5 items found
Sculley and colleagues showed in 2015 that machine learning code is only a small fraction of a real system. This post turns that diagram into clickable tiles and explains what the surrounding platform actually does.
A seasoned fine-tuner’s take: teaching Mistral-7B sarcasm with LoRA, 4-bit quantization, and ~700 examples-because adding wit to chatbots is fun and useful.
A deep dive into Retrieval-Augmented Generation (RAG) and how it's revolutionizing AI responses by combining context-aware retrieval with sophisticated text generation.
A deep dive into the data preparation pipeline behind Gmail's Smart Compose feature
Models with just 135M to 7B parameters started outperforming their heavyweight counterparts on real-world tasks. Learn why smaller is smarter and how SLMs deliver 10x cost reduction with 5x faster inference.