in datasets/video_db.py [0:0]
def __init__(self,
return_video=True,
video_root=None,
video_fns=None,
video_clip_duration=1.,
video_fps=25,
video_transform=None,
return_audio=True,
audio_root=None,
audio_fns=None,
audio_clip_duration=1.,
audio_fps=None,
audio_fps_out=None,
audio_transform=None,
return_labels=False,
labels=None,
return_index=False,
mode='clip',
clips_per_video=1,
max_offsync_augm=0,
):
super(VideoDataset, self).__init__()
self.num_samples = 0
self.return_video = return_video
self.video_root = video_root
if return_video:
self.video_fns = chararray(video_fns)
self.num_samples = self.video_fns.shape[0]
self.video_fps = video_fps
if video_transform is not None:
if not isinstance(video_transform, list):
video_transform = [video_transform]
self.video_transform = video_transform
self.return_audio = return_audio
self.audio_root = audio_root
if return_audio:
self.audio_fns = chararray(audio_fns)
self.num_samples = self.audio_fns.shape[0]
self.audio_fps = audio_fps
self.audio_fps_out = audio_fps_out
self.audio_transform = audio_transform
self.return_labels = return_labels
if return_labels:
self.labels = np.array(labels)
self.return_index = return_index
self.video_clip_duration = video_clip_duration
self.audio_clip_duration = audio_clip_duration
self.max_offsync_augm = max_offsync_augm
self.clips_per_video = clips_per_video
self.mode = mode