Skip to content

Conversation

@ATrueLight4
Copy link
Contributor

Summary

--no-warmup loads larger models faster, --verbose-prompt, see the actual prompts

Type of Change

  • Feature
  • Bug fix
  • Refactor / Code quality
  • Performance improvement
  • Documentation
  • Tests
  • Security fix
  • Build / Release
  • Other (specify below)

AI Assistance

  • This PR was created or reviewed with AI assistance

Testing

Manual

Related Issues

Relates to #5175
Discussion: LINK (if any)

Screenshots/Demos (for UX changes)

Before:

After:

Submitting a Recipe?

Email:

jamadeo and others added 30 commits November 11, 2025 08:20
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
@michaelneale michaelneale changed the base branch from main to micn/embedded-llama-server November 14, 2025 01:46
@michaelneale michaelneale requested a review from a team as a code owner November 14, 2025 01:46
@ATrueLight4
Copy link
Contributor Author

so I am in a quandary... I think my PR for adding --verbose-prompt and --no-warmup should be closed and deleted...

I noticed that the llama-embedded branch was updated and stdio and stderr are now going to null, so no need for --verbose or --verbose-prompt... And if I add the --no-warmup to a new PR, for not spawning a window for llama-server when running the GUI in Windows the PR I just created is no longer needed...

Next I just tested not spawning a window, which never happened in CLI, and the change now no longer spawns a window for the llama-server in GUI under Windows and with the --no-warmup both CLI and GUI starts the larger gpt-oss in time to respond to first prompt given...

Problem... GUI version of goose still doesn't kill the llama-server process at least under Windows. The CLI did, does, and still does, stop the llama-server when the goose CLI stops

But with no window means user would need to open taskmanager or something to kill the llama-server process

I can create a PR with this change to not spawn a window for the llama-server under Windows, but making sure the GUI version of goose kills the llama-server process should probably occur first...

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.