Should I use the default Gemma 3 12b tokenizer with this model when loading it on Oobabooga? LLamacpp_HF still requires it to expose all the sampling parameters.
Yes, it should work with default.
· Sign up or log in to comment
