This LoRA was trained (with consent) on 36 images by the legendary master of prompts @CranialOrigami. It does not exactly duplicate his numerous styles, it is more of a creative tribute to his vision that you can play with.
Description
Trained for 2000 steps using the prodigy optimizer with a batch size of 4 and a grad accum of 4, offers a slightly different take on the style, perhaps a bit more faithful to the training material.
FAQ
Comments (2)
Fun LoRA but you have a hash crash with Porsche 911 GTS (2024) on Forge. The PNG info looks ok when viewed, so you think everything's ok. Then when you transfer it to T2I, it puts the wrong LoRA name in the prompt. I have a couple other LoRA that do this too.
Did you use Ostris toolkit? I read this on Reddit:
"Ostris AI toolkit had been putting invalid hashes on Lora, at some point (not sure if it's fixed), and forge was using that value as the Lora hash in the metadata, which made a whole lot of images point to the wrong resource when uploaded.
Now, they're checking the hash and correcting it on the upload if it matches the bad hash, but the file needs to replace the original for that to work, hence advising to redownload it."
I did use ostris for this one... hmm
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.