Skip to content
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

Fix Llava-NeXT / Llava-NeXT Video / Llava-OneVision's token unpadding mismatch #35779

Merged
merged 2 commits into from
Jan 24, 2025
Merged

Fix Llava-NeXT / Llava-NeXT Video / Llava-OneVision's token unpadding mismatch #35779

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c1e740d
Fix Llava OneVision's token padding
sheryc Jan 19, 2025
b78c385
Fix Llava next and Llava next video's token unpadding for consistency
sheryc Jan 20, 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
4 changes: 2 additions & 2 deletions src/transformers/models/llava_next/processing_llava_next.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,11 +200,11 @@ def _get_unpadded_features(self, height, width, patches_height, patches_width, s
original_aspect_ratio = width / height
current_aspect_ratio = current_width / current_height
if original_aspect_ratio > current_aspect_ratio:
new_height = (height * current_width) // width
new_height = int(round(height * (current_width / width), 7))
padding = (current_height - new_height) // 2
current_height -= padding * 2
else:
new_width = (width * current_height) // height
new_width = int(round(width * (current_height / height), 7))
padding = (current_width - new_width) // 2
current_width -= padding * 2

Expand Down
4 changes: 2 additions & 2 deletions src/transformers/models/llava_next_video/processing_llava_next_video.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -247,11 +247,11 @@ def _get_unpadded_features(self, height, width, patches_height, patches_width, s
original_aspect_ratio = width / height
current_aspect_ratio = current_width / current_height
if original_aspect_ratio > current_aspect_ratio:
new_height = (height * current_width) // width
new_height = int(round(height * (current_width / width), 7))
padding = (current_height - new_height) // 2
current_height -= padding * 2
else:
new_width = (width * current_height) // height
new_width = int(round(width * (current_height / height), 7))
padding = (current_width - new_width) // 2
current_width -= padding * 2

Expand Down
5 changes: 3 additions & 2 deletions src/transformers/models/llava_onevision/processing_llava_onevision.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -219,6 +219,7 @@ def _get_number_of_features(self, orig_height: int, orig_width: int, height: int
num_image_tokens = unpadded_features + newline_features + base_features
return num_image_tokens

# Adapted from transformers.models.llava_next.processing_llava_next.LlavaNextProcessor._get_unpadded_features
def _get_unpadded_features(self, height, width, patches_height, patches_width, scale_height, scale_width):
"""
Get number of features for a given image with height/width. LLaVA-NeXT is different from LLaVA
Expand All @@ -231,11 +232,11 @@ def _get_unpadded_features(self, height, width, patches_height, patches_width, s
original_aspect_ratio = width / height
current_aspect_ratio = current_width / current_height
if original_aspect_ratio > current_aspect_ratio:
new_height = int(height * (current_width / width))
new_height = int(round(height * (current_width / width), 7))
padding = (current_height - new_height) // 2
current_height -= padding * 2
else:
new_width = int(width * (current_height / height))
new_width = int(round(width * (current_height / height), 7))
padding = (current_width - new_width) // 2
current_width -= padding * 2

Expand Down