ASR AI GALLERY

[Dynamic Prompt] portrait of {a|an American|an Asian|a Latin} girl wearing MLB major league team jersy, {LA Dodgers|New York Yankees|Boston Red Socks|Chicago White Sock|Atlanta Braves|Baltimore Oriole|Chicago Cubs|Milwaukee Brewers|Pittsburgh Pirates|Philadelphia Phillies|Washington Nationals}, big {boobs|breasts|bust}, tight form-fitting jersey showing cleavage, simple dark background, ambient spotlight, team logo of the jersey I used clip_l and 5xxl_fp8_e4m3fn text encoders, no matter how I created images by using an XL checkpoint. As you can see, the correct logo and name text are shown on the images. When you use a Flux checkpoint, clip-L and a text encoder are requested mandatorily. However, using text encoder for XL models improves making correct text and logo. But not always, it's better than nothing.