juggernaut XL -
Checkpoint
31
8642Details
Support
ControlNet | I2I/T2I
The wild journey continues with Juggernaut, but this time based on SDXL. Just like with Juggernaut's SD 1.5 version, I will take you on the entire journey, so Juggernaut's output will change significantly over the next updates.
Please don't be too harsh with your rating; we are still in an early stage of Juggernaut XL
Time for the first update of Juggernaut XL. As I've already announced, I will process the training into several medium-sized datasets in the future and add them to the base model. For the first update, with the help of Dreamlook.ai, I was able to train the first "Side" dataset (58k Steps). The base model has undergone an additional 50k Steps, and in total, we are now at 328k Steps.
At first glance, this seems quite a lot, but there is still an incredible amount of work to be done. Currently, only 25-30% of what I had planned is trained.
With this update, all areas of the model should be improved (Skin, Eyes, Faces, NSFW, etc.). Especially the NSFW area still requires a lot of further training. Particularly, the male anatomy is giving me a lot of headaches :D
The Sample Images are without any refiner or Upscale. Just the RAW Output like usual
All in all, I'm much happier with Version 2 and hope you enjoy generating images as much as I did when testing the model ;)
Show MoreGallery