【Stable Diffusion】AnyLoraCleanLinearMix_ClearVAE and LoRA to generate the image of Hatsume Mei of My Hero Academia

スポンサーリンク

This time I tried to use AnyLoraCleanLinearMix_ClearVAE and LoRA to generate an image of Hatsume Mei of My Hero Academia!
This article is a memorandum.

contents

スポンサーリンク

abstract

Model, LoRA and prompts used to generate images of Hatsume Mei of My Hero Academiai on google colab and diffusers

1.requirement

Google Colab
Diffusers transformers==4.26.0
model : YesMix LoRA: Hatsume Mei | My Hero Academia
LoRA:Detail Tweaker LoRA

2.result

positive prompt

highres, hatsume mei(mha), 1girl, pink hair, bangs, goggles on head, yellow eyes, parted lips, blush, makeup, light smile, TANK TOP, denim pant, cityscape, light rays, glow, thighs, collarbone, narrow waist, (masterpiece), wallpaper, upper body, street photography, depth of field,

negative prompt

(worst quality, low quality:1.4), monochrome, zombie, (interlocked fingers:1.2)

seed=52528641545129

seed=53243691904945

seed=60438584027558

スポンサーリンク

3.refarence

civitai.com

【Stable Diffusion】I checked how the generated images change when diffusers change samplers

スポンサーリンク

This time we compared the images generated by each sampler that can be selected by diffusers.

contents

スポンサーリンク

abstract

How the generated image changes when diffusers change samplers.

1.実行環境

Google Colab
Diffusers transformers==4.26.0
model : yayoi_mix
sampler：ddpm / ddim / pndm / lms / euler_anc / euler / dpm

2.result

positive prompt

positive = best quality 8k 32k masterpiece photo of pretty japanese women ultra_highres (detailed beautiful skin face focus)

negative prompt

negative=distorted face, badhandv4:0.6, extra fingers, fewer fingers, extra digit, bad hands, bad finger, head cut off, Easy Negative, disfigured, ugly, mutation, deformed, deformed face, long face, cross-eyed, long neck, cross-eyed, mutated hands, polar lowres, bad proportions, gross proportions, abdominal stretch, glans, fused fingers, bad body, malnourished,  

ddpm

ddim

pndm

lms

euler_anc

euler

dpm

スポンサーリンク

3.refarence

github.com
huggingface.co
zenn.dev

【Stable Diffusion】AnyLoraCleanLinearMix_ClearVAE and LoRA to generate an image of Mikan Yuki from To LOVE Ru

スポンサーリンク

This time I tried to generate an image of Mikaan Yuki from To LOVE Ru with AnyLoraCleanLinearMix_ClearVAE and LoRA!
This article is just a reminder.

contents

スポンサーリンク

abstract

Model, LoRA and prompts used to generate images of To LOVE Ru's Yuki-Mikan on google colab and diffusers

1.requirement

Google Colab
Diffusers transformers==4.26.0
model : AnyLoraCleanLinearMix_ClearVAE
LoRA: Mikan Yuuki (結城 美柑) - To Love Ru (とらぶる)
LoRA:Detail Tweaker LoRA

2.result

positive prompt

mikanyuuki,mikan yuuki, solo, 1girl, medium breasts, (brown eyes:1.7), brown hair, hair ornament, hair scrunchie, long hair, pink scrunchie, scrunchie, (flat chest:1.2), BREAK green skirt, plaid, plaid skirt, sainan high school uniform, school uniform, skirt, shirt, white shirt, sweater vest, (yellow sweater vest:1.5),BREAK looking at viewer,BREAK indoors, classroom, (masterpiece:1.2), best quality, high resolution, unity 8k wallpaper, (illustration:0.8), (beautiful detailed eyes:1.6), extremely detailed face, perfect lighting, extremely detailed CG, (perfect hands, perfect anatomy),

negative prompt

 (worst quality, low quality:1.4), monochrome, zombie, (interlocked fingers:1.2), bad-hands-5, mutated hands, liquid fingers, extra fingers, missing fingers

seed=4550826000845

seed=20866449928235

seed=21675827484143

スポンサーリンク

3.refarence

civitai.com

【Stable Diffusion】AnyLoraCleanLinearMix_ClearVAEとLoRAで水原千鶴の服装を調整してみた

スポンサーリンク

以前に彼女お借りしますの水原千鶴の画像生成を行いました
今回は服装を調整してみました
この記事は備忘録です

目次

スポンサーリンク

この記事でわかること

google colabとdiffusers上で彼女お借りしますの水原千鶴の画像を生成する際に使用したモデルとLoRAとプロンプト

1.実行環境

Google Colab
Diffusers transformers==4.26.0
model : AnyLoraCleanLinearMix_ClearVAE
LoRA: Mizuhara Chizuru 彼女、お借りします Ichinose Chizuru 水原千鶴 Kanojo, Okarishimasu
LoRA:Detail Tweaker LoRA

2.生成結果

positive プロンプト

brown hair, long hair, bangs,brown eyes, medium breasts, BREAK knit sweater, long sleeves, long skirt, Exquisite visuals, high-definition, masterpiece, best quality

negative プロンプト

(worst quality, low quality:1.4), monochrome, zombie, (interlocked fingers:1.2), bad-hands-5, mutated hands, liquid fingers

seed=21203534824388

seed=45245481642650

seed=90468563403281
　　

スポンサーリンク

3.参考

civitai.com

【Stable Diffusion】AnyLoraCleanLinearMix_ClearVAEとLoRAでToLoveるの結城美柑の画像を生成してみた

スポンサーリンク

今回はAnyLoraCleanLinearMix_ClearVAEとLoRAでToLoveるの結城美柑の画像生成をやってみました
この記事は備忘録になります

目次

スポンサーリンク

この記事でわかること

google colabとdiffusers上でToLoveるの結城美柑の画像を生成する際に使用したモデルとLoRAとプロンプト

1.実行環境

Google Colab
Diffusers transformers==4.26.0
model : AnyLoraCleanLinearMix_ClearVAE
LoRA: Mikan Yuuki (結城 美柑) - To Love Ru (とらぶる)
LoRA:Detail Tweaker LoRA

2.生成結果

positive プロンプト

mikanyuuki,mikan yuuki, solo, 1girl, medium breasts, (brown eyes:1.7), brown hair, hair ornament, hair scrunchie, long hair, pink scrunchie, scrunchie, (flat chest:1.2), BREAK green skirt, plaid, plaid skirt, sainan high school uniform, school uniform, skirt, shirt, white shirt, sweater vest, (yellow sweater vest:1.5),BREAK looking at viewer,BREAK indoors, classroom, (masterpiece:1.2), best quality, high resolution, unity 8k wallpaper, (illustration:0.8), (beautiful detailed eyes:1.6), extremely detailed face, perfect lighting, extremely detailed CG, (perfect hands, perfect anatomy),

negative プロンプト

 (worst quality, low quality:1.4), monochrome, zombie, (interlocked fingers:1.2), bad-hands-5, mutated hands, liquid fingers, extra fingers, missing fingers

seed=4550826000845

seed=20866449928235

seed=21675827484143

スポンサーリンク

3.参考

civitai.com

【Stable Diffusion】YesMix and LoRA generated images of Kizuna Ai.

スポンサーリンク

I have previously generated cosplay images of Kizuna Ai.
technoxs-stacker.hatenablog.com

This time I tried to generate an animated image of Kizuna Ai with YesMix and LoRA!
This article is a memorandum.

contents

スポンサーリンク

abstract

LoRA and prompt with model used to generate Kizuna Ai images on google colab and diffusers.

1.requirement

Google Colab
Diffusers transformers==4.26.0
model : YesMix LoRA: Kizuna Ai キズナアイ / Kizuna Ai inc Vtuber
LoRA:Detail Tweaker LoRA

2.result

positive prompt

asterpiece, best quality, highres, 1girl, kizuna ai, long hair, brown hair, multicolored hair, short shorts, floating hair, pink hairband, white shorts, detached sleeves, navel, sailor collar, streaked hair, pink hair, blue eyes, white thighhighs, medium breasts, lace-trimmed sleeves, sleeveless shirt, striped, white sailor collar, bowtie, hair bow, swept bangs, white shirt, lace-trimmed legwear, running, park, smile, betterhands,

negative prompt

negative_hand-neg, (worst quality, low quality:1.4), monochrome, zombie, (interlocked fingers:1.2), missing fingers, bad-hands-5, mutated hands, liquid fingers

seed=74478136894945

seed=62712405028113

seed=34007951865529

スポンサーリンク

3.refarence

civitai.com

【Stable Diffusion】diffusersのsampler変更時に生成画像がどう変化するか確認してみた

スポンサーリンク

今回はdiffusersで選択できるsamplerごとに生成される画像の比較を行ってみました

目次

スポンサーリンク

この記事でわかること

diffusersのsampler変更時に生成画像がどう変化するか

1.実行環境

Google Colab
Diffusers transformers==4.26.0
model : yayoi_mix
sampler：ddpm / ddim / pndm / lms / euler_anc / euler / dpm

2.生成結果

positive = best quality 8k 32k masterpiece photo of pretty japanese women ultra_highres (detailed beautiful skin face focus)

negative=distorted face, badhandv4:0.6, extra fingers, fewer fingers, extra digit, bad hands, bad finger, head cut off, Easy Negative, disfigured, ugly, mutation, deformed, deformed face, long face, cross-eyed, long neck, cross-eyed, mutated hands, polar lowres, bad proportions, gross proportions, abdominal stretch, glans, fused fingers, bad body, malnourished,  

ddpm

ddim

pndm

lms

euler_anc

euler

dpm

スポンサーリンク

3.参考

github.com
huggingface.co
zenn.dev

4.所感

個人的にはpndmとeuler_ancの結果が良いかと思います