GGUF-quantized version of esalahterus/juicio, a Qwen2.5-7B model fine-tuned to answer questions about Indonesian law. Converted using Unsloth for fast, local inference with llama.cpp-compatible tools (llama.cpp, Ollama, LM Studio, etc.).

Available Model Files

File	Quantization
`Juicio.Q8_0.gguf`	Q8_0 (near-full precision, largest file, highest fidelity)

Example Usage

llama-cli -hf esalahterus/juicio-gguf --jinja

Ollama

An Ollama Modelfile is included in this repo for easy deployment — download the repo and run:

ollama create juicio -f Modelfile
ollama run juicio

Intended Use & Limitations

Same as the base esalahterus/juicio model: intended for general questions about Indonesian law for educational/informational purposes, not a substitute for professional legal advice. May hallucinate legal citations — always verify against primary legal sources.

Acknowledgements

Base fine-tuned model: esalahterus/juicio
Original base model: Qwen2.5-7B-Instruct by Alibaba Cloud
Training datasets: fathurfrs/qna-hukum-indonesia, ShoAnn/legalqa_klinik_hukumonline
GGUF conversion accelerated with Unsloth

This model was finetuned and converted to GGUF format using Unsloth.