How to fine - tune this model?
#5
by
Spijoker
- opened
What methods can be used to achieve this?
I found that its performance is poor after adding custom policies. Or even if I modify the original three policies, the impact is not significant. I want to conduct Lora fine-tuning on it to expand the detection range.
Hi @Spijoker ,
Please go through Fine-Tuning Gemma Models in Hugging Face. And also please refer to this codebase that provides a script specifically designed for fine-tuning the Gemma 3 model using only the Hugging Face framework. These resources be useful for fine-tuning the shieldgemma-2-4b-it model.
Thank you.