bug with chat

#681
by vadebaf930 - opened

there some bug in chat that it uses to much compute because of non useful rendering , openwebui has a good implementation and you can also select the text even when the new tokens are coming in openwebui , you can use that approach

could you tell me if the issue still occurs ? @vadebaf930 I changed some stuff in our markdown parsing implementation, hopefully it's better now

I've added even more fixes for this :) hopefully should be solved now. will close, we can reopen if you still have issues

nsarrazin changed discussion status to closed

it was fixed before but it happening again with every new tokens the compute on the site increasing and after some time it becomes unresponsive and also theres some problem with the chat whenever in the import it gives "react" it show "eact" like this "import { useEffect } from 'eact';"
also can you add the thinking non thinking mode to the qwen 3 models you just have to add /think or /no_think in the end of the prompt

Sign up or log in to comment