| # Flux Prompt Enhancer | |
| A specialized Ollama model trained to convert plain text into dynamic image prompts. This fine-tune version is specifically designed to generate more vibrant and creative outputs compared to boring AI image generators. This model builds upon [chuanli11/Llama-3.2-3B-Instruct-uncensored](https://github.com/chuanli11/Llama-3.2-3B-Instruct-uncensored) for enhanced prompt generation capabilities. | |
| ## Usage | |
| Use this template: | |
| Create a **flux prompt** from the given text: {your_prompt} | |
| Settings: | |
| temperature: 0.5 | |
| top_p: 0.9 | |
| ## Key Features | |
| - Creates detailed, content-rich prompts | |
| - Tends toward bold, imaginative interpretations | |
| - Works best with short, simple inputs | |
| ## Ollama | |
| - [Llama-3.2ft Flux-Prompting v0.2](https://ollama.com/poluramus/llama-3.2ft_flux-prompting_v0.2) | |
| - [Llama-3.2ft Flux-Prompting v0.4](https://ollama.com/poluramus/llama-3.2ft_flux-prompting_v0.4) | |
| - [Llama-3.2ft Flux-Prompting v0.5](https://ollama.com/poluramus/llama-3.2ft_flux-prompting_v0.5) | |
| ## Notes | |
| - When providing a prompt to the model, natural English sentences typically result in outputs that are closer to what you described. In contrast, using SDXL-style prompts like "1girl, fighting, etc." can lead to more varied and less predictable outputs. | |
| - If you want more controlled outputs, add "Don't be wild" to your prompt | |
| - Keep prompts brief for optimal results | |