Skip to content
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.

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

Handle Different Context Lengths Between LLM and Embedding Models #12

Open
nkkko opened this issue Sep 25, 2024 · 2 comments
Open

Handle Different Context Lengths Between LLM and Embedding Models #12

nkkko opened this issue Sep 25, 2024 · 2 comments

Comments

@nkkko
Copy link
Member

nkkko commented Sep 25, 2024

Is your feature request related to a problem? Please describe.
Embedding models typically have smaller context windows than LLMs, which can limit the quality of embeddings generated for large contexts. This discrepancy can result in incomplete or less meaningful embeddings for the context used in devcontainer.json generation.

Describe the solution you'd like

  • Propose novel way of handling this challange.
  • Ensure that the primary context used by the LLM doesn't lose its semantic meaning when split or summarized for embedding purposes.
  • Adjust the logic dynamically based on the maximum token limits of the LLM and embedding models specified in the .env file.

Describe alternatives you've considered

  • Using the same context for both LLM and embeddings, which is not feasible due to the context length differences.
  • Manually tweaking the context content, which is not scalable or efficient.

Additional context
Consider edge cases where critical files like README.md might need special handling to ensure they are adequately represented in both the LLM and embedding contexts. Thorough testing is essential to ensure the adjustments maintain the semantic integrity and utility of the generated devcontainer.json configurations.

@nkkko
Copy link
Member Author

nkkko commented Oct 9, 2024

/bounty $20

Copy link

algora-pbc bot commented Oct 9, 2024

💎 $20 bounty • Daytona

Steps to solve:

  1. Start working: Comment /attempt #12 with your implementation plan
  2. Submit work: Create a pull request including /claim #12 in the PR body to claim the bounty
  3. Receive payment: 100% of the bounty is received 2-5 days post-reward. Make sure you are eligible for payouts

If no one is assigned to the issue, feel free to tackle it, without confirmation from us, after registering your attempt. In the event that multiple PRs are made from different people, we will generally accept those with the cleanest code.

Please respect others by working on PRs that you are allowed to submit attempts to.

e.g. If you reached the limit of active attempts, please wait for the ability to do so before submitting a new PR.

If you can not submit an attempt, you will not receive your payout.

Thank you for contributing to daytonaio/devcontainer-generator!

Add a bountyShare on socials

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant