This a a reworked data set of my previous Alt4One V1 LoRa. But trained for PONY checkpoints. So it should produce better results.
I'm looking to refine the data more for the next versions and hopefully a FLUX version. I'm unable to buy buzz as the site one accept my card or my Google Pay. Shrug
Description
FAQ
Comments (4)
1.7GB LoRa? 300MB is good enough
Yeah, that's fair. I'm still new to training. I need to figure out the best settings to keep the size small, but retain a lot of the details and features. If you have any advice, that would be awesome.
@SkullBaron Go and check out CivitAI settings, after tagging images there are training settings to change. I use CivitAI for training LoRa (to save money on my pc and electricity :P ) and those settings seems really good. I usually change clip skip to 2 and max resolution depending on dataset. Later I will try this LoRa of Yours ;) Ahh and You can do this without any buzz.
@Nostra6666 Yeah, I didn't have much success with the default settings, but the currents settings i use produce decent results, but large file sizes. I'll keep looking into it, thanks. Yeah, also train on here to save my pc the strain and electric. XD Thanks again, I'll look into it.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.