I got to the point where my browser (Brave/Linux Mint/Cinnamon) stopped loading the thread. I used 400tys of the 1 million tokens and that’s the end of it. Is there anything I can do to continue using the model, but normally (without problems from the browser or others, because I don’t know what was supposed to block GAIS from loading)? From ~200tys of tokens, it started to slow down until that point when the thread loads, but a browser error appears (or as now, the tab is not responding…
2 Likes
It did exactly the same thing for me, but at 218K, that specific browser tab stopped completely, or almost completely.
I’m trying to switch from GAIS to Vertex AI - I loaded the entire GAIS history (I have it saved on a google drive) into the prompt and will see if the problem recurs.
[edit] At the moment, prompt VERTEX accepts 537 thousand tokens and running smoothly.