Uploads from VictoryCat

Currently Playing: Using Vector Memory In SillyTavern

In this video we go over how we can use Vector Storage inside of SillyTavern to better handle memory and lorebooks. This can be used to have much longer chats without losing context, and also to use Lorebooks without needing to put keywords on every entry. Timeline: 0:00 - Intro 0:28 - Using Local (Transformers) Vectorization 1:53 - Settings explained 5:01 - Using other vectorization options 5:58 - Ollama vectorization 6:54 - Downloading vectorization model for ollama 7:59 - Removing Ollama models 8:50 - Multilingual vector embeddings 9:11 - Setting up vectorization for spanish 10:30 - Closing Here is the video where we connect Ollama to SillyTavern: https://youtu.be/2eE_HZ0KYz8 This can be very useful for running different vectoriaztion sources locally, but it is by no means needed. "Local (Transformers)" will work even without ollama being set up. In this video we go over how Lorebooks work in SillyTavern: https://youtu.be/mlf1PrMxh74 Lorebooks pair wonderfully with vectorization! I highly recommend using both techniques in conjunction. #sillytavern #tutorial #vectormemory #vectorrag #ragmemory #llm #llmcontext #llmmemory #lorebooks #ai