A Lora made using google colab this model has 3 modes. Shantae pirate/og and a neutral. Suggested prompt for Pirate: shantae \(pirate\), 1girl, solo, pointy ears, purple hair, smile, navel, dark skin, dark-skinned female, blue eyes, long hair, jewelry, open mouth, earrings, beach, ;d, outdoors, harem pants, looking at viewer, pants, cleavage, o-ring, large breasts, choker, vest, o-ring top, bandana, stomach, midriff, high ponytail, bracer, bare shoulders, collarbone, thigh strap, red pants, shoes, open vest, vest Suggested for OG: shantae \(og\), wide hips, gigantic breasts, high ponytail, blue eyes, desert, bandeau, o-ring, red harem pants, outdoors, dark-skinned female, smile and the neutral is for any other outfit. Any suggestions to improve are appreciated. Images generated in anything5.

Shantae

A google colab version has 10 epochs expanded data set, and all pictures have been retagged. It has 3 different modes for two different outfits and a dressup version.

Shantae-10.safetensors

v4.0

Shantae V1

Recommended settings:Make sure you are aware on the usage instructions of LORAVAE: vae-ft-mse-840000-ema-pruned (for realistic models &amp; 2,5D), kl-f8-anime2, blessed2.vae (for anime models)Sampler: DPM++ SDE Karras or DPM++ 2M KarrasSteps: 20 to 35Clipskip: 2Upscaler : img2img in comfyui using 4x-UltraMix_SmoothCFG Scale : 5 to 9 (mostly 7 and 8)LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.8)Denoise strength for img2img: 0.4 to 0.7

夏日限定海滩激情写真

Recommended for use with VAE (to improve the quality of the generation

Files with SHA256: 52cd6744c0b4990a8e181f9b810afc8e5c4a8077fde02963c4dc06ef8e20bf5c