Instructions to use lodestones/Kromoss with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use lodestones/Kromoss with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("lodestones/Kromoss", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
DIfference from OPD-priming and how well it can do nsfw without lora compared to the base model? asking for a friend.
Base MOSS can do many, not good but still.. nsfw outputs..
Are NSFW data included on Kromoss as well? and also whats the difference between the OPD-priming and the model without that name? There is no info on the description about that.
Base MOSS can do many, not good but still.. nsfw outputs..
Are NSFW data included on Kromoss as well? and also whats the difference between the OPD-priming and the model without that name? There is no info on the description about that.
Yes it can create NSFW images.
As for your second question, from what I could gather from the commit history:
- Standard Kromoss is the stable fine-tuned checkpoint (update 11,000).
- OPD-priming is a newer experimental checkpoint (update 69,800, five epochs) trained further using GLM-generated captions and bounding-box layout information.
The additional training appears to focus on improving scene and layout understanding, although that doesn't necessarily mean better overall image quality.
The developer also mentioned in discussion #11 that OPD-priming is still experimental and should be used with the DMD2 LoRA. They recommend the standard checkpoint for now. There is a Basic workflow for Comfyui (the one I am using right now).
Hope that helps!