Learning to wield large language models
summary
Explains how to pick and run LLMs with concrete steps: use Anthropic 75k+ context when you must paste entire docs, use Hugging Face Spaces or Transformers to run Falcon 40B locally or on cloud, and use Google Cloud’s tuning UI to upload a JSON of instruction/context/response to fine-tune a model. Shows an end-to-end pattern: chunk PDFs by paragraph, ask an LLM to generate Q&A pairs, build a JSON training file, fine-tune, and serve the tuned model; notes cost (~$0.0007/token) and hardware tradeoffs (local M2 Pro/GPUs).
AI-written summary · about
chapters (8)
- 0:00Introduction to LLMs
- 1:30How LLMs work
- 6:47LLM landscape overview
- 10:21Google Cloud APIs demo
- 15:38Hugging Face models
- 20:12Fine-tuning vs in-context learning
- 25:10Building FAQ bots
- 29:00Q&A
related talks
43mLarge Language Models for Devs: from zero to your first LLM application
43mCreating Your Own Copy with a Fine-Tuned LLM
46mUnder the Hood of LLMs — all those questions you never dared ask about them
31mDeploying Large Language Models with Ease
29mLocal LLM fine-tuning: a practical example
28mFine-Tuning Small Language Models with Python
22mIntroduction to Large Language Models (LLM) on Kubernetes
20mSelf-hosting AI LLMs - a beginners guide