Instructions to use 0RisingStar0/FantaStel with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use 0RisingStar0/FantaStel with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("0RisingStar0/FantaStel", 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
- Draw Things
- DiffusionBee

U-Net mixed model specialized for fantasy landscape backgrounds(especially floating islands). Pastel style art for background.
FP16 Pruned version(No EMA).
(Quality change may occur in very small details on buildings' textures)
Recommended prompts :
(masterpiece, best quality, excellent quality), ((1girl, solo, cowboy shot)), (fantasy landscape, fictional landscape)
EasyNegative, cowboy, cowboy hat, cowboy western, cowboy boots, cowboy hat, fat, moss, phone, man, pedestrians, extras, border, outside border, white border, watermark, logo, signature
(EasyNegative is a negative embedding : https://huggingface.co/datasets/gsdf/EasyNegative)
Recommended settings :
Sampler : DPM++ 2M Karras
Sampling steps : 24
Resolution : 768x512
CFG Scale : 9.5
Upscale is a must-do!! Otherwise, you won't get intended results.
Upscaler : Latent (nearest)
Hires steps : 0
Denoise : 0.6
Upscale 2x
Recommended VAEs :
orangemix.vae.pt
Mixed models :
AikimiXPv1.0, colormixed, Counterfeit V2.0, Counterfeit V2.5, Dreamlike Diffusion 1.0, HighRiseMixV2, mouseymix-lignepatsel, powercolorV1, RoboeticInkPunkDreamShaperChromaV5
(Thanks to everyone who made above models!)
Feel free to give feedbacks as you wish, I'll try and work around with it.
- Downloads last month
- 58