Is the Gemma4‑12B text encoder just the standard model, or is it a version fine‑tuned by your team?
#43
by makisekurisu-jp - opened
Can I directly use the standard Gemma4‑12B model to replace the one in your repository?
For example, by using the GGUF‑quantized version:
https://huggingface.co/unsloth/gemma-4-12b-it-GGUF
The provided Gemma4‑12B text encoder is our custom fine tuned version and it can't be replaced by a standard gemma.
art-alex changed discussion status to closed