• Attendance: 15

AI in the news and discussion #

• NVIDIA stock had good earnings bump, and discussion of suppliers behind the chip makers
• Using Outlines to constrain LLM output
• Fine-tuning vs embeddings. Fine-tuning is still showing good results vs popularity of embeddings.
• LoRa fine tuning.
• Textual Inversion
• Word Embeddings
• Auto encoder
• Discussing how image generation with auto encoders works

Projects #

• Fine tuning a llm model on spiritual texts, using 4-bit LoRa, with llama.
• Llama has bigger community and better results than Falcon
• Where to get started with local llm dev? Llama cpp.
• Mojo, programming language, superset of python with performance features
• Commercial projects not running on OpenAI?
• CivitAi
• supernormal - AI meeting notes
• wizard 7b llama fine tuned
  • Soggy tee-shirt problem
• Open AI plugin Code interpreter, great for giving coding power to open-ai

Published 23 Aug 2023