/kinetics/700_2020/{split}/k700_2020_{split}_path.txt400600700z=https://s3.amazonaws.com/kinetics/400/annotations/{split}.csvz=https://s3.amazonaws.com/kinetics/600/annotations/{split}.csvzBhttps://s3.amazonaws.com/kinetics/700_2020/annotations/{split}.csvNrootframes_per_clip num_classessplit frame_ratestep_between_clips transform extensions.downloadnum_download_workers num_workers_precomputed_metadata _video_width _video_height_video_min_dimension_audio_samples_audio_channels_legacy output_formatrcpt|dgd|_||_| |_||_||_|r(t d||_d|_d}| r;tdtj|||_t|dgd |_| r|jt|9|jt|j\|_}t#|j||d |_|j$Dcgc]}|d  }}t'||||| | | ||||| |_||_ycc}w) Nr(r")arg valid_valueszUsing legacy structureunknownTHWCz2Cannot download the videos using legacy_structure.r))trainvaltest) is_valid_filer)r0r2r3r4r5r6r8)rr(r-r/r&r7print split_folderr) ValueErrorrjoindownload_and_process_videossuper__init__rclassesrsamplesr video_clipsr,)selfr&r'r(r)r*r+r,r-r.r/r0r1r2r3r4r5r6r7r8 class_to_idxx video_list __class__s rrHzKinetics.__init__\s10*+=Wlm$$8!   * + $D "DJ"M !UVV $ $ 6D '7IabDJ   , , . #%1$2C2C%D" l#D$5$5|Z_cd $(LL1qad1 1%     !#%'!5)+'  #2s D3c(tj}|jtj}td||z dz |jtj}td||z dz td||z dz y)zEDownloads all the videos to the _root_ folder in the expected format.z%Elapsed time for downloading in mins <z$Elapsed time for processing in mins zElapsed time overall in mins N)time_download_videosrB_make_ds_structure)rLtictoctoc2s rrFz$Kinetics.download_and_process_videosswiik iik 5c R7GH !yy{ 4tczR6GH -s b/@Arctj|jrtd|jdtj|j d}tj|j d}|j |jj|j}tj|tj|}t|s t||t|5}|jjDcgc]#}t j"j%|d%}}ddd|j&d k(rD]}t)|||jyt+t,||j}t/|j&} | j1|ycc}w#1swYxYw) adownload tarballs containing the video to "tars" folder and extract them into the _split_ folder where split is one of the official dataset splits. Raises: RuntimeError: if download folder exists, break to prevent downloading entire dataset again. zThe directory z[ already exists. If you want to re-download or re-extract the images, delete the directory.tarsfilesr)z/,:)safeNr)rexistsrC RuntimeErrorrEr& _TAR_URLSr(formatr)basenamerropenread splitlinesurllibparsequoter/rrrrmap) rLtar_pathfile_list_path split_urlsplit_url_filepathfilerlist_video_urlspartpoolprocs rrTzKinetics._download_videoss ;;t(( ) !2!2 34]^ 99TYY/499g6NN4#3#34;;$**;M !YY~t}}Y7OP12 N 3 $ % jPTPYPYP[PfPfPhiv||11$U1CiOi j  $ $ )' P,T8T=N=NO P8Xt/@/@ADD556H LL /j j js7!G(F;G;GG c tj|jd}ttj||jds=t |j |jj|j|tj||jd}d}t|5}tj|}|D]}|j|dt|dt|d}|d jd d jd d jdd jdd }tjtj|j |dtj|j |} tj"| stj| tj|j || dddy#1swYyxYw)umove videos from split_folder/ ├── clip1.avi ├── clip2.avi to the correct format as described below: split_folder/ ├── class1 │ ├── clip1.avi annotationsz.csvr\z{ytid}_{start:06}_{end:06}.mp4 youtube_id time_starttime_end)ytidstartendlabel _'()T)exist_okN)rrEr&rr)r_ANNOTATION_URLSr(rarccsv DictReaderintreplaceosmakedirsrCisfile) rLannotation_pathrs file_fmtstrcsvfilereaderrowfrzdownloaded_files rrUzKinetics._make_ds_structures))DII}=tyyTZZL:MNO ..t/?/?@GGdjjGY[j kiiDJJ>sBGOOPSUWX``adfhi DIId&7&7?$O"&))D,=,=q"A;;/JJ' $"3"3UA>    s:C:G758G77Hc.|jjSr)rKmetadatarLs rrzKinetics.metadatas(((rc6|jjSr)rK num_clipsrs r__len__zKinetics.__len__s))++ridxc|jj|\}}}}|j|d}|j|j|}|||fS)Nr)rKget_cliprJr,)rLrvideoaudioinfo video_idxrzs r __getitem__zKinetics.__getitem__sY(,(8(8(A(A#(F%udI Y'* >> %NN5)EeU""r)r#r>NrN)avimp4FrrNrrrrrFTCHW)rN)__name__ __module__ __qualname____doc__r`rr strrrr rr boolr rrHrFrTrUpropertyrrr r __classcell__)rPs@rr!r!s7tUT^I ONS!$("#(,&4$%:>$% #)@#CI@#@# @#  @# SM @# @#H%@##s(O@#@#"@#@# (S#X7@#@#@# "!@#"#@#$%@#&'@#()@#* +@#D B0rsd  >>.^^#!;c;c;;;a#}a#r