<p>After seeing the new detailed prompting methods flux uses to generate images, I decided to see if it was possible to train other models to accept JoyCaption style prompts. I will attempt to train multiple common non-Flux checkpoints like Pony, SDXL, and SD15 to accept more advanced prompts. Check each version's version details to see what it does and how it should be used.</p>

JoyCaption style prompting for non flux models

<p>Sdxl version trained at 256X256 on KohayaSS at a high learning rate at around 3200 steps. Seems to work best at around 0.5 strength, any higher than that it and it has difficulty with prompt adherence. Seems to increase text accuracy on most SDXL models, but on lighting and turbo models, it doesnt help all that much. </p>

Files with SHA256: 7608fd3f0f4a600f4d757027b6a56096051aa3592d3e79c81dd750f06a0d2053