import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("Yntec/NovelAIRemix", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]NovelAIRemix
NovelAI mixed with SD1.5.
Sample and prompt:
sitting elementary girl, Pretty CUTE, gorgeous hair, Magazine ad, iconic, 1943, Cartoon, sharp focus, 4k. beautiful art on canvas by kyoani and ROSSDRAWS and ross tran. DETAILED CHIBI
Check out:
https://huggingface.co/Yntec/NovelAI
Recipe
SD1.4Full + fp16 - no-ema = SD1.4 (https://huggingface.co/Yntec/NovelAIRemix/resolve/main/sd-v1-4-fp16-no-ema.safetensors)
SD1.5Full + fp16 - no-ema = SD1.5 (https://huggingface.co/Yntec/DreamLikeRemix/resolve/main/v1-5-pruned-fp16-no-ema.safetensors)
Add Difference (SD1.4 + (SD1.4 - SD1.5)*1)=SD1.5Essence (https://huggingface.co/Yntec/NovelAIRemix/resolve/main/SD1.5Essence.safetensors)
Weighted Sum (SD1.5Essence * (1 - 0.7) + NovelAIFull * 0.7) = NovelAISD1.5
Weighted Sum (NovelAISD1.5 * (1 - 0.7) + NovelAISFW * 0.7) = NovelAIRemix
- Downloads last month
- 156