Updated: Feb 24, 2026
conceptFor my fellow Spanish speakers, I turned the meme into a concept LORA. I used Flux Klein to create anime variations of the image and build a 50-images dataset. As usual, I trained on WAI v12.0. Use the following trigger words: joder_esto_si_es_cine, portrait, smoke, holding cigarette, hand on own face, profile, looking to the side.
The LORA weight works fine around 0.7-0.9. Most of the sample images here used 0.85. The samples were generated with 1536 x 864 px resolution; use upscalers if you want to go below this (something around 1800 x 1000 has worked well for me too).


