# stablediffusion anythingv3 aiart anime animeart animegirl aiartists childhoodfriendgirl giveaway airdrop nft 幽香 textualinversion novelai sao10th waifudiffusion novelaidiffusion wifudiffusion airmpit aiイラスト ai絵 ai術師さんと繋がりたい

ボツ絵の供養。

MODEL: v1-5-pruned-emaonly.ckpt
Steps: 100, Sampler: Euler, CFG scale: 11
Seed: 24805162, Face restoration: GFPGAN
Size: 1408x1024, Denoising strength: 0.6, First pass size: 704x512

2 6
2022-11-13

Another Anything V3 generation. I may have set the cfg too high though.. but it made for a cool aesthetic.

0 8
2022-11-13

【#Childhoodfriendgirl(cfg)】
AIによるデジタルアート‼️
複数所持でメインキャラクターの未来のストーリーNFTがエアドロ🎉✨

🎁賞品🎁

NFT × 2名様
(#Airdrop)

参加条件
✅フォロー



✅❤️& RT &3Tag

✅Discord
https://t.co/L55Pcqd0JT

⏰24h

115 122
2022-11-11

🎉Childhood friend girl (cfg) Giveaway

CFG AL × 10名様🎉

複数所持でエアドロ！（未来のストーリーNFT⁉️）

総供給量1111
ミントの価格AL: 0.012 ETH Public:0.017 ETH

✅Follow


✅❤️＆ RT & tag3
Discord ⬇️
https://t.co/ztZZJDsUjP

⏰48h

46 48
2022-11-10

23k steps @ high CFG.
Either the updates to UI/VAE or the embedding itself don't like it when it's close to something like "perfect face" in the prompt, it really messes with the result. See the first two images. The second pair are better ones...
https://t.co/AxNibT64u4

0 1
2022-11-08



AIによるデジタルアート✨
複数所持でメインキャラクターの未来のストーリーNFTがエアドロ⁉️

AL × 10名様

✅参加条件
1️⃣フォロー


2️⃣❤️&RT

✅Mint price
AL: 0.012 ETH
Published:0.017 ETH
✅Mint
11/13 12:00 UTC

⏰24h

40 48
2022-11-07

skipping the clamping-by-%ile and just denormalizing CFG20's latents to CFG7.5's abs().max() is very similar to reducing cond_scale, but not quite.

think it works out something like:
(uncond + (cond - uncond) * 20)/(20/7.5)
versus
uncond + (cond - uncond) * (20/(20/7.5))

0 0
2022-11-07

99.99999999999%ile dynthresh
I think this shows that "clamping out n%ile outliers" is only important when you have excessive outliers. the rest of the battle is "what range of values do you span". hence denormalizing the latents to CFG7.5's abs().max(), gives us a safer range.

0 2
2022-11-07

each step: compute CFG7.5. for each channel: flatten, center on mean, grab abs().max()
compute CFG20. each channel: flatten, center on mean, compute 99.x%ile of abs(), pick larger of %ile or of CFG7.5's max. clamp channel by that. normalize. multiply by CFG7.5's max.
code coming

0 6
2022-11-07

made a new algorithm for dynamic thresholding in
enables us to set CFG scale high (e.g. 20) without clipping; keep dynamic range / subtlety in shadows, highlights
we refer to a known-good CFG (7.5)'s dynamic range, which helps us pick a ceiling.
detail to follow

7 63
2022-11-07

CFG 스케일 값을 좀 더 올려서 만들어낸 결과물.
앞선 이미지들은 대략 6~8 언저리였던것 같은데 이쪽은 13 전후.
후방주의 태그를 달았더니 대부분 18금이 되어서 제대로 건질건 그리 많지 않지만 그래도 마음에 드는것 위주로.


0 10
2022-11-06

前張左手很怪，遮罩這個部分拉高cfg和步數重新繪製

0 0
2022-11-06

都同樣seed cfg=5 感覺可以用了

0 0
2022-11-06

CFG=3＝，面孔不清楚

0 0
2022-11-06

got inspired by a fren and trying the same prompt for myself as well.

prompts in ALT!

CFG scale may vary!

8 32
2022-11-05

[本日の成果(18)]
黒エルフ戦士。片方はどちらかというと盗賊ですが……。
"CFG Scale"と"Denoising strength"はあんまりイジっていなかったが、召喚結果がだいぶ変化することに今更気づく。

0 4
2022-11-05

【#Childhoodfriendgirl(#cfg)】

AIによるデジタルアート‼️
複数所持でメインキャラクターの未来のストーリーNFTがエアドロ🎉✨

NFT購入でプライベートDiscordに入れる✨✨

🎁賞品🎁

👑AL × 10名様

参加条件
✅フォロー



✅❤️&RT&3Tag

⏰24h

112 121
2022-11-04

After all of that is done, you can just use the model like any other model. It's REALLY resistant to oversaturation and general crappiness at high CFG values, so try to go no lower than 12. There's a list of negative prompt stuff that people recommend, and that's also in alt-text

0 1
2022-11-04