Fine-tuning vs RAG on OSS @ AGI House

*Theme* Fine-tuning vs RAG! Let's compare the effectiveness of fine-tuning relative to RAG for generative AI apps built on open-source models. Your goal will be to demonstrate your app is better at a task than the base model was on it's own, using either fine-tuning, RAG, or a combination. Your project will be evaluated on how well it accomplishes the task, and how much better it is than the base model. Tracks: 1. Fine-tuning with the Together API. 2. Using RAG with Together AI and Langchain. 3. Implement both fine-tuning and RAG together, showing results independently and combined. Together AI Prize: $2,500 in Together credits and AirPods for all winning team members. LangChain Prize: Access to LangSmith platform, early access to Hosted LangServe, and LangChain swag. WANDB Prize: Merch + Airpods Max or Oculus. Rules: *All attendees must ship code.* Please do not invite other guests. Attendees must be very experienced hackers. 12:00pm Welcome to AGI House 12:05pm Keynotes & Surprise Speakers 12:05pm - 12:30pm Tri Dao (Chief Scientist @ Together.AI - Incoming Assistant Professor @ Princeton) & Albert Gu (Assistant Professor @ CMU) - Mamba: Linear-Time Sequence Modeling with Selective State Spaces 12:30 - 12:40pm Harrison Chase - Founder & CEO @ LangChain - Three levels of RAG: from Q&A to a reflective research assistant 12: 40 - 12:55pm Alex Volkov (AI Evangelist @ Weights & Biases )- Fine-tuning like the pros with WandB 12:55pm Project proposals from hackers 1:10pm Hacking Begins! 1:30pm Lunch is Served 4:00pm Project Check-In 6:00pm Dinner is Served 8:00 - 10pm Demos * Sponsored by: Together AI, Weights & Biases, LangChain