google

/

gemma-3-4b-it

like 1.22k

Follow

Google 46.5k

For anybody looking to finetune it in native PyTorch for either Research or for better control over the params and training without the HF trainer class, you may look at my repo: https://github.com/beingdutta/Gemma-3-Multimodal-Finetuning-Using-Native-PyTorch/tree/master