Skip to content

Optimize generation loop performance - #54

Merged
faridani merged 1 commit into
mainfrom
codex/investigate-and-fix-inference-speed-issue
Oct 19, 2025
Merged

Optimize generation loop performance#54
faridani merged 1 commit into
mainfrom
codex/investigate-and-fix-inference-speed-issue

Conversation

@faridani

Copy link
Copy Markdown
Owner

Summary

  • avoid re-tokenizing prompts during generation and run the model under torch.inference_mode() for faster inference
  • update the server callback and tests to match the streamlined run_generation signature

Testing

  • pytest

https://chatgpt.com/codex/tasks/task_e_68f49ae1ab308329a094c66e196849d8

@faridani
faridani merged commit f3657cc into main Oct 19, 2025
7 of 9 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant