So many new model architectures, I can't keep up. Krea 2 is definitely going to have a place in my toolbox though. It might be the best current T2I model.
For best results, follow the recommended suggestions in the About This Version section.
v1.0 seems to be very good at content but it has the problem of producing unprompted .
v1.0 Alt has addressed the unprompted issue to some extent but is weaker at .
v2.0 Training Base is my test bed for new training.
v2.0 RC1 is a potential release candidate that builds on Training Base.
v2.0 RC2 (unpublished) is an improvement on RC1, adding some more training and replacing Krealism in the early part of the recipe with regular Krea 2 Turbo.
v2.0 Final is based on the original v1.0, abandoning the ALT, Training Base and RC lineages completely.
v3.0 is in development now.
Beginning with v2.0, there are at least 5 versions available for download:
BF16
FP8
INT8 convrot
Q4_0 GGUF
FP8 All-in-One Checkpoint (text encoder and vae included)
Description
This version is directly based on the original v1.0 which I liked the best even though it had flaws. This version attempts to make progress on fixing those flaws but it might be one or two more versions before I can fully stamp them out.
The ALT, Training Base and RC lineages are being abandoned. I liked them but ultimately I decided that they were the wrong direction to take the model.
Sampler/Scheduler: euler/beta, dpmpp_2m/simple, er_sde/bong_tangent, sa_solver/simple
CFG: 1.0
Steps: 4-8
The workflow is embedded in all of my images.
FAQ
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.