Workflow Introduction
This workflow is based on the Qwen3.5 9B GGUF large model. It can reverse-engineer input images and instructions to generate prompts compatible with other large models. It runs entirely locally without relying on the internet, supports various instruction requirements, and delivers fast reverse-engineering performance.
Model Requirements
- Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf
- Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf
Download these two models to the \models\LLM folder.
Download link: https://huggingface.co/terrylet/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive/tree/main
Usage Instructions
This workflow is designed to be used within the DFCine Infinite Canvas software. In a text node, select the Qwen3.5 Reverse Prompt Workflow, then connect the image and text instruction you want to reverse-engineer to this node. Enter the required prompt settings and click Execute to run.
