Modelpublic

JupiterJil/cyberspace-qwen25-7b-gguf

sourceHugging Faceupdated 5mo agoView on Hugging Face

0likes11downloads

Model Card

Cyberspace Presales Co-Pilot — Qwen 2.5 7B (Q4KM)

Fine-tuned Qwen 2.5 7B Instruct for generating Cyberspace Limited technical proposals.

—Base: Qwen/Qwen2.5-7B-Instruct
—Method: LoRA via Unsloth on RunPod A100
—Dataset: 132 structured proposals (real + augmented), ChatML format
—Template-bleed mitigation: train_on_responses_only with category-specific system prompts

bash

ollama create cyberspace-qwen25-7b:latest -f Modelfile
ollama run cyberspace-qwen25-7b:latest