Base:Qwen/Qwen2.5-7B-Instruct Format:GGUF, quantization Q8_0 (~8.1 GB) Adapter/Training: Uses a custom adapter trained on the author’s curated materials for the domains above and merged into the base before quantization.
Quickstart (Ollama)
Use the Modelfile so the chat format, identity, and scope limits are correct.
bash
1ollama create aaai -f https://huggingface.co/miladatefi/aaai-7b-dapt-gguf/raw/main/Modelfile
2ollama run aaai