Glossary: Running Open Models Locally

KV heads

The number of key/value sets per attention layer. Fewer KV heads means a smaller KV cache.