Deploy Dedicated GPU server to run AI models

Deploy Model
Skip to main content
⋅ ⊣ Realities Edge XL ⊱ ⋅ LIGHTNING + Turbo! - 👁‍🗹V7 (BakedVAE)👁‍🗹 thumbnail

⋅ ⊣ Realities Edge XL ⊱ ⋅ LIGHTNING + Turbo! - 👁‍🗹V7 (BakedVAE)👁‍🗹 API Documentation

by ModelsLab
  • Baked FIXED FP16 VAE

  • New architecture and training direction, taking advantage of complex merging operations in Supermerger and training of LoRAs

  • Much better EYES, faces and better skin texture.

  • Better Anatomy and clearer mid to far faces in crowds

  • Excellent at 20 steps but up to 40 is sometimes better!

  • Probably even more perfect hands with easy (short) prompts, the longer (more complex) the prompt, the higher the risk of bad hands

  • DON'T use SD1.5 type NEGATIVES with it, just super-short negatives is if you need them. Always put nothing or just "blurry" for better, more high quality outputs.

realitiesedgexllightningturbo-v7bakedvae
Open Source ModelUnlimited UsageLLMs.txt
API PlaygroundAPI Documentation