Modelpublic

worldboss/lauki-qna-gemma3-270m-it-gguf

sourceHugging Facegemmaupdated 27d agoView on Hugging Face

0likes113downloads

Model Card

Lauki Phones Q&A — Gemma 3 270M GGUF

Q8_0 GGUF export of a LoRA SFT of `unsloth/gemma-3-270m-it` on the Lauki Phones customer-support set (`worldboss/lauki-qna`).

Trained with Unsloth in small-talk-finetuning/finetune-unsloth.

Files

—gemma-3-270m-it.Q8_0.gguf

llama.cpp

bash

llama-cli -m gemma-3-270m-it.Q8_0.gguf \
  --temp 1.0 --top-k 64 --top-p 0.95 --repeat-penalty 1.0 \
  -p "<start_of_turn>user\nDoes Lauki Phones support eSIM?<end_of_turn>\n<start_of_turn>model\n"

Gemma 3 recommended sampling: temperature 1.0, topk 64, topp 0.95 (Unsloth Gemma 3 guide).

Ollama

bash

huggingface-cli download worldboss/lauki-qna-gemma3-270m-it-gguf gemma-3-270m-it.Q8_0.gguf --local-dir .
ollama create lauki-qna-gemma3 -f Modelfile
ollama run lauki-qna-gemma3 "Does Lauki Phones support eSIM?"

If Modelfile is in this repo, it already points at gemma-3-270m-it.Q8_0.gguf.

Hub

https://huggingface.co/worldboss/lauki-qna-gemma3-270m-it-gguf