Improve user prompt for text-to-image generation

999 6 months ago

Readme

This model is a customized version of a small 2B language model (gemma-2b-instruct) by giving a new system prompt. It aims to demonstrate a small model can enhance your prompt for text-to-image generation. By using this model, a more detailed and informative prompt is generated, which can lead to better and more accurate image generation results.

You can run the model directly in the command prompt with your initial prompt input.

User Prompt

ollama run gnokit/improve-prompt "digital art depicts a lonely figure standing in front of a mega structure"

Enhanced Prompt

Digital art depicts a lonely figure standing in front of a vast, shimmering metropolis with towering skyscrapers
reaching towards the sky. The figure's silhouette is sharp and defined, their expression etched with loneliness
and sadness. The cityscape below is bustling with activity, but the figure stands apart, dwarfed by the towering
structures. The digital art evokes a sense of vastness, isolation, and the overwhelming presence of urban life.