1 image = few seconds(8 steps/cfg scale 1.5~2.5), more "clear" colors compared to normal sdxl turbo LoRA/sdxl lightning LoRA

For Pony/SD XL anime models and realistic(juggernautXL and similars) models

UPDATE v3: lightning version(sdxl_lightning_8step_lora), same speed of v2 but better quality

UPDATE v2: With a low cfg(1.5~2.5) now the images are not blurred like in v1

shiny_style_xl_turbo is a LoRA based on sdxl turbo, you can use the TURBO with any stable diffusion xl checkpoint, few seconds = 1 image(3 seconds with a nvidia rtx 3060 12 GB with 1024x768 resolution)

Merged: sd xl lightning LoRA + selected images

Recommended anime checkpoint: base_anime_style_xl, true_anime_style_xl, Animagine XL V3.1, AAM XL (Anime Mix), autismmixSDXL_autismmixConfetti, Confetti Comrade Mix, Pony Diffusion and similars

Recommended realistic checkpoint: juggernautXL and similars

Weight: 1 recommended

Sampling method: LCM(install animatediff extension if you don't see it in sampling list)/Euler A

Sampling steps: 8+

v2/v3:CFG Scale: 1.5~2.5

v1:CFG Scale: 3~5(cfg with higher value than "normal" sd xl turbo)

Stable Diffusion XL Lightning LoRa: https://huggingface.co/ByteDance/SDXL-Lightning

Stable Diffusion XL Turbo LoRA: https://civarchive.com/models/215485

Stable Diffusion XL Turbo Checkpoint : https://huggingface.co/stabilityai/sdxl-turbo

If you want to support me, check out my Patreon: https://www.patreon.com/Shiroppo

Follow me on civitai(!!!): https://civarchive.com/user/shiroppo

Follow Me on Twitter/X: https://x.com/ShiroppoTwit

Pixiv: https://www.pixiv.net/users/96490888

---