以textual inversion为首,这类以小图集训练新concept的方法是我最担心的,打开了潘多拉的魔盒。画风偷麻了属于是。

0 0

Today I'll react on Twitch to the new Iono's Pokémon name and appearence! Then all my opinions in the TEXTUAL REACTIONS channel on my Discord server!
What will it be?

1 13

Love everlasting from Tom King and Elsa Charretier is a twist on the classic romance genre of comics that puts our heroine first as she has a meta-textual journey through the romance tropes of the comics world. This might be King’s best ongoing work.

5 33

(Stable Diffusion)
雑に描いた1個の画像を Textual Inversion で少しだけ学習すると、作られる絵がなんとなく片寄って便利かも? 画風ガチャ。

画像は順に、学習した絵 → 3 iteration目のptで出力される絵 → 18 it目 → 321 it目

呪文は「1girl, cat ears, (pt名:1.00)」

0 0

Trying some different style
I am still wondering what would be a the best way to learn a style, textual inversion, dreambooth, vae, hypernetwork?



1 12

¨Yoga Lesson Number XII¨


Appropriation from the traditional design of the card ¨The Hanged man¨ to decontextualize its meaning to illustrate a waggish point of view of my reality.
....

12 7

¨Yoga Lesson Number XII¨


Appropriates from the traditional design of the card ¨The Hanged man¨ to decontextualize its meaning to illustrate a waggish point of view of my reality.
....

13 5

For this reason, doing a single circle “self portrait” takes on special meaning and purpose. Especially since its coherently contextualized in the description.

0 12

試しにウマ娘の画像使ってTextual Inversionでファインチューニングしてみたらなかなか良き

1 8

Note: this is a hyperbole. Perhaps some of you will understand and contextualize fan sentiments when given a more extreme comparison.

This is just to illustrate how Amy's fans feel and clarify complaints concerning her. This can apply to other characters too, not just Amy Rose.

3 13

GN 🔥

Thank you all for getting me to front page today. I refreshed the auctions to be a price drop - starts at 125 and drops to 15 tez. 👀👀

and I am happy to announce that my tests has good results!!

https://t.co/zACLPphVDi

17 58

recontextualising old photos

81 926

v1.3 + <muacca>
呪文等はaltに記載。

Stable Diffusion WebUIでTextual Inversionでの追加学習が結構簡単にできるようになっていたので再チャレンジ!
VRoid版Muaccaさんの画像で学習したconceptを使って出力してみました!今回はうまく行っている!気がする!

3 26

🔴OMG!!! RESERVE MET!! 🔥🔳🔥

The 24-hour auction has started with the bid of 0.088 $ETH by amazing !!

Thank you so much, I'm honored and I hope to see you in my future drops ser! 🤍🙏🤍

🔳 'Uncontextual Typologies'

Link👇👇

2 19

今日1日学習させたデータから
1枚目を作成し、
img2imgでさらに修正

Textual Inversion 15,000stepでこんな感じ

0 1

No, I actually gave it four pictures of myself and taught it the token "jurph" using textual inversion.

0 4

'Uncontextual Typologies'

LAST 3 🔥🔳🔥

https://t.co/BDQZMUBooh

2 6

textual inversionとかプロンプトエンジニアリングとか色々試した結果、私の好みにナイフ投げのナイフくらいは刺さる絵をたまに出してくれるようになったので満を持して参考資料を出力させてる
これを模写してさらに私の好みに近づけるための研究をこれからやる。

1 8

・本日のSD追加学習進捗
での 追加学習、今日は とあまり変わらないやり方
1枚目の 出力画像5種を学習させて2,3,4枚目を出力
従来のtextual inversionよりかなり理想的に学習してくれる(品質が良い) 絵同士の中間を取った絵を出力してくれる

39 328