Skip to content

[Bugfix] use right truncation for non-generative tasks #12050

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

Merged
DarkLight1337 merged 2 commits into from
Jan 15, 2025
Merged

[Bugfix] use right truncation for non-generative tasks #12050

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
337758d
:bug: use right truncation for non-generative tasks
joerunde Jan 14, 2025
4ce2673
:test_tube: add simple truncation test
joerunde Jan 14, 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: 4 additions & 0 deletions vllm/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -357,6 +357,10 @@ def __init__(self,
supported_tasks, task = self._resolve_task(task, self.hf_config)
self.supported_tasks = supported_tasks
self.task: Final = task
if self.task in ("draft", "generate"):
self.truncation_side = "left"
else:
self.truncation_side = "right"

self.pooler_config = self._init_pooler_config(override_pooler_config)
self.logits_processor_pattern = logits_processor_pattern
Expand Down
3 changes: 2 additions & 1 deletion vllm/transformers_utils/tokenizer_group/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,8 @@ def init_tokenizer_from_configs(model_config: ModelConfig,
max_input_length=None,
tokenizer_mode=model_config.tokenizer_mode,
trust_remote_code=model_config.trust_remote_code,
revision=model_config.tokenizer_revision)
revision=model_config.tokenizer_revision,
truncation_side=model_config.truncation_side)

return get_tokenizer_group(parallel_config.tokenizer_pool_config,
**init_kwargs)
Expand Down
Loading