FlowerTune LLM Leaderboard
Embrace Federated LLM Fine-Tuning and Secure Your Spot on the Leaderboard!
How to participate
If you are interested in taking part in the FlowerTune LLM Leaderboard, here are the instructions to follow.
Rules
Remember to read the rules before contributing. We aim to evaluate solutions under a realistic FL environment and so establish a fair competition, which is why we have a number of real-world constraints.
Read more
Read our FAQs
Got any questions? Check out the FAQs. If you can't find what you're looking for, head over to Flower Discuss, or Flower Slack. The Flower team will be happy to assist you!
How to participate
Follow the instructions to take part in the FlowerTune LLM Leaderboard:
- 01Choose your challenge
- Join the Flower Slack and say “Hi” in channel #flowertune-llm-leaderboard.
- Check out the challenge boards , then choose a challenge.
- Follow the instructions in the challenge page to start your project.
- 02Design your LLM
- Propose and implement your own approach to federated fine-tuning on top of the provided template code.
- Run fine-tuning experiments using flwr run ( instruction ).
- 03Evaluate your model
- Run the evaluation on your fine-tuned LLM ( instruction ).
- Check out the baseline performance and current ranking status on the challenge pages.
- 04Submit your project
- Submit your project through the provided Google Form or Tencent Survey .
- Secure your spot on the leaderboard!
