Artificial Intelligence #210

Hey, in this issue: how to finetune LLMs in 2024 with Hugging Face; Meta’s free Code Llama model closes the gap with GPT-4; implement a sparse mixture-of-experts language model from scratch; function calling datasets, training, and inference; improve LLMs with proxy-tuning; and more.

The sponsors of this issue are Marqo and True Positive Inc.

• [Tutorial] How to finetune LLMs in 2024 with Hugging Face
• [Nature] Loneliness and suicide mitigation for students using GPT3-enabled chatbots
• Meta’s free Code Llama AI programming tool closes the gap with GPT-4
• [Sponsored] Marqo - the vector search platform

Marqo is defining end-to-end vector search. Unlike a vector DB, Marqo handles everything from embedding to vector storage and retrieval. Use Marqo to convert unstructured data into vectors by leveraging SOTA ML models; get better search results and more performant speeds. Search the way you think with Marqo.

• makeMoE: Implement a sparse mixture-of-experts language model from scratch
• [Video] Function calling datasets, training, and inference
• [Project] Exploring ColBERT with RAGatouille
• [Sponsored] The best way of learning machine learning in 2024
• [Tutorial] Improve LLMs with proxy-tuning
• [Explained] Sampling for text generation

More than 750,000 subscribers are reading this newsletter. If you are building an AI or a data product or service, you can become a sponsor of one of the future newsletter issues and get your business featured in the newsletter. Feel free to reach out to [email protected] for more details on sponsorships.

Enjoy the newsletter? Please help us make it bigger and better by sharing it with colleagues and friends.