metadata
license: creativeml-openrail-m
tags:
- diffusers
- safetensors
- text-to-image
- stable-diffusion
- stable-diffusion-diffusers
- StableDiffusion
- image-generation
- SDXL
- SDXLv1.0
- image-generation
- sdxl-lightning
- distilled
language:
- en
library_name: diffusers
Natural Vision Lightning
Lightning Version(s) of Natural Vision
Model Info
- Base Model - NaturalVision v1.0 + SDXL Lightning
- VAE - sdxl-vae-fp16-fix
- Type - Full-parameter fine-tune, FP16
- Medium - Photorealistic, Cinematography, Photography, Generalist
- Lightning Generation Parameter Recommendations:
- v1-v2 8step:
- CFG: 4-6
- Steps: 8-9
- v2.5 8step:
- CFG: 2-3
- Steps: 8-9
- Sampler: Euler
- Schedule Type: SGM Uniform
- All 4step versions:
- CFG: 1.5-2
- Steps: 4-5
- Sampler: Euler
- Schedule Type: SGM Uniform
- v1-v2 8step:
- Recommended Generation Dimensions:
- 1344x768 (16:9) β Cinematic Film Stills
- 1536x640 (21:9) β Ultrawide Cinematic Film Stills
- 1152x896 (4:3) β Fullscreen
- 1216x832 (3:2) β Mobile landscape
- 1024x1024 (1:1) β Square
- 1024x704 (11:16)
- 768x1344 (9:16) β Tall (Instagram stories / snapchat)
- 896x1152 (3:4)
- 832x1216 (2:3) β Mobile Portrait
- 704x1024 (16:11)
Latest Version: v2.5
Available Lightning Variants: 8step, 4step
Civitai Link: NSFW WARNING