Skip to content

make load_and_preprocess_images accept grayscale image #133

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
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

make load_and_preprocess_images accept grayscale image #133

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
94cfaa4
make load_and_preprocess available to grayscale image
HyeonJaeGil May 9, 2025
e701e5e
Merge branch 'facebookresearch:main' into main
HyeonJaeGil Jul 28, 2025
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
7 changes: 4 additions & 3 deletions vggt/utils/load_fn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,7 +94,7 @@ def load_and_preprocess_images_square(image_path_list, target_size=1024):
return images, original_coords


def load_and_preprocess_images(image_path_list, mode="crop"):
def load_and_preprocess_images(image_path_list, mode="crop", grayscale=False):
"""
A quick start function to load and preprocess images for model input.
This assumes the images should have the same shape for easier batching, but our model can also work well with different shapes.
Expand All @@ -105,6 +105,7 @@ def load_and_preprocess_images(image_path_list, mode="crop"):
- "crop" (default): Sets width to 518px and center crops height if needed.
- "pad": Preserves all pixels by making the largest dimension 518px
and padding the smaller dimension to reach a square shape.
grayscale (bool, optional): If True, convert images to grayscale. Default is False.

Returns:
torch.Tensor: Batched tensor of preprocessed images with shape (N, 3, H, W)
Expand Down Expand Up @@ -140,14 +141,14 @@ def load_and_preprocess_images(image_path_list, mode="crop"):
img = Image.open(image_path)

# If there's an alpha channel, blend onto white background:
if img.mode == "RGBA":
if not grayscale and img.mode == "RGBA":
# Create white background
background = Image.new("RGBA", img.size, (255, 255, 255, 255))
# Alpha composite onto the white background
img = Image.alpha_composite(background, img)

# Now convert to "RGB" (this step assigns white for transparent areas)
img = img.convert("RGB")
img = img.convert("L" if grayscale else "RGB")

width, height = img.size

Expand Down