An embedding represents a learned concept through a special term in your prompt.
Typical 1
Textual inversion learns a representation that a compatible text encoder can use. An embedding may contain several vectors, even when you refer to it with one term.You can ask the Horde to insert the term into the positive or negative prompt, or add it yourself. The strength field applies when the Horde inserts it for you.Use an embedding made for your model family. Follow its author's instructions for the trigger term and intended use.A request includes at most twenty embeddings. The studio identifies an embedding by its model entry, rather than a specific published version.