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

PaliGemma: Make sure to add <eos> to suffix if <image> is present in text #35201

Merged
merged 1 commit into from
Dec 19, 2024
Merged

PaliGemma: Make sure to add <eos> to suffix if <image> is present in text #35201

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
10 changes: 5 additions & 5 deletions src/transformers/models/paligemma/processing_paligemma.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -287,11 +287,6 @@ def __call__(
elif not (isinstance(images, list) and isinstance(images[0], list) and is_valid_image(images[0][0])):
raise ValueError("images must be an image, list of images or list of list of images")

if suffix is not None and _is_str_or_image(suffix):
suffix = [suffix]
if suffix is not None:
suffix = [sfx + self.tokenizer.eos_token for sfx in suffix]

input_strings = [
build_string_from_input(
prompt=prompt,
Expand All @@ -314,6 +309,11 @@ def __call__(
)
expanded_samples.append(expanded_sample)
input_strings = [f"{sample}\n" for sample in expanded_samples]

if suffix is not None and _is_str_or_image(suffix):
suffix = [suffix]
if suffix is not None:
suffix = [sfx + self.tokenizer.eos_token for sfx in suffix]
pixel_values = self.image_processor(images, **output_kwargs["images_kwargs"])["pixel_values"]

# max_length has to account for the image tokens
Expand Down