Loading…
12
Spent 6 months fine tuning my image prompts by hand, found out about LoRA training and felt like a fool
I was writing these long 200 word prompts for Stable Diffusion trying to get a consistent art style for a comic I'm working on, messing with weights and negative prompts for a couple hours a night. Then a guy in a Discord server asked why I wasn't just training a LoRA on my 30 reference images, and I felt so dumb because it took him 20 minutes to set up what I'd been fighting for half a year. Is fine tuning with LoRAs really that much better than prompt engineering, or does it depend on what you're making?
0 comments
Log in to join the discussion
Log In0 Comments
No comments yet
Be the first to share your thoughts on this discussion.