Tedy50

OK a quick solution to this problem is to block that thing with addblocker using block element function LOL

Yes that woud be pretty useful. it is not fun to lose your work when you accidentally close the browser window. Maybe there should be some rolling buffer of 100...

To me the same issue sometimes happens on windows as well. it is hard to reproduce but sometimes loading new model does not unload old

I think at this time exl2 format is most attractive and may become most popular. GPTQ is currently unavoidable because that's the only format which supports training. but comparing performance...

> Though for now I use exl2 because it's just so fast.) yes that's exactly why it is most attractive as it is running on exlama engine which is in...

I think the main thing to mention here is that training can be done only on transformers. you need to disable exllama kernels and select load in 4 bit if...

This is not only content restriction but the models now is very stupid as it cannot understand jokes and constantly talks about safety and other stuff. Even in the demo...