Is it possible to use a verctorDB for the embedings?

[nobody4t]

when I play, I have to load the local data again and again when to start. I wonder if it is possible to use a vectorDB to store all the info.

And it is possible to seperate vectorBD and model? So that I can keep feeding more and more data to DB without affecting other parts.

[imClumsyPanda]

Now the web UI is implemented with the separated method, cli_demo has not been updated. Will update in recent update.

Wang Dong @.***>于2023年4月25日 周二14:03写道：

when I play, I have to load the local data again and again when to start. I wonder if it is possible to use a vectorDB to store all the info.

And it is possible to seperate vectorBD and model? So that I can keep feeding more and more data to DB without affecting other parts.

— Reply to this email directly, view it on GitHub https://github.com/imClumsyPanda/langchain-ChatGLM/issues/171, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABLH5ES6CUFDLEOFSS57N43XC5SLLANCNFSM6AAAAAAXKQGO7I . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[nobody4t]

@imClumsyPanda I have inspected the webui.py. It looks like no vectorDB used too, just files.

I wonder what do you mean by "seperated method"?

[imClumsyPanda]

There’s a selectbox to choose which knowledge base to use

Wang Dong @.***>于2023年4月26日 周三10:12写道：

@imClumsyPanda https://github.com/imClumsyPanda I have inspected the webui.py. It looks like no vectorDB used too, just files.

I wonder what do you mean by "seperated method"?

— Reply to this email directly, view it on GitHub https://github.com/imClumsyPanda/langchain-ChatGLM/issues/171#issuecomment-1522662757, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABLH5EUNBIRZOSGKLXZLIVDXDCACBANCNFSM6AAAAAAXKQGO7I . You are receiving this because you were mentioned.Message ID: @.***>