Type | |
Stats | 763 352 |
Reviews | (156) |
Published | Apr 14, 2023 |
Base Model | |
Hash | AutoV2 8C424E4ADF |
Lora trained on Musubi from Sekirei
Master model anime: Based65, meinaunreal
2,5D: fantasticmixReal_V30 (this one is also more realistic), ReV Animated
negative embeddings: embedding:bad-hands-5, embedding:ng_deepnegative_v1_75t, embedding:bad-artist, embedding:EasyNegative
Recommended settings:
Make sure you are aware on the usage instructions of LORA
VAE: orangemix.vae (for 2,5D pictures), vae-ft-mse-840000-ema-pruned (for realistic models), kl-f8-anime2, blessed2.vae (for anime models)
Sampler: DPM++ SDE Karras or DPM++ 2M Karras
Steps: 20 to 35
Clipskip: 1 or 2
Upscaler : img2img in comfyui using 4x-UltraMix_Smooth
CFG Scale : 5 to 9 (mostly 7 and 8)
LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.7)
Denoise strength for img2img: 0.4 to 0.7
If you want the exact same settings, just drag and drop the picture into ComfyUI because this time I included the metadata of the pictures
Trained with based65 checkpoint
for the pictures i used comfyui
for the most accurate pictures i would recommend using this costum node for comfyui: cutoff-for-comfyui
or this for automatic1111: no-more-color-contamination