371a1b952acb0d2cccaf3ece7f594763d3f02e03,datasets/videodataset.py,VideoDataset,__getitem__,#VideoDataset#Any#,94

Before Change

        if self.temporal_transform is not None:
            frame_indices = self.temporal_transform(frame_indices)
        clip = self.loader(path, frame_indices)
        if self.spatial_transform is not None:
            self.spatial_transform.randomize_parameters()
            clip = [self.spatial_transform(img) for img in clip]
        clip = torch.stack(clip, 0).permute(1, 0, 2, 3)

        target = self.data[index]
        if self.target_transform is not None:
            target = self.target_transform(target)

After Change

        if self.temporal_transform is not None:
            frame_indices = self.temporal_transform(frame_indices)

        if isinstance(frame_indices[0], list):
            clips = []
            for one_frame_indices in frame_indices:
                clips.append(self.loading(path, one_frame_indices))

            return clips, [target for _ in range(len(clips))]

        else:
            clip = self.loading(path, frame_indices)

            return clip, target

    def __len__(self):
        return len(self.data)

Instances