Support for Gemma-2 models

[nerdimite]

Hi, I wanted to ask if the support for Gemma2 models can be added in this binding? Specifically I was trying to load the gemma2-2b-it model but wasn't able to because the ModelType enum only supports 2b and 7b and passing gemma 2 weights in the path doesnt work.

Also, if it's possible, I think it would be better to make these bindings compatible directly with gemma.cpp's CLI args instead of custom enums mapping to an integer, so that this package in itself remains compatible with any new model types that come in the original gemma.cpp. Like directly passing the cpp args ./gemma --tokenizer tokenizer.spm --model gemma2-2b-it --compressed_weights 2.0-2b-it-sfp.sbs in the Gemma() class object

[namtranase]

Hi @nerdimite, thank you for using the package, we will try to update gemma-cpp-python to the latest version compatible with the gemma.cpp within this week!

[JustinSo1]

Has any progress been made on this? I would love to see it implemented