Regards for all your love and support ❤
avoiding negatives can improve result on SDXL
Description
FAQ
Comments (28)
named XL but Base Model set to SD 1.5
fixed, thx for hint
I've been waiting for this for a long time :D
Awesome. I've been waiting for this. Thank you !
very nice, my favorite model made it to SDXL!
ooohhhhh, now that's good news. the best is now XL....
OMG! I've been waiting for this for a loooong time. I'm a big fan of your models and now there is a SDXL version :D
I have a question: Your 1.5 models were trained checkpoints and with this one it says its a merged checkpoint. May I ask what checkpoints were used in the merge?
After testing it I think it is indeed a merge. It doesn't resemble the 1.5 models at all. It's lacking a lot of texture and definition which the 1.5 had
I was looking for an honest comment. Yes this is not as good as the 1.5 I will wait until an update. Once this hits the level and surpasses 1.5 this will be the standard.
@thesilvermoth @zGenMedia since the finetuning of sdxl models is far more sensitive and captioning is different from 1.x/2.x models it is a lot of work to redo all the work, and since I’m out of time I don’t will dig into do much more finetuning.
Yes finally!
wo!!!yes
'bout time
Many thanks for all the hard work! Looking forward to trying it (after the download over my snail-pace connection finally finishes :-)
...... and congratulations on the new addition to your family!!!!!!
Ohhh nice, thanks!
And massive congratulations on the baby!!!!👍
Amazing!!!
Can you please upload this on huggingface as well? (in huggingface format, whatever it means).
Reason is I'd like very much to use it in SDNext with LCM, but that requires a specific format, not safetensors.
Please see: https://github.hscsec.cn/vladmandic/automatic/issues/2481#issuecomment-1804932075
i am losing my mind at how great this checkpoint is... absolute game changer for me
Every time I use this the model it outputs an image covered in a staticky and weird random color pattern, almost like its corrupted. Has anyone else encountered this problem or is it only me? How do I fix it additionally? Also during the generation I can see that its creating the image fine but at the last second it becomes all messed up, and switching to refiner doesn't stop it.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.