Finally I decided to version my SD 1.5 LoRa with the artistic style of the Spanish illustrator Alfonso Azpiri for Pony XL. He was well known for his erotic Lorna comics and for creating more than 200 Spanish video game covers in the 80s and 90s. He also published his comics in the prestigious magazine 'Heavy Metal', and also made comics for young audiences like those of his character Mot. His graphic style is very characteristic and attractive.

This LoRa is trained over Pony XL V6 checkpoint, and works well with streghts of 0.7-0.9. The sweet spot is between 0.8-0.9. Don't forget to use the Pony prompt style to have good results. The usual prompt should start with 'score_9, score_8_up, score_7_up' and i recomend to use this as Negative Prompt:

score_6_up, score_5_up, score_4_up, score_3_up, EasyNegative, (worst quality, low quality:1.4), text, watermark, signature, artist name, artist logo, censored, ugly, bad hands, bad anatomy, bad proportions, simple background, toon, cartoon, boring background

EasyNegative is a negative embedding for the negative prompt mostly for SD 1.5 and totally optional, but it has positive effects on the quality of the images here too.

You can control the style of the generated images using the labels 'retro artstyle, watercolor /(medium/)' and/or 'traditional media, retro artstyle, 1980s (style)' to mimic the watercolor style used by Azpiri on some of his works, or without the named labels to use a more detailed style. In the example images you can see the difference between both styles.

This LoRa can reproduce some of Azpiri's characters quite well despite not having been specifically trained to do so. For example:

• Thaca Reem:
  score_9, score_8_up, score_7_up, 1girl, woman with long curly black hair, dark skin, jewelry, solo, black hair, jewelry, big breasts, hair over one eye, gold pasties, gold chains, makeup, pelvic curtain, earrings, armlet

• Sally Gothic:
  score_9, score_8_up, score_7_up, 1girl, solo, sexy women with red medium hair, blue eyes, black corset, bustier, polearm, cleavage, makeup, eyeshadow, red lipstick, red nails, nail polish, striped shocks, garter straps, detached striped sleeves, striped thighhighs, underwear, earrings, jewelry, medium breasts, nipple slip, black panties, black thong, navel

• Bethlehem Steele:
  score_9, score_8_up, score_7_up, 1girl, solo, large breasts, green eyes, red lipstick, blonde long hair, bangs, makeup, eyeshadow, pelvic curtain, gold pasties, half closed eyes

• Lorna (with her orange bodysuit outfit):
  score_9, score_8_up, score_7_up, 1girl, long blonde hair, blue eyes, weapon, gun, orange bodysuit, big breasts, cleavage, earrings, jewelry, holding weapon, holding rifle with both hands, red lipstick, makeup, eyeshadow, gloves, belt with holsters

You can even generate Mot (the big brown monster with tiny horns from the same name comics), but it's more a hit or miss:

score_9, score_8_up, score_7_up, fat brown monster with tiny horns and tail, claws, Mot

You can see some of Azpiri's characters in the example pictures. If you use Forge or ComfyUI, it's good to enable Perturbed-Attention Guidance (PAG)... It helps to improve some aspects of the images.