Modelpublic

perfectPresentation/rcrc-chat-v3-gemma-1b-rag-hanifa-GGUF

sourceHugging Facegemmaupdated 5mo agoView on Hugging Face

0likes29downloads

Model Card

RCRC Hanifa GGUF (Gemma 3 1B)

Quantized GGUF of `perfectPresentation/rcrc-chat-v3-gemma-1b-rag-hanifa` for llama.cpp and Ollama. Original FP16 weights remain at the source repo.

This is the chatbot for RCRC's Hanifa Standard knowledge base, fine-tuned with RAG-style instruction data across MSA, Najdi, Hijazi, and English.

Files

File	Quant	Size (approx)	Notes
`rcrc-hanifa-gemma3-1b-F16.gguf`	F16	~2.0 GB	Reference precision
`rcrc-hanifa-gemma3-1b-Q8_0.gguf`	Q8_0	~1.0 GB	Near-lossless
`rcrc-hanifa-gemma3-1b-Q5_K_M.gguf`	Q5KM	~720 MB	Balanced
`rcrc-hanifa-gemma3-1b-Q4_K_M.gguf`	Q4KM	~620 MB	Recommended for laptops

Quick start — llama.cpp

bash

hf download perfectPresentation/rcrc-chat-v3-gemma-1b-rag-hanifa-GGUF rcrc-hanifa-gemma3-1b-Q4_K_M.gguf --local-dir .
./llama-cli -m rcrc-hanifa-gemma3-1b-Q4_K_M.gguf -cnv

Quick start — Ollama

bash

hf download perfectPresentation/rcrc-chat-v3-gemma-1b-rag-hanifa-GGUF rcrc-hanifa-gemma3-1b-Q4_K_M.gguf Modelfile --local-dir ./rcrc-hanifa
cd rcrc-hanifa
ollama create rcrc-hanifa -f Modelfile
ollama run rcrc-hanifa

Chat template (Gemma 3)

<start_of_turn>user
{user_message}<end_of_turn>
<start_of_turn>model
{response}<end_of_turn>