Skip to content

Conversation

@david-thrower
Copy link
Owner

Merge in recent optimizations.

  1. Try Additional permutations of generation settings.
  2. Coerce jit_compile=True
  3. Clean up the prompt samples.

More experimentation on example text prompts.
Trigger cicd tests...
Try more permutations of hyperparmas with the new dataset.
Try top_p 0.85 temperature 0.65, cut underperforming prompts.
Try a sample with lower penalty for repetition. temperature 0.8, top_p 0.99
Added comment about perplexity score cutoff for production.
Temporarily manually set jit_compile to True to check compatibility.
…opy-of-branch-254-updated-hpo-script-for-cicd-scale-testing
@david-thrower david-thrower merged commit 6c7924a into 254-more-optimizations-to-notgpt-hpo-script Oct 10, 2025
1 check failed
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.

2 participants