Welcome to Byte Goose AI

This section serves as a consolidated reference for modern AI models and adaptation techniques. It collects practical and theoretical information on how different classes of models work, how they are trained, and how they can be adapted to specific tasks or domains.

This documentation covers:

• Model families (LLMs, multimodal models, embedding models)
• Training vs. fine-tuning vs. post-training adaptation
• Fine-tuning strategies and when to use them
• Embeddings, retrieval-augmented generation, and hybrid approaches
• Evaluation, limitations, and operational considerations

The intent is to provide a clear mental model for selecting and combining techniques based on constraints such as data availability, cost, latency, and maintainability.