Replies: 1 comment 2 replies
-
Hi @abhuva One of my initial focuses was on making sure that notes weren't unnecessarily embedded, as I didn't want excess costs to prevent people from having the opportunity to utilize the technology. This #47 (comment) explains some of the details. Toggling on A file created and deleted before being closed and opened again should only use a negligible amount of tokens. You can double-check this with those console logs. Additionally, some of this depends on the underlying "embedding strategy," which I would consider still under active development. That "embedding strategy" plays a critical role in the performance of this plugin. This part also affects how the "efficiency layer" would need to be designed. But, as that "embedding strategy" gets refined, adding additional layers focused on efficiency will make more sense. Something I have been considering implementing would be a "snooze" button, preventing new embeddings for a preset amount of time. Is that something you think would make sense for your use case? Thanks, |
Beta Was this translation helpful? Give feedback.
-
I just wanted to ask what exactly happens to the embeddings if i delete (or modify) a file?
My guess is that its beeing updated (as i see constant calls to OpenAI Embeddings when i work on my vault).
For the future i think it would be great to also have a bit more control over this (if one wants) - by giving me the option to only update the embeddings when i want it (by using a command) - instead of doing it continuisly in the background.
It might save a bit on the tokens (altough embeddings seems to be quite cheap), but its also sometimes unneccessary if i just work on my vault, wich i do often (work on a file and delete it later once i am done).
Beta Was this translation helpful? Give feedback.
All reactions