Sign In

Mikey (The Fuccons / OH! Mikey) マイキー (オー!マイキー)

10

Download

1 variant available

SafeTensor

36.11 MB

Verified:

Type
LoRA
Stats

201

10

Reviews
Published

Dec 17, 2023

Base Model

SD 1.5

Training
Steps: 2,035
Epochs: 185
Usage Tips
Clip Skip: 2
Strength: 0.8
Hash
AutoV2
63E201E6AB
Trigger Words
brown hair, (blue jacket:1.3), necktie, blue shorts, green striped socks, black shoes, multicolored clothes, multicolored legwear, big smile, teeth, blue eyes, forehead, (open mouth:0.3)
default creator card background decoration
Followers - 429

429

Likes - 4387

4.4K

00076-289082121 breakdomainFp16NoEmaNoVae_R2462RealisticFP16 6af0bd7153 DPM++ 3M SDE Karras 20 3.png

I was really curious to see how the OH! Mikey characters would look like as LoRAs. I may make others later. I won't make all of them, though. The dataset was 17 screenshots depicting various angles.

オー!マイキーのキャラをLoRAにしたらどうなるかなーと前から気になってたんです。後で他のキャラも作るかもです。全員は無理だと思いますが。データセットはいろんな角度からのスクショ17枚でした。

Recommended weight (おすすめウエイト):

  • 0.65 - 9

Base prompts for Realistic Model (リアル風の基本タグ):

  • Positive (ポジティヴ): brown hair, (blue jacket:1.3), necktie, blue shorts, green striped socks, black shoes, multicolored clothes, multicolored legwear, big smile, teeth, blue eyes, forehead, (open mouth:0.3)

  • Negative (ネガティヴ): (looking at viewer, looking at camera, eye contact with viewer, white background, hair between eyes, open jacket:1.5)

Base prompts for Anime Model (アニメ風の基本タグ):

  • Positive (ポジティヴ): (very short hair:1.5), brown hair, swept bangs, (blue jacket:1.3), necktie, blue shorts, green striped socks, black shoes, multicolored clothes, multicolored legwear, big smile, teeth, blue eyes, forehead, (open mouth:0.3)

  • Negative (ネガティヴ): (looking at viewer, looking at camera, eye contact with viewer, white background, hair between eyes, open jacket, bob cut:1.5)

Optional prompts (好みによるタグ):

  • green hat, (brown jacket:1.5)

  • (photorealistic:1.5), mannequin

Problems (問題点):

  • Patterns and minor details may sometimes be off.

  • Many anime models seem to often draw the hair as a bob cut.

  • Dataset image quality isn't high, so the images may be a little blurry.

  • パターンと細かい詳細がたまにずれてます。

  • アニメのモデルなら髪型がよくボブとして描かれます。

  • データセットの画像クォリティーはあまり高くないので、出るイメージは少しボケてるかもです。

Models used for screenshots(スクショに使ったモデル):

I used FreeU for the screenshots. スクショには全部FreeUを使いました。

  • BreakDomain NoEma NoVae R2462 Realistic FP16

  • ExpMixLine v3

  • CalicoMix DC v4.0