Hands-On GenAI LLM Training Overview
- Generative AI
- Enterprise h2oGPTe
- H2O Cascade
This lesson walks through the agenda for a hands-on generative AI and large language model training, built around Python notebooks rather than theory. It covers using retrieval augmented generation with Enterprise h2oGPTe's API, preparing datasets manually and with libraries like LLM Data Studio, and fine-tuning models starting from the Hugging Face open-source ecosystem through to H2O LLM Studio. It previews building a generative AI app with H2O's open-source Wave ecosystem and closes with evaluating large language models in code and through an active leaderboard. It notes the course ends with a certification quiz.
Key takeaways
- What the course covers, from RAG basics through dataset prep, fine-tuning, app building, and evaluation
- How retrieval augmented generation gets implemented hands-on using Enterprise h2oGPTe's Python API
- Why dataset prep is covered three ways, manually, through libraries, and using LLM Data Studio
- How fine-tuning is taught starting from Hugging Face's open-source tools before moving to H2O LLM Studio
- The course closes with a certification quiz after covering model evaluation and a leaderboard
Start owning your AI