A finetune of gemma4 to write in a more humanlike style. I used GAN style RL instead of primarily SFT to see if it performs better.
Doesn't seem to be better for either writing and loses more instruction following that sft w/ lower lr :(
- Downloads last month
- 99
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support