bug free
#2
by
etemiz
- opened
Is this done with the latest bug free llama.cpp ?
where do you get this mythical latest bug free llama?
lol. I mean they fixed some things for llama 4?
I don't know, what specifically do you refer to?
We used llama.cpp master to quant them after the llama 4 PR was merged. I'm not aware of any llama.cpp llama 4 change that would benefit from requantizing the model. Can you please point out what bug you are talking about? I'm following llama.cpp changes quite closely as I mentain ouer own llama.cpp fork so unless I missed something there should be no quality diffrence between this and a model quantized on the latest llama.cpp master.