Skip to content

fix(data): prompt-length filtering crashes on VLM dataset with apply_chat_template #2126

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

fix(data): prompt-length filtering crashes on VLM dataset with apply_chat_template #2126

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a4560f6
fix(data): reuse stored multimodal_inputs in length filter
Meihan-chen Jun 23, 2026
0e0ec78
del test
Meihan-chen Jun 23, 2026
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
6 changes: 3 additions & 3 deletions slime/utils/data.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,11 +105,11 @@ def filter_long_prompt(origin_samples: list[Sample], tokenizer, processor, max_l
if len(input_ids) <= max_length:
filtered_samples.append(sample)
if multimodal:
from slime.utils.processing_utils import process_vision_info
from slime.utils.processing_utils import build_processor_kwargs

for sample in multimodal:
multimodal_inputs = process_vision_info(sample.prompt, processor)
processor_output = processor(text=sample.prompt, **multimodal_inputs)
processor_kwargs = build_processor_kwargs(sample.multimodal_inputs)
processor_output = processor(text=sample.prompt, **processor_kwargs)
input_ids = processor_output["input_ids"][0]
if len(input_ids) <= max_length:
filtered_samples.append(sample)
Expand Down
Loading