This LoRA was trained from the the work of the artist Ironlily.
Check them out if you like this style, they make funny images : https://x.com/ironlily_art
FLUX:
Trained using 37 images, captions were made using florence-2-large-ft + fixing manually each of them.Adding "a cartoon of" at the beganing of your prompt can force the style if you get realistic results with a specific prompt. It normally doesn't need a trigger word since it wasn't trained using one.I added the training settings + captioning txt file to the ressource if you want to check them out.Illustrious:
Trained using 126 images, captions made with joytag and then cleaned manually.Description
trained with noobAI vpred v10.
FAQ
Comments (3)
Absolutely in love with this LoRA! Very cute, good quality, combines beautifully with artists tags. Great job!
Thank you very much for the compliment, the training part is mostly thanks to @wwHEY for their advice and letting me see an example of their LoRa captioning.
Your images are always a treat to see, please don't stop making them, was quite bummed out when you vanished for a few weeks, glade to see you are back <3
@Poochilli Our mom wwHEY managed to help everyone. Praise to her, haha! :^)
Likewise! I love your works too - they’re always neat and high quality.
Also hope to see more LoRAs from you.
Thanks for the kind words! <3