Description
FAQ
Comments (5)
Mothra! ♥ Ohrai's work always has such gravitas. This style lends work such weight and dynamic excitement. Thank you so much for training a LoRA on Ohrai.
Was this only trained on space ships? I think the space ships are fine, but the portrait/scene art is so cool...
Yes this was trained on spaceships i know but portrait have to many face in one images its like 6 characters on one image it's hard to train if you can create dataset cut characters to pieces i can create this lora for free
@Eepol Thank you for offering. I think I'm going to try to train a Qwen style lora, since that can parse very complex captions. I've been trying to collect data to train it on, but the resolution available out there isn't great... I found a couple of high res posters though. I just really love those old Japanese mystery novel covers.
Also, I'm surprised by how well this old SDXL lora works. I think Qwen could pull it off more consistently.
https://civitai.com/models/132998/noriyoshi-ohrai-style-sdxl



















