def get_augmentation(mode, input_size):
if mode == 'RGB':
return torchvision.transforms.Compose(
[GroupMultiScaleCrop(input_size, [1, .875, .75, .66])])
elif mode == 'Flow':
return torchvision.transforms.Compose(
[GroupMultiScaleCrop(input_size, [1, .875, .75])])
elif mode == 'RGBDiff':
return torchvision.transforms.Compose(
[GroupMultiScaleCrop(input_size, [1, .875, .75])])
def __init__(self, cfg: DictConfig):
super().__init__()
self.train_gulp_dir = Path(cfg.data.train_gulp_dir)
self.val_gulp_dir = Path(cfg.data.val_gulp_dir)
self.test_gulp_dir = Path(cfg.data.test_gulp_dir)
self.cfg = cfg
channel_count = (3 if self.cfg.modality == "RGB" else 2 *
self.cfg.data.segment_length)
common_transform = Compose([
Stack(bgr=self.cfg.modality == "RGB"
and self.cfg.data.preprocessing.get("bgr", False)),
ToTorchFormatTensor(div=self.cfg.data.preprocessing.rescale),
GroupNormalize(
mean=list(self.cfg.data.preprocessing.mean),
std=list(self.cfg.data.preprocessing.std),
),
ExtractTimeFromChannel(channel_count),
])
self.train_transform = Compose([
GroupMultiScaleCrop(
self.cfg.data.preprocessing.input_size,
self.cfg.data.train_augmentation.multiscale_crop_scales,
),
GroupRandomHorizontalFlip(is_flow=self.cfg.modality == "Flow"),
common_transform,
])
self.test_transform = Compose([
GroupScale(self.cfg.data.test_augmentation.rescale_size),
GroupCenterCrop(self.cfg.data.preprocessing.input_size),
common_transform,
])
def get_augmentation(self):
scales_i = [1, .875, .75, .66]
scales_m = [1, .875, .75]
scales_r = [1, .875, .75]
print('Augmentation scales_i:', scales_i)
print('Augmentation scales_m:', scales_m)
print('Augmentation scales_r:', scales_r)
transform_i = torchvision.transforms.Compose(
[GroupMultiScaleCrop(self._input_size, scales_i),
GroupRandomHorizontalFlip(is_mv=False)])
transform_m = torchvision.transforms.Compose(
[GroupMultiScaleCrop(self._input_size, scales_m),
GroupRandomHorizontalFlip(is_mv=True)])
transform_r = torchvision.transforms.Compose(
[GroupMultiScaleCrop(self._input_size, scales_r),
GroupRandomHorizontalFlip(is_mv=False)])
return transform_i, transform_m, transform_r
def get_augmentation(self):
if self._representation in ['mv', 'residual']:
scales = [1, .875, .75]
else:
scales = [1, .875, .75, .66]
print('Augmentation scales:', scales)
return torchvision.transforms.Compose([
GroupMultiScaleCrop(self._input_size, scales),
GroupRandomHorizontalFlip(is_mv=(self._representation == 'mv'))
])
def get_augmentation(self):
scales = [1, .875, .75, .66]
return torchvision.transforms.Compose([
GroupMultiScaleCrop(self._input_size, scales),
GroupRandomHorizontalFlip(is_mv=False)
])