Function decode

slowfast/datasets/decoder.py:327–446

Decode the video and perform temporal sampling. Args: container (container): pyav container. sampling_rate (int): frame sampling rate (interval between two sampled frames). num_frames (int): number of frames to sample. clip_idx (int): if clip_idx

(
    container,
    sampling_rate,
    num_frames,
    clip_idx=-1,
    num_clips=10,
    video_meta=None,
    target_fps=30,
    backend="pyav",
    max_spatial_scale=0,
    use_offset=False,
    sparse=False,
    total_frames=None,
    start_index=0
)

def decode(
    container,
    sampling_rate,
    num_frames,
    clip_idx=-1,
    num_clips=10,
    video_meta=None,
    target_fps=30,
    backend="pyav",
    max_spatial_scale=0,
    use_offset=False,
    sparse=False,
    total_frames=None,
    start_index=0
):
    """
    Decode the video and perform temporal sampling.
    Args:
        container (container): pyav container.
        sampling_rate (int): frame sampling rate (interval between two sampled
            frames).
        num_frames (int): number of frames to sample.
        clip_idx (int): if clip_idx is -1, perform random temporal
            sampling. If clip_idx is larger than -1, uniformly split the
            video to num_clips clips, and select the
            clip_idx-th video clip.
        num_clips (int): overall number of clips to uniformly
            sample from the given video.
        video_meta (dict): a dict contains VideoMetaData. Details can be find
            at `pytorch/vision/torchvision/io/_video_opt.py`.
        target_fps (int): the input video may have different fps, convert it to
            the target video fps before frame sampling.
        backend (str): decoding backend includes `pyav` and `torchvision`. The
            default one is `pyav`.
        max_spatial_scale (int): keep the aspect ratio and resize the frame so
            that shorter edge size is max_spatial_scale. Only used in
            `torchvision` backend.
    Returns:
        frames (tensor): decoded frames from the video.
    """
    # Currently support two decoders: 1) PyAV, and 2) TorchVision.
    assert clip_idx >= -1, "Not valied clip_idx {}".format(clip_idx)
    try:
        if backend == "pyav":
            frames, fps, decode_all_video = pyav_decode(
                container,
                sampling_rate,
                num_frames,
                clip_idx,
                num_clips,
                target_fps,
                use_offset=use_offset,
            )
        elif backend == "torchvision":
            frames, fps, decode_all_video = torchvision_decode(
                container,
                sampling_rate,
                num_frames,