Skip to content

Add preserve_border flag to vjepa2_preprocessor (default off) to avoid extra border trimming during eval #73

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Lightet wants to merge 1 commit into
base: main
Choose a base branch
from
+8 −5
Open

Add preserve_border flag to vjepa2_preprocessor (default off) to avoid extra border trimming during eval #73

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions evals/hub/preprocessor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,12 +4,12 @@
# LICENSE file in the root directory of this source tree.


def _make_transforms(crop_size=256):
def _make_transforms(crop_size=256, preserve_border=False):
from ..video_classification_frozen.utils import make_transforms

return make_transforms(crop_size=crop_size, training=False)
return make_transforms(crop_size=crop_size, training=False, preserve_border=preserve_border)


def vjepa2_preprocessor(*, pretrained: bool = True, **kwargs):
def vjepa2_preprocessor(*, pretrained: bool = True, preserve_border: bool = False, **kwargs):
crop_size = kwargs.get("crop_size", 256)
return _make_transforms(crop_size=crop_size)
return _make_transforms(crop_size=crop_size, preserve_border=preserve_border)
5 changes: 4 additions & 1 deletion evals/video_classification_frozen/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ def make_transforms(
crop_size=224,
num_views_per_clip=1,
normalize=((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),
preserve_border=False
):

if not training and num_views_per_clip > 1:
Expand All @@ -44,6 +45,7 @@ def make_transforms(
motion_shift=motion_shift,
crop_size=crop_size,
normalize=normalize,
preserve_border=preserve_border
)
return _frames_augmentation

Expand All @@ -61,11 +63,12 @@ def __init__(
motion_shift=False,
crop_size=224,
normalize=((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),
preserve_border=False
):

self.training = training

short_side_size = int(crop_size * 256 / 224)
short_side_size = int(crop_size * 256 / 224) if not preserve_border else crop_size
self.eval_transform = video_transforms.Compose(
[
video_transforms.Resize(short_side_size, interpolation="bilinear"),
Expand Down