Skip to content

fix(vllm): Enhance VLLMModel context size handling for batch inputs#1205

Open
paulovsantanas wants to merge 1 commit intohuggingface:mainfrom
paulovsantanas:main
Open

fix(vllm): Enhance VLLMModel context size handling for batch inputs#1205
paulovsantanas wants to merge 1 commit intohuggingface:mainfrom
paulovsantanas:main

Conversation

@paulovsantanas
Copy link
Copy Markdown

  • Updated context size calculation to use the longest prompt in the batch for truncation decisions, ensuring that shorter prompts do not skip necessary truncation.
  • Improved logging messages to provide clearer context on truncation actions based on the longest sample in the batch.

This change aims to enhance the model's handling of varying input lengths and improve the clarity of warnings related to context size management.

@paulovsantanas
Copy link
Copy Markdown
Author

Issue: #1204

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant