Evaluate adapters for style fine-tuning in WebIDE

One of the potential solutions to bring custom style guides to LLMs is to train a LoRA on the code that follows a specific style.

This issue is to train LoRA on ai-gateway, integrate it to AI-Gateway and test in WebIDE. The code to train LoRA is completed and can be find here on ai-gateway.

Definition of Done

  1. Train LoRA on AI-Gateway
  2. Hack together a solution to integrate it to AI-Gateway
  3. Evaluate LoRA manually via WebIDE
Edited by Ekaterina Nikonova