Trained on 80 images of Emma Watson
SDXL Update: New model trained on 120 images of Emma Watson around 3,600 steps. This provides very accurate likeness and flexibility.
SDXL: I am still fine-tuning my training process on SDXL it seems to be a little more fussy than training on 1.5. I lowered the learning rate which I feel captures her likeness better ^^.
Description
FAQ
Comments (6)
only 4kb? Wow. I love the new direction of very small loras
I'm a little confused at the SDXL embedding. Is there something I'm missing to allow for TIs in SDXL? It never comes up as available to me and I just (probably erroneously) thought they weren't available for SDXL.
Hey there Starlight ! Glad to see you're back ^^
I'm glad someone's taking a great shot at SDXL embeddings. I don't get why it's been rare for SDXL. Anyway your techniques seems to be on point. Keep on the good work.
can you share the training details of sdxl textual inversion.
Hey there
I wonder, have you ever solidified a good workflow for SDXL TI training? If yes, it would be amazing if you could do a write-up on it.
Also, how long did this training last for you?
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.