Retro Ad Flux is a Flux LoRA trained on mid-century print advertisements, blending elements of iconic photorealistic illustrations and nostalgic photographs. I used 216 images 3 repeats 15 epochs and 0.0001 learning rate. These were all captioned using Joy Caption Batch.
When prompting use natural language describing the ad layout, you can specify things like headlines, subheadings and image styles and for the most part it works as expected. Be creative, it's possible to create many different styles, the model appears to be receptive to mentioning stylized font faces and other attributes.
I used very similar settings as I have with my previous LoRAs so I won't go into detail for the sake of my time and brevity, low learning rates with less repeats and good data work best for simple styles.
I'll post a v2 version of this LoRA soon with a focus on the cropped images particular, excluding the text. This should help capture the style of imagery without the ad layouts.