license: creativeml-openrail-m
library_name: diffusers
pipeline_tag: text-to-image
tags:
- stable-diffusion
Clam
"Clam" is named from using modes' name.
It has better performance in txt2img when using positive prompts describing what you want without any quality tags, and you can use EasyNegative as negative prompt.
Clam2 has more detailed background but it is less stable than Clam, so I recommend to try to use Clam first.
How to Use
here is the parameter recommended to use in stable diffusion
VAE: animevae.pt [f458b5c6]
Sampler: DPM++ 2M Karras or DPM++ 2S a Karras
Step: 15~30
Hires Upscale: 2,Latent Denoising:0.5~0.8
CFG Scale: 7~15
Examples
From Clam:
From Clam2:
Parameters:
1girl, skirt, purple eyes, plaid, pantyhose, shirt, blush, long hair, twintails, plaid skirt, bow, solo focus, brown hair, collared shirt, white shirt, school uniform, bangs, hair between eyes, cardigan, pleated skirt, 1boy, blue bow, long sleeves, open mouth, low twintails, indoors, brown cardigan, fang, very long hair, plaid bow, grey skirt, bowtie, curtains, cheek pinching, breasts, sleeves past wrists, blonde hair
Negative prompt: EasyNegative
Steps: 28, Sampler: DPM++ 2M Karras, CFG scale: 9, Seed: 974084531, Size: 512x768, Model hash: e15065ad94, Model: Clam2, Denoising strength: 0.7, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: Latent
From Clam:
From Clam2:
Parameters:
1girl, solo, long hair, hair ornament, looking at viewer, blush, hairclip, bangs, brown eyes, long sleeves, brown hair, upper body, parted lips, book, holding, white shirt, holding book, shirt, hair between eyes, black hair, hand up, sleeves past wrists, light particles, smile, sweater, indoors, pillow, puffy long sleeves, lens flare, sidelocks, wings, white sweater, shiny, shiny hair, <lora:standingFullBodyWithBackgroundStyle_v10Offset:1>
Negative prompt: EasyNegative
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 9, Seed: 2578216481, Size: 512x768, Model hash: f49368d588, Model: Clam, Denoising strength: 0.7, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: Latent (nearest)
From Clam:
From Clam2:
Parameters:
1girl, solo, hair ornament, hair flower, flower, blue eyes, holding, looking at viewer, jewelry, earrings, white hair, theresa apocalypse, bangs, long hair, bare shoulders, dress, chinese clothes, white flower, sky, parted lips, hair between eyes, night sky, breasts, sleeveless, upper body, night, hair bow, bow, small breasts, sleeveless dress, china dress, detached sleeves, black dress, wand, star (sky), holding wand, star (symbol), long sleeves, lantern
Negative prompt: EasyNegative
Steps: 26, Sampler: DPM++ 2S a Karras, CFG scale: 7, Seed: 2773744106, Size: 512x768, Model hash: f49368d588, Model: Clam, Denoising strength: 0.7, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: Latent (nearest)
Clam
Use Models
Sorry that I forgot to record using models' weight.
- Color Box Model [a80fee84]
- Ligne Claire Anime Diffusion V1 [ee5b91da]
- Anything-V3.0-pruned [2700c435]
- tmndMix-tmndMixPlus [65b8d0b2]
- momoko-e [a2a802b2]
Clam2
I remerge models from above, and add some another models that I think they may generate more detailed results, but it really destroyed the stability of Clam.
Use Models
| Models | Weights |
|---|---|
| NyanMix_230303Absurd2 [466a84a7] | 0.1152 |
| Anything-V3.0-pruned [2700c435] | 0.08257536 |
| EimisAnimeDiffusion_1-0v [4f828a15] | 0.0589824 |
| LigneClaireAnimeDiffusion [ee5b91da] | 0.2029824 |
| pastelmix-better-vae [260de5e8] | 0.18 |
| momoko-e [a2a802b2] | 0.1 |
| IripasteDiffusionSD1_lines [8aa756f3] | 0.073728 |
| colorBoxModel_colorBOX [a80fee84] | 0.0589824 |
| TmndMix_TmndMixPlus [65b8d0b2] | 0.12754944 |