Comment by aarkay

The inference seems to be taking very long and it gets stuck at times forever (> a few minutes). My guess is that this is happening due to the load? How fast is this when it's not under load? Is it as fast as ChatGPT?