Open munjalmukul opened 3 months ago
How Its Working Currently Right now If I am adding same document twice its first searching memories and then updating, adding or deleting memory.
How It Should be There should be a layer at embedding search level so it reject the duplicate document to save llm cost.
There are times when duplicate documents are uploaded by user so I dont want to send that document to llm so cost can be saved.
@Dev-Khant I am working on this enhancement, can you assign me this this issue? Looking forward to share the PR soon.
Sure @NavyaAlapati13, Thanks for working on it.
🚀 The feature
How Its Working Currently Right now If I am adding same document twice its first searching memories and then updating, adding or deleting memory.
How It Should be There should be a layer at embedding search level so it reject the duplicate document to save llm cost.
Motivation, pitch
There are times when duplicate documents are uploaded by user so I dont want to send that document to llm so cost can be saved.