A guide on how to train embeddings with textual inversion to learn a person's likeness. The guide assumes the use of the Automatic1111 Web UI and basic knowledge of embedding-related terminology. The guide explains what embeddings are, their advantages over other options such as models, hypernetworks, and LoRAs, and how they work by creating keywords that trick the model into creating the desired output. The guide also provides suggested settings for training embeddings and explains how to fix common problems.