Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

「scale」を含む日記 RSS

はてなキーワード: scaleとは

2024-10-04

Temu Affiliate Program [acp8567098] | Join & Earn $100000 In a Month

Temu Affiliate Program [acp8567098] | Join & Earn $100000 In a Month

Have you thought about joining the Temu affiliate program? You may have seen lots of buzz in TikTok and other social media platforms about the Temu affiliate program which pays a higher commission rate than any other e-commerce platform. So if you thought about joining the Temu Influencer or Temu Affiliate program then you will find all the details here.

I've seen my fair share of affiliate programs come and go. But when I stumbled upon the Temu affiliate program, I knew I had found something special. In this article, I'll share my personal journey with Temu's affiliate program and provide you with all the insider knowledge you need to succeed in this exciting opportunity.

Having worked with countless affiliate programs across various niches, I can confidently say that Temu's offering stands out from the crowd. Whether you're a seasoned affiliate marketer or just starting your journey in the world of online income, this guide will help you navigate the ins and outs of the Temu affiliate program and maximize your earning potential.

Temu Affiliate Program

The Temu affiliate program is a game-changer in the world of e-commerce affiliate marketing. Launched by the rapidly growing online marketplace Temu, this program offers a unique opportunity for content creators, influencers, and marketers to earn substantial commissions by promoting Temu's vast array of products.

What sets Temu's affiliate program apart is its generous joining bonus, which ranges from $55 to $75 depending on your social media following. This initial boost can be a great motivator for new affiliates and a nice addition to the earnings of established marketers.

One of the most impressive aspects of the Temu affiliate program is its global reach. Available in 66 countries, it opens up a world of possibilities for international marketers and those targeting diverse audiences. This wide availability ensures that you can tap into various markets and maximize your earning potential on a global scale.

Temu Affiliate Program Review

After participating in the Temu affiliate program for several months, I can confidently say that it's one of the most rewarding programs I've encountered. The combination of high-quality products, competitive commission rates, and user-friendly tools makes it a top choice for affiliates of all levels.

One of the standout features of the program is its intuitive dashboard. As someone who's dealt with clunky, outdated interfaces in the past, I was pleasantly surprised by how easy it was to navigate Temu's affiliate portal. Tracking your earnings, accessing promotional materials, and generating affiliate links are all straightforward processes that don't require a tech degree to figure out.

Another aspect I appreciate is the wide range of products available for promotion. From trendy fashion items to innovative gadgets and home decor, there's something for every niche and audience. This variety has allowed me to diversify my promotional efforts and tap into different markets, significantly boosting my overall earnings.

The support team at Temu has also been exceptional. Whenever I've had questions or encountered issues, their responsive and knowledgeable staff have been quick to assist. This level of support is crucial, especially when you're trying to maximize your earnings and optimize your strategies.

What is Temu Affiliate Program?

The Temu affiliate program is a partnership opportunity offered by Temu, a popular e-commerce platform known for its wide range of affordable products. As an affiliate, you earn commissions by promoting Temu's products and driving sales through your unique affiliate links.

What makes Temu's program unique is its focus on social media and content creation. They understand the power of influencer marketing and have tailored their program to cater to this modern approach. Whether you're a blogger, YouTuber, Instagram influencer, or TikTok star, Temu provides the tools and support you need to monetize your audience effectively.

The program operates on a cost-per-action (CPA) model, meaning you earn a commission when a user clicks on your affiliate link and makes a purchase. This model ensures that you're rewarded for driving actual sales, not just traffic.

How to Join Temu Affiliate Program?

Joining the Temu affiliate program is a straightforward process. Here's a step-by-step guide based on my personal experience:

Visit the Temu affiliate program website.

Click on the "Join Now" or "Apply" button.

Fill out the application form with your personal and professional details.

Provide information about your promotional channels (website, social media accounts, etc.).

Read and agree to the terms and conditions.

Submit your application.

Wait for approval (usually takes 1-3 business days).

Once approved, log in to your affiliate dashboard and start promoting!

Remember, Temu values quality over quantity when it comes to affiliates. Make sure to highlight your unique selling points and explain how you plan to promote their products effectively in your application.

How to Sign Up for Temu Affiliate Program

The sign-up process for the Temu affiliate program is similar to joining, but here's a more detailed breakdown of what you can expect:

Navigate to the Temu affiliate program homepage.

Look for the "Sign Up" or "Create Account" button and click it.

Choose between individual and business account options.

Enter your email address and create a strong password.

Provide your full name, address, and phone number.

Select your primary promotion method (blog, social media, email marketing, etc.).

Enter your website URL or social media handles.

Specify your niche or the main categories of products you plan to promote.

Describe your promotional strategy and expected monthly sales.

Agree to the program's terms and conditions.

Verify your email address by clicking the link sent to your inbox.

Complete any additional verification steps if required.

Once you've completed these steps, your application will be reviewed by the Temu team. If approved, you'll gain access to your affiliate dashboard and can start generating affiliate links.

Temu Affiliate Codes

Temu affiliate codes are unique identifiers assigned to each affiliate. These codes are incorporated into your affiliate links and help Temu track the sales you generate. Here's what you need to know about Temu affiliate codes:

Uniqueness: Each affiliate receives a unique code to ensure accurate tracking and commission attribution.

Integration: Your affiliate code is automatically included in the links generated through your dashboard.

Custom Links: You can create custom short links that include your affiliate code for easier sharing on social media.

Multiple Codes: Some affiliates may receive multiple codes for different campaigns or platforms.

Code Protection: Never share your affiliate code publicly to prevent unauthorized use.

Performance Tracking: Use your affiliate code to monitor the performance of different promotional channels.

Special Promotions: Temu occasionally provides special promo codes for affiliates to boost sales during specific periods.

Pro Tip: I've found that creating memorable, niche-specific custom links with my affiliate code embedded has significantly increased my click-through rates, especially on platforms like Instagram and TikTok where direct linking can be challenging.

How Much Is Temu Affiliate Program Commission Rate?

One of the most attractive aspects of the Temu affiliate program is its competitive commission rates. While the exact rates can vary depending on product categories and promotional periods, here's a general overview:

Standard Commission: 5-10% on most products

Electronics: 3-5%

Fashion and Accessories: 7-12%

Home and Garden: 6-10%

Beauty and Personal Care: 8-15%

It's worth noting that Temu often runs special promotions where commission rates can be boosted by an additional 1-3%. Keep an eye on your affiliate dashboard for these opportunities.

In my experience, the key to maximizing your earnings is to focus on high-ticket items and product categories with higher commission rates. I've had great success promoting fashion accessories and home decor items, which tend to have higher margins and, consequently, better commission rates.

Temu Affiliate Program Payments Method

Temu understands the importance of timely and convenient payments for affiliates. They offer several payment methods to cater to affiliates from different countries:

Direct Bank Transfer: Available in most countries, this is the most common and often the fastest method.

PayPal: Perfect for international affiliates, PayPal offers quick and easy transfers.

Wire Transfer: Ideal for large payments, especially for high-earning affiliates.

Digital Wallets: Some regions have access to popular digital wallet options.

Check: Available in select countries, though less common due to longer processing times.

Payment Threshold: The minimum payout threshold is typically $50, which is quite reasonable compared to many other programs I've worked with.

Payment Schedule: Temu processes payments on a monthly basis, usually within the first week of the following month.

Currency: Payments are generally made in USD, but some local currency options may be available depending on your location.

Pro Tip: I always recommend setting up at least two payment methods. This ensures you have a backup in case there are issues with your primary method, preventing any delays in receiving your hard-earned commissions.

Temu Affiliate Program Link

Your Temu affiliate program link is your golden ticket to earning commissions. Here's what you need to know about creating and using these links effectively:

Link Generation: You can create affiliate links directly from your dashboard for any product on Temu.

Deep Linking: Temu supports deep linking, allowing you to create affiliate links for specific product pages, categories, or even search results.

Link Cloaking: You can use link cloaking services to create shorter, more appealing links that still contain your affiliate code.

Link Tracking: Temu provides detailed analytics for each of your links, helping you understand which products and platforms are performing best.

Mobile Optimization: All Temu affiliate links are mobile-friendly, ensuring a seamless experience for users on smartphones and tablets.

Social Media Integration: Temu offers special tools to create social media-friendly links that work well on platforms with link limitations.

QR Codes: You can generate QR codes for your affiliate links, perfect for offline promotions or print media.

Best Practices:

Always disclose your affiliate relationship when sharing links.

Test your links regularly to ensure they're working correctly.

Use compelling call-to-actions (CTAs) with your links to encourage clicks and purchases.

What is Temu Affiliate Influencer Program?

The Temu affiliate program has a special focus on influencers, recognizing the power of social media in driving sales. As someone who's worked extensively in influencer marketing, I can attest to the effectiveness of Temu's approach. Here's what you need to know about their influencer program:

Tiered System: Temu uses a tiered system based on follower count and engagement rates. Higher tiers often receive better commission rates and exclusive perks.

Custom Campaigns: Influencers can work with Temu to create custom campaigns tailored to their audience.

Product Seeding: Top-performing influencers may receive free products to review and promote.

Exclusive Discounts: Influencers often get access to exclusive discount codes to share with their followers.

Performance Bonuses: Temu offers performance-based bonuses for influencers who exceed their sales targets.

Content Support: The program provides content ideas, best practices, and sometimes even professional content creation support.

Cross-Promotion: High-performing influencers may be featured on Temu's official social media channels.

To make the most of the influencer program, focus on creating authentic, engaging content that showcases Temu products in real-life scenarios. I've found that behind-the-scenes content and honest reviews tend to perform particularly well.

How To Become a Temu Affiliate?

Becoming a successful Temu affiliate involves more than just signing up and sharing links. Here's a strategy I've developed over my years in affiliate marketing, tailored specifically for Temu:

Choose Your Niche: Focus on product categories that align with your audience's interests and your expertise.

Build Your Platform: Whether it's a blog, YouTube channel, or Instagram account, create a strong foundation for your content.

Create Valuable Content: Develop high-quality, informative content that genuinely helps your audience.

Understand Your Audience: Use analytics tools to understand what products and content resonate with your followers.

Optimize for SEO: If you're running a blog, make sure your content is optimized for search engines to attract organic traffic.

Engage with Your Community: Respond to comments and messages to build a loyal following.

Stay Updated: Keep up with Temu's latest products, promotions, and affiliate program updates.

Experiment with Different Formats: Try various content types like unboxing videos, product comparisons, and style guides.

Be Transparent: Always disclose your affiliate relationship to maintain trust with your audience.

Track and Analyze: Regularly review your performance metrics and adjust your strategy accordingly.

Remember, success in affiliate marketing doesn't happen overnight. It requires consistency, creativity, and a genuine desire to provide value to your audience.

How to Make Money on Temu Affiliate Program?

Based on my experience, here are some effective strategies to maximize your earnings with the Temu affiliate program:

Leverage Seasonal Promotions: Temu often runs special deals during holidays and shopping events. Plan your content calendar around these to boost sales.

Create Product Bundles: Showcase how different Temu products work together to increase average order value.

Utilize Email Marketing: If you have an email list, send curated product recommendations to your subscribers.

Explore Different Social Platforms: Don't limit yourself to one platform. Test your promotions across various social media channels to find what works best.

Host Giveaways: Partner with Temu to host product giveaways, increasing engagement and attracting new potential customers.

Produce Comparison Content: Create videos or blog posts comparing Temu products with similar items from other retailers.

Focus on High-Commission Products: While it's important to promote products your audience Permalink | 記事への反応(0) | 01:15

2024-09-22

辛さを表す単位として最も一般的なのはスコヴィル値」(Scoville scale)


スコヴィル値とは

スコヴィル値は、唐辛子の辛さを測定するための単位です[1][3]。正式にはスコヴィル辛味単位(Scoville heat units、略称SHU)と呼ばれます[3]。この単位は、1912年アメリカ化学者ウィルバースコヴィルが考案したスコヴィル味覚テストに由来しています[3]。

測定方法

当初の測定方法:

現在の測定方法:

代表的唐辛子スコヴィル値

1. 鷹の爪: 40,000~50,000 SHU[4]

2. 韓国辛味唐辛子: 20,000~50,000 SHU[4]

3. タバスコソース: 2,500~5,000 SHU[4]

4. ハバネロ: 100,000~350,000 SHU[4]

5. キャロライナ・リーパー: 約3,000,000 SHU[4]

スコヴィル値の特徴

スコヴィル値は、辛さを数値化することで、異なる唐辛子の辛さを比較したり、料理加工食品の辛さを表現したりする際に広く使用されています。ただし、個人の感じ方には差があるため、あくまでも参考値として捉えることが重要です。

Citations:

[1] https://aminaflyers.amina-co.jp/list/detail/1191

[2] https://www.unagi-koubou.jp/contents/blog/2087.html

[3] https://ja.wikipedia.org/wiki/%E3%82%B9%E3%82%B3%E3%83%B4%E3%82%A3%E3%83%AB%E5%80%A4

[4] https://canaeru.usen.com/diy/p944/

[5] https://macaro-ni.jp/29198

[6] https://jirei.bzlog.jp/othertopics/trivia-unit5/

2024-09-09

グラビア動画の新しい楽しみ方を発見してしまった

DMMTV を PCブラウザで見てる時しかできない(かつHTML/CSSリテラシ必要な)芸当なんだが、

ブラウザ開発者ツールでvideoタグスタイルを変更して拡大表示するんだ。

大画面で観るのもいいが、小画面で局所ズームするのもなかなかエロいぞ。

transform: scale(2);

倍率はまあ 2~4 くらいの好きなところでよろしい。

(ちなみにブラウザズーム機能を使っても、局所拡大はできない。動画サイズCSSで相対指定されているので、画角は画面の大きさを超えられないんだ)

まり大きくしすぎると見たいところまでフレームアウトしっぱなしになるので、3以上はおすすめしない。

次に、どのあたりを拡大するかだが、それはこのように指定する。

transform-origin: 50% 75%;

原点をXY相対座標で指定するだけだ。上の例だとヨコは中央、タテは画面下寄り4/3の点を中心に拡大する。股間に注目したい人向きの設定だ。オッパイに注目したい人はY軸を中ほどにするといいだろう。

単純拡大なので解像感が落ちてしまうのが難点だが、接写好きの人は一度試してみてほしい。

2024-07-21

anond:20240721172506

おお、なるほどやな。Premiere Pro使っとると、そんなん分からへんことあるよな。ちょっとだけ考えてみるわ。

まず、今の説明からするとシーケンス作成ちゃんと700 x 1000のサイズでできとるみたいやな。でも、使おうとしてるときサイズが変わるのは、いくつか原因が考えられるで。

1つ目は、作成したシーケンスを他のメインシーケンスドラッグ&ドロップするときに、自動スケールが変更される場合や。Premiere Proはデフォルト設定で、新しくドラッグするシーケンス現在プロジェクト解像度に合うように自動的に調整することがあるんや。

その場合やったら、プロジェクト設定の中で「Scale to Frame Size」ってのをオフにせなあかん。これでドラッグしたシーケンスがそのままのサイズで表示されるようになるはずや。

もう一つの可能性は、シーケンス同士のフレームサイズ比率が違うからPremiere Proが自動的にサイズを調整しとるんかな。それは、メインのシーケンスとサブシーケンスが違う解像度を持っとるときに起こりやすいんや。

どっちにしても、エフェクトコントロールスケールをいじることで対応可能やけど、基本設定で解決できると楽やんな。

上手くいかなかったらもう少し詳しい情報教えてくれたらまた考えるで。知らんけど。

2024-07-03

スカのプレイリストが200曲になったので

また公開してみる

コイツの続きになります https://anond.hatelabo.jp/20240416183932

スカ好きもっと増えろ

101, 浅草ジンタ/蒲田行進曲

102, 浅草ジンタ/すたこらさっさと早よ逃げよ

103, 浅草ジンタ/自・燦々

104, 上坂すみれ/七つの海よりキミの海

105, 奥村愛子/いっさいがっさい・密命ヴァージョン

106, オレスカバンド/AZAMI Spirit

107, 東京スカパラダイスオーケストラ/めくれたオレンジ

108, 東京スカパラダイスオーケストラ10-FEET/閃光

109, 東京スカパラダイスオーケストラ/Pride Of Lions

110, 東京スカパラダイスオーケストラ/Burning Scale

111, 東京スカパラダイスオーケストラ/The Last Ninja

112, A Contratiempo Ska-Jazz/Kamusari

113, beat sunset/なまくらさ

114, beat sunset/Progress

115, Blechreiz,Memoria Insuficiente Ska/Bumble Bee

116, Bluekilla/Misirlou

117, Desorden Público/Música de Fiesta

118, Distemper/All Colors Crew

119, Eastern Standard Time/Sei Pazzo

120, El Pacto/Ciclope Tin Tan

121, Fanfare Ciocarlia/Cruzzzando el Compooo

122, Frau Doktor/Birks Works

123, Gypsy Ska Orquestra/El Majestuoso Imperio de Satanas

124, Karamelo Santo,Los Calzones/El Ritmo Indecente

125, La Parranda Magna/Crueldad Animal

126, LÄ-PPISCH/MAD GIRLS

127, Let's Go Bowling/Mayhem

128, Let's Go Bowling/Hot-Buttered

129, Los Calzones/Caras y Caretas

130, Los Calzones/La Felicidad

131, Los Calzones/Satanica

132, Los Calzones/Un Rey Ska

133, Los De Abajo/De Marcha

134, Los Tres Puntos/Perpetua

135, Madness/Night Boat to Cairo

136, MAHO堂/ステキ

137, Mocca Garden/บรึ๋ย (กลัวนะครับ OST ปอบหน้าปลวก)

138, Mocca Garden/ฮักสาวตาซัง

139, mule train/INTENCE SKA

140, Nancy Ska Jazz Orchestra/Laisse tomber less filles

141, New Lion Ska/Quizas tu amor

142, New York Ska-Jazz Ensemble/Blowout

143, No Water Please/MA&PA

144, Off Beat Xperiment/Intentions

145, Oi-SKALL MATES/I can't Stay with You

146, Oi-SKALL MATES/Enjoy yourself

147, Oi-SKALL MATES/TALIO

148, Operation Ivy/Yellin' In My Ear

149, Out Of Control Army,Dr.Shenka/Al Fondo del Mar

150, Out Of Control Army,The Locos/Quién devuelve el tiempo?

151, Out Of Control Army/Lilith

152, Plegue Of Happiness/REBEL YOUTH

153, Plegue Of Happiness/TUNJAL

154, POTSHOT/BE ALIVE

155, Quit Ska Jazz/Cold Duck Time

156, Random Hand/Devil With a Microphone

157, Reel Big Fish/Don't Stop Skankin'

158, ROUTE85/Over the Limit

159, Russkaja/Dobrij Abend

160, Russkaja/Rock'n Roll Today

161, Russkaja/Zirk

162, Russkaja/Love Revolution

163, Russkaja/Gop Stop

164, Salon Victoria/Salome

165, Salon Victoria/Tumbao Fregao

166, Salon Victoria/Mr.Yamamoto

167, Save Ferris/Come on Eileen

168, Sessiones Reggae/Barracas Ska Tango

169, SHOW-SKA/CHAMBARA MAN

170, SHOW-SKA/ナオミの夢

171, SKA FREAKS/From This Room

172, Ska'n'Ska/Money

173, Ska-P/El Chupacabras

174, Skarlatines/Rebel Girl

175, SkaZka Orchestra/Bormental

176, Sonora de Llegar/Hipo ska

177, Sonora Skandalera/Eslabón

178, St.Petersburg Ska-Jazz Review/Monochrome Memories

179, St.Petersburg Ska-Jazz Review/4.70

180, St.Petersburg Ska-Jazz Review/Four

181, St.Petersburg Ska-Jazz Review/Trip Back To Childhood

182, The Autocratics/Solid State SKA

183, The Determinations/TIPSY

184, The Determinations/New day

185, The Interrupters/Rumors and Gossip

186, The Interrupters/Babylon

187, The Locos/Espacio Exterior

188, The Middle Volga Social Club/Give Me the Way

189, The OLDTONES/Tocamos SKA

190, The OLDTONES/Monopolistic

191, The Rude Monkey Bones/La persecución

192, The Scofflaws/Rudy's Back

193, The Skatalites/Bamako Ska

194, The Specials/Nite Klub

195, TRI4TH/FULL DRIVE

196, Vallanzaska/Si si si no no no

197, Voodoo Glow Skulls/Fat Randy

198, Voodoo Glow Skulls/Left For Dead

199, Киоск/Маршрут 400

200, ผดุง ทรงแสง (แจ๊ส),ThepPhithak/ป๊าด 8

おススメ引き続き募集中です

個人的には、2000年以降のジャマイカスカとか、上海あたりの大編成中華スカバンドとか、ホーミー入りのモンゴルスカとかあれば聞いてみたいところ

2024-05-10

anond:20240510180014

今、世間一般認識はどうなっているかの話をしているだけ

2022年の侵攻は2014年から続くロシアによる一連の侵攻の一部だというのが世間一般認識ですね

少なくともお前個人が一人で勝手にわめいてることよりはWikipediaの方がはるかに「世間一般認識」に近いわな

https://en.wikipedia.org/wiki/Russo-Ukrainian_War

The Russo-Ukrainian War is an ongoing war between Russia and Ukraine, which began in February 2014. Following Ukraine's Revolution of Dignity, Russia occupied and annexed Crimea from Ukraine and supported pro-Russian separatists fighting the Ukrainian military in the Donbas war. The first eight years of conflict also included naval incidents, cyberwarfare, and heightened political tensions. In February 2022, Russia launched a full-scale invasion of Ukraine and began occupying more of the country.

 

それを前提にして元々の話に戻って反論は?

ウクライナ侵攻も始まる前はありえないと言われていたのが

2014年からずっと戦争状態なのにあり得ないとか言っていたヤツが頭おかしいだけだね

2024-04-12

https://edition.cnn.com/2024/04/11/sport/ippei-mizuhara-shohei-ohtani-investigation/index.html

Any winnings were not placed in Ohtani’s account, but rather in Mizuhara’s personal account, Estrada said.

Mizuhara even impersonated Ohtani to try to get a bank to approve large wire transfers to bookmakers, Estrada said.

Mr. Mizuhara committed fraud on a massive scale,” the prosecutor said.

Authorities obtained recordings of phone calls during which Mizuhara allegedly lied to bank employees about being Ohtani and gave them personal biographical information in order to gain approval to transfer the money, Estrada said.

賞金は大谷の口座ではなく、水原個人口座に振り込まれたとエストラーダ氏は言う。

エストラーダ氏によれば、水原氏は大谷選手なりすまして、ブックメーカーへの多額の電信送金を銀行承認させようとさえしたという。

水原氏は大規模な詐欺を働いた」と検察は述べた。

エストラーダ検察官によれば、水原容疑者が送金の承認を得るために、銀行従業員大谷であると嘘をつき、個人的な経歴を話したとされる電話の録音を当局は入手したという。

2024-02-21

anond:20240221003832

直訳版



A long time ago, in a certain country, there was a bastard named Cinderella. The reason why she is called Cinderella is because she sprinkled her own ashes on her own head at her mother's funeral, which led to her being called the ash-coverer, which is why her real name is Cinderella. I lost

The husband who lost his wife was also a bastard, and the day after the funeral he cheated on a woman three years younger than him, had sex with her, and even got married.

I have decided. Then, to her Cinderella

She now has an older sister and a younger sister who are not related by blood.

Her sister was a bitch too, she came to Cinderella's house

After a few days, she thought, ``Wouldn't it be more convenient to have an outbuilding?'' and cut down Cinderella's mother's grave in her garden and the hundreds of years old tree that was next to her grave, and built an outbuilding. We forced construction.

One day, the bastard's family receives news of a ball from the castle. But before Cinderella could read it, her sister broke the news and said, ``There's no way you can go.'' Cinderella didn't seem to have any intention of going either, saying, ``I had no intention of going in the first place.I'd rather go to hell.''

Then, on the day her mother was admitted to the hospital to give birth, her sister went to the ball. In the garden, the construction of an outbuilding was progressing as usual.

Let’s light a fire here and make a bonfire.”

She thought about that, but Cinderella stopped her. Suddenly, a witch appeared

``Cinderella, why aren't you going to the ball?'' asked the witch.

"What's the point of telling Teme?"

Cinderella returned with an ax in her hand

Then, for some reason, Cinderella was wrapped in a jewel-like dress.

"If you don't go to the ball, that dress will kill you. Now go."

Cinderella doesn't want to die in a place like this, so she reluctantly decides to go to the castle.

In the castle, some asshole like her sister was looking for a one-night stand and was dancing wildly. No one really praised others, and in their hearts they all seemed to think that they were the center of the world.

Cinderella didn't dance, but killed her time by smoking a cigarette at the end of her room. There, she was approached by a man wearing formal clothes that were flashier and more expensive than anyone else, and said to her, ``Miss, won't you dance?''

She is Cinderella with a cigarette in her mouth

She said, ``I only dance with the Grim Reaper and bad luck.''

she answered. For some reason, the man fell in love with her at first sight, and she said, ``Let's have some fun in the other room.Here, come.'' She tried to force Cinderella to take her to another room in the castle. However, Cinderella pressed the cigarette she was holding into her hand.

"Her hand slipped," she laughed.

When she pressed the cigarette, the sound in her voice was so pitiful that no one tried to get close to him or dance with him. Of course, no one even knows that such a pitiful man is the prince who owns this castle.

Eventually, she had enough, Cinderella threw her shoes on her stairs and went back to her house in her bare feet. Her shoes were made of glass, so they shattered into pieces.

The next day, while her servant was cleaning the castle, she found one shoe on the stairs. Obviously thinking this was strange, she secretly examined her shoes and discovered that they had been smeared with a deadly amount of poison.

Rumors spread within the castle that this must be an attempt to kill the prince, and soon a large-scale investigation was conducted. Even though it's a survey, it's simple.

The idea was to create a replica of a shoe and imprison the person who matched it perfectly as a suspect.

And one after another, people were imprisoned just because of their size. Some of them had not been to the ball, but they thought it must be a lie and were forcibly taken away.

Eventually, an investigation came to Cinderella's house. That's when we all looked at her mother and child after she was discharged from the hospital, and of course her sister didn't fit.

So all that was left was Cinderella. Then, her sister said to Cinderella, ``Hell suits you better than this world.''

She was of course the perfect size for Cinderella.

However, she protested her innocence, saying that I hadn't gone to the ball and there was no dress anywhere in the house to prove it.

However, the prince showed him an empty pack of cigarettes that had been discarded in a trash can and said, ``This is the same brand that the criminal was smoking that time.Why?''

Then Cinderella grinned and

I unbuttoned all the buttons on her jacket.

She said, ``After all, I feel like dancing with the Grim Reaper. Right, little prince?'' she said.

By the time the princes realized it, it was already too late.

The grenade that fell at his feet blew up Cinderella and the prince.

The time was around noon. Bells rang throughout the city and residents prayed for lunch.

The flying fragments of the shoe became a weapon, a sad fragment that could only hurt someone, and no one could imagine its original form.

end

2023-11-19

Notice

■ Regarding the free release of the new species previewed and the reason for making the support site the main platform for releasing new species:

While the free release on the Kurobinega website is at my discretion, I'm considering doing so roughly once every three or two paid releases.

The following might not be particularly relevant to English-speaking fans, or in other words, to everyone. The reason I'm focusing on early releases is that I want to limit the exposure of MGE and continue creating content on a smaller scale. First of all, I don't want MGE to be a hugely popular restaurant that anyone can enter, nor do I wish to become an internet celebrity. For me, MGE is akin to a reasonably popular local restaurant where like-minded individuals who share my interests gather. By scaling down, my production speed has increased. The period from Dragonewt to Tai Sui was shorter, and the next species should be released sooner. What matters most to me is creating the MGE I want to craft, drawing pictures, and developing games in the world of MGE. I'm delighted when people who value MGE's themes as I do get to see it. Receiving money on support sites like these, allowing me to spend more time on production, is truly appreciated.

Using the stuffed animal analogy again, it truly makes me happy when someone who genuinely loves and cherishes stuffed animals sees the ones I've created. Yet, I can't see those who say they adore stuffed animals but also tolerate and respect those who take pleasure in mutilating them as people who share the same love for stuffed animals as I do. I don't have the inclination to actively show them the stuffed animals I've made.

I believe English-speaking individuals, when witnessing a stuffed animal's head being torn off or the pleasure derived from it, would straightforwardly label it as "crazy." However, it's different in Japan, especially on social media. Even if they dislike watching the stuffed animals being hurt, they'll say, "Let's respect those who enjoy tearing them." This, despite seeing me, someone desperately protecting my stuffed animals, getting attacked by such individuals. There's a prevailing notion in the otaku community, especially on Japanese SNS (Social Media), that "All fetishes must be respected." Speaking out against this, even as a victim, can result in backlash. Not everyone thinks this way, but it's the dominant mindset. Disheartened by this, I didn't want to show them MGE anymore. When you grow to dislike a place, it's only natural to distance yourself, and that's what I've done.

However, this is a problem in Japan. In English-speaking regions, fans typically call out what they dislike, labeling it as "crazy." I have many English-speaking friends who've genuinely helped me in trying times. The MGEwiki admin is also a dear friend. Depending on the English-speaking fan community's dynamics, I might consider releasing content on MGEwiki a month after the early release.

2023-11-05

anond:20231104194027

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0265798

ASDTD比較により、ASD場合、生の眼の接触中に右側の頭頂葉活動が減少し、右側の腹側側頭頭部の活動が増加し、生の眼の接触に対する非典型的な神経システムによる脳間の一貫性が低下することが示されました。ASDにおける眼の接触中の右側頭頂葉領域の低活性は、社会的パフォーマンスゴールドスタンダード尺度と神経反応との相関に関連しており、社会的能力が低下するにつれて、右側頭頂葉領域での実際の眼の接触に対する神経応答も低下することを示唆しています

この研究の主な結果は以下の4つの主要な発見からなり、ASDにおける生の対話的な顔を見る際の代替神経プロセス仮説を支持しています

1) ASDでは腹側後頭頭領域内の神経システム活性化されており、一方、通常発達者(TD)では生の対話的な顔を見る際には頭頂頭領域と外側前頭領域が関連していました。

2) ASDでは、視覚の取得要因も生の顔処理の困難に寄与する可能性があると示唆され、眼球固定の位置の変動が増加し、生の顔に対する瞳孔反応も増加していました。

3) 生の眼と眼の接触中に右側の頭頂頭領域での神経応答の反対相関とADOS-2およびSRS-2のスコアとの関連からASDに関連した社会的パフォーマンス生物学的基盤が示唆されています

4) TDにおける生の眼と眼の接触に対する脳間同期はASDでは観察されなかったため、共有の顔情報の脳間伝送の減少を示唆しています

注釈

1. “ASD”:自閉症スペクトラム障害(Autism Spectrum Disorder)の略語で、神経発達における病的な特徴を示す一連の症状の集まりです。

2. “腹側後頭頭領域”:脳の特定領域で、視覚情報処理に関連しています

3. “TD”:通常発達(Typically Developed)のことで、ASDではない一般個人を指します。

4. “視覚の取得要因”:視覚情報収集や処理に影響を与える要因。この文脈では、ASDにおいて視覚情報収集の異常が示唆されています

5. “眼球固定の位置の変動”:視線が注がれた対象への眼球の移動や位置の変化を指します。

6. “瞳孔反応”:瞳孔の大きさや収縮が外部刺激に対してどのように変化するかを示す反応。この文脈では、生の顔に対する瞳孔の反応がASDにおいて増加していることが指摘されています

7. “ADOS-2”:自閉症診断観察スケジュール、第2版(Autism Diagnostic Observation Schedule, 2nd Edition)の略語で、ASDの診断に使用される標準的評価ツールです。

8. “SRS-2”:社会的適応尺度、第2版(Social Responsiveness Scale, Second Edition)の略語で、社会的適応能力評価する尺度です。

9. “脳間同期”:異なる脳領域間で情報が同時に伝達されること。この文脈では、ASDTDの間で脳間同期の違いが指摘されており、ASDにおいて共有の顔情報の脳間伝送が低下している可能性が示唆されています

2023-10-20

Refuting the “Israeli” Claims Regarding Evading their responsibility for the Massacre of the Gaza Baptist Hospital

Today, the “Israeli” Occupation Entity is trying to evade their responsibility for the crime of bombarding the “Baptist” Arab National Hospital in Gaza City, which the Israeli” Occupation Forces (IOF) committed on the evening of Tuesday, 17 October 2023. This heinous crime was committed against innocent Palestinian civilians while taking the hospital as a shelter from the flames of the “Israeli” brutal aggression, which left nowhere safe in the besieged enclave. It is necessary to affirm that, the Baptist Hospital belongs to the Anglican Episcopal Church in Jerusalem, and was built before the occupation of Palestine.

It is obvious that the “Israeli” enemy has been spreading lies since the very beginning of his destructive war on our people, when he with no single evidence claimed that the Palestinian resistance killed children, cut off heads and raped women. In continuation of this series of lies, they tried to evade their crime, attributing it to one of the resistance factions. Accordingly, we present some of the conclusive evidence to prove the “Israeli” Occupation Entity responsible for this heinous crime:

1) It is no secret that the IOF, several days ago, threatened several hospitals in the Gaza Strip, contacting each hospital separately and requesting their evacuation and holding the hospitals directors responsible for the consequences of neglecting the threats. In fact, there are dozens of clear statements from the IOF spokespersons in this regard.

2) Since the beginning of the ongoing aggression, the “Israeli” army has ignored the principle of distinguishing between civilian and military targets. Thus, the bombardment has systematically targeted emergency services, ambulances, civil defense facilities, schools, mosques, and churches.

3) The IOF contacted the directors of 21 hospitals in the Gaza Strip, especially those are located in the Gaza and the North of Gaza governorates, most notably: (Al-Awda, the Indonesian, Kamal Adwan, the Kuwaiti, Al-Quds, and Al-Mamadani), asking them to evacuate immediately, given that the hospitals are located within the geographical scope of “Israeli” military operations. In this regard, the official spokespersons for the IOF plus a number of hospital directors conducted interviews live on Al Jazeera, revealing the premeditated intention of the IOF to target hospitals in the Gaza Strip.

4) On 14 October, 2023, at 20:30, the IOF fired two shells towards the Baptist Hospital, and the next morning they called the hospital director, Dr. Maher الصفحة 2 من 3

Ayyad, saying to him: “We warned you last evening with two shells, so why have you not evacuated the hospital until this moment?!” Following that call, The hospital director contacted the bishop of the Evangelical Church in Britain and informed him of the incident, who in turn contacted international organizations before sending the hospital a message reassuring them that they could remain in the hospital. Yet with no warning, the IOF airstrikes returned, on Tuesday evening, to carry out the massacre against the hospital and the shelter-seekers of the displaced innocent Palestinian civilians.

5) Immediately after the massacre, the IOF spokesperson quickly published a statement on his page on the “X” and “Telegram” platforms at 21:17, which stated, “We had warned the evacuation of the Baptist Hospital and five other hospitals so that the Hamas terrorist organization would not take as a safe haven”. That statement is a clear claim of this massacre, and it is documented with a “screenshot” image attached to his page on Telegram, but he quickly deleted the post after seeing the massive scale of the massacre for the large number of victims, and the angry responses of the Arab, regional and international communities. Later on, he disavowed it, publish another

statement denying that he had issued the first statement.

6) Before and during the event, the resistance factions did not fire any missiles at the occupied territories, the “Israeli” sirens did not activate, and the Iron Dome missiles were not launched. Moreover, dozens of reconnaissance drones do not leave the sky of the Gaza Strip (365 km2 ), photographing and monitoring every inch around the clock. If the massacre was due to the resistance’s missiles, as the “Israeli” Occupation Entity falsely claim, why would not they show one picture their claims?!

7) The “Israeli” Occupation Entity claimed that this massacre was caused by a missile launched by the Islamic Jihad. However, we would wonder how they could identify and distinguish between the missiles of the resistance factions immediately after they were launched?!

8) The IOF military system documents and records all their operations by day, hour, minute and second, and in all previous times their media outlets came out to announce or deny much less severe massacres than this massacre, so what made them wait more than 4 hours before declaring their irresponsibility other than weaving scenarios of falsification, lies and deception?!

9) It is obviously known that the resistance’s missiles are somewhat “primitive” and do not have the destructive power that kills hundreds at one time. And throughout the history of previous confrontations and the current confrontation, no resistance missile has caused a tenth of this number of “Israeli” deaths.

10) The only video scene documenting the moment of the explosion reveals that the mass of flame and the sound of the explosion are identical to other “Israeli” bombs throughout the days of the ongoing aggression, which with no doubt proves that the “Israeli” origin of the missile.

الصفحة 3 من 3

The deliberate attack on hospitals is a war crime, stipulated in Article (8, 2, b, 9) of the Statute of the International Criminal Court. Accordingly, Hamas officially calls on the Prosecutor of the International Criminal Court to open an investigation into this crime, which falls within the framework of a genocidal war, in accordance with the provisions of Article 6 of the Basic Law. These bloody acts are also based on the policy of denial pursued by the “Israeli” Occupation Entity, as they deny the existence of the Palestinian people.

There is no doubt that impunity fuels crime, and investigating crimes is a way to protect people. Furthermore, if a legal and judicial response is necessary, it is above all that the response to this crime must be humanitarian and global. Such actions must shock the conscience of the world, or else there would no longer be a reliable international community if the “Israeli” Occupation government free to decide to bomb hospitals!

In conclusion, we are facing a genocidal massacre committed by IOF against children, women, and the elderly. The “Israeli” Occupation Entity is the only responsible for the crime, no matter how much they try to weave lies and fabrications to get away with it, as they always do.

The Islamic Resistance Movement

HAMAS-Palestine

Wednesday, 18 October 2023

本日ハマスpdf形式でテレグラムにアップしたやつの全文

ガザ・バプテスト病院虐殺責任回避に関するイスラエルの主張に反論する」

って題名

2023-07-10

anond:20230710122636

スケールする

スケールする」とは、物事プロジェクトの規模を拡大するという意味ビジネス用語ベンチャースタートアップ企業の界隈で用いられる日本ビジネススラング。「スケールアップ」。

もともと英語の「scale」は、「規模」「目盛り、物差し、定規」「尺度」といった名詞意味しか持たない。しかし、「規模の経済性」を意味する和製英語の「スケールメリット」の言葉使用が広まり日本語としての「スケール」の言葉自体に「規模が大きいこと」のニュアンスが含まれるようになった。それが動詞化したものが「スケールする」である

事業規模が拡大することを「スケールする」、規模が拡大せず維持停滞することを「スケールしない」などと表す。システムネットワーク用語「スケーラビティ」をビジネス用語転用した意味類似する。対義語として、縮小することを「シュリンクする」という。

2023-05-15

[]なんか勝手に見ているページから情報を抜き出すんだな…

GPT-4 AI Tutor Prompt: A tool for creating personalized learning experiences using GPT-4 and LangchainJS, a drag and drop UI for building LLM flows.

Pandas AI: A Python library that adds generative AI capabilities to Pandas dataframes, making them conversational.

FigmaChain: A set of Python scripts that use GPT-3 to generate HTML/CSS code from Figma designs. It also has a chatbot interface for interactive code generation.

Sui: A smart contract platform with high performance and an asset-oriented programming model based on the Move language.

Jailbreak for ChatGPT: A project that allows ChatGPT to predict the future, opine on controversial topics, and assess what is true. It may help us understand LLM bias.

ML Observability in a Notebook: A tool to uncover insights, surface problems, monitor, and fine-tune generative LLM, CV and Tabular models.

System Design Primer: A resource to learn how to design large-scale systems and prepare for the system design interview. It includes Anki flashcards.

Next.js 13 App: An open source application that showcases the new features of Next.js 13, such as router, server components, and more.

TypeScript: A superset of JavaScript that compiles to clean JavaScript output.

Media Downloader: A desktop utility to download images/videos/music/text from various websites, and more.

AI Pull Request Fixer: A tool to fix issues with AI-generated pull requests, powered by ChatGPT.

Chat Chat: An app that lets you deploy your own AI interface and chat with different LLMs in one place.

2023-05-02

(full body,whole body),ruined city,post apocalypse city,solo,(pastel pink buruma and bloomers and tube top:1.1),[boob],have a gun,

使用モデル:Anything V4.5 ネガティブ:(flill,short pants,skirt:1.1),back,gym,nipple,skirts,pussy,(以下Pixai内デフォルト)

Sampling Steps:23Sampling Method:Euler a、CFG Scale:11Seed:2934948792

2023-04-30

[][] 白緑 +1 タッチ続唱 ノーン 7-0

緑白マルチと後援の+1シナジーで序盤を防ぎ、3マナ培養オーラマナ加速から続唱の動きが強い。

5マナ続唱はそれだけでも青タッチする価値があるから必然マナ補助と素出しが狙える緑が基盤になる。etbだけじゃなくて後続も続唱するのズルい。

ノーン様は3回出て勝ったからまあボムゲー。相手にノーン様だされて爆発することもあるから許せ。うち2回が白青騎士で膠着してたところにドーンだからタフ2環境だなあという印象。

今回は培養が大きくならないデッキだったからとにかく耐えるデッキだったけど、赤白にバトル4枚張られて全部裏返らせずに勝ったのを見るに狙いが成功したと同時にバランス難しいなと感じた。


2パック2手目にノーン様が流れるんだからクイックドラフトは魔境だぜ


Deck

1 Invasion of Gobakhan (MOM) 22

1 Alabaster Host Intercessor (MOM) 3

1 Phyrexian Awakening (MOM) 30

1 Kami of Whispered Hopes (MOM) 196

2 Blighted Burgeoning (MOM) 177

2 Golden-Scale Aeronaut (MOM) 15

2 War Historian (MOM) 214

2 Botanical Brawler (MOM) 220

1 Streetwise Negotiator (MOM) 207

1 Shanna, Sisay's Legacy (MUL) 124

1 Realmbreaker's Grasp (MOM) 33

1 Tandem Takedown (MOM) 208

1 Timberland Ancient (MOM) 210

1 Elesh Norn, Grand Cenobite (MUL) 68

1 Invasion of Dominaria (MOM) 21

1 Sandstalker Moloch (MOM) 203

1 Blossoming Sands (MOM) 268

1 Knight of the New Coalition (MOM) 25

1 Imoti, Celebrant of Bounty (MUL) 108

1 Serpent-Blade Assailant (MOM) 205

7 Plains (SIR) 279

1 Island (SIR) 282

8 Forest (SIR) 291

Sideboard

1 Skittering Surveyor (MOM) 264

1 Kor Halberd (MOM) 27

2 Bonded Herdbeast (MOM) 178

1 Furnace Host Charger (MOM) 140

1 Invasion of Muraganda (MOM) 192

1 Fertilid's Favor (MOM) 186

1 Enduring Bondwarden (MOM) 14

1 Scrollshift (MOM) 34

1 Ayara, Widow of the Realm (MOM) 90

1 Render Inert (MOM) 123

2 Aerial Boost (MOM) 2

1 Ayara, First of Locthwain (MUL) 78

1 Serpent-Blade Assailant (MOM) 205

1 Bloated Processor (MOM) 93

1 Crystal Carapace (MOM) 183

1 Kithkin Billyrider (MOM) 24

2023-04-21

スケーリンニュアアイズscale in your eyes

お前の目に鱗

転じて目から鱗を落として解像度を高めよの意

2023-04-07

anond:20230407152201

スマホ連動の体重計を買う。3000円くらいからある。InFieldのAI×体組成計とか。おすすめAnkerのEufy Smart Scale P2 Pro

2023-03-14

Chat GPTお絵描きAIが大盛況でみんなきゃっきゃしてるが無限Scaleする存在人類は勝てるとは思えない。スタートレックボーグ、集合意識みたいな存在が出てきたわけ。人類組織を大きくするのに縦とか横にマネージャーとかコンサルとか飲み会とかゴマすりを差し込む必要がかの存在には全く必要ない。これから向こう数年でAIプロジェクト管理させる話が出てくるのは間違いないが、今まで見たことがないような規模のものなったとき人類は対抗できるのか

2022-11-06

ちょっと気になってた Stable Diffusion を試してみた。Web UIかいうやつ。

Anaconda を入れていたおかげか、環境構築はすんなりと行ってしまった。

GTX1060、VRAM 6G の環境

出力画像サイズ 512x512、パラメーターはよくわかんないけど初期値のまま。

品質と速度のトレードオフに関わるとこは Sampling Steps = 20、CFG Scale = 7 あたり?)

これで、20秒~30秒以内にはそれなりの画像が出る。わりと面白い

GTX1060 でもまあまあ遊べるやんな。

そんで、次は RTX3060 12G に換装して遊んでみるんよ。

2022-10-16

NovelAIが重すぎるからローカル環境にNAI環境を構築する(2022年10月16日版)(追記あり)

せっかく課金したのにユーザが増えまくっているのか滅茶苦茶重くなっていて最悪。

から流出したモデルを使ってローカルでNAI環境を構築する。

ネットには情報もだいぶ転がってるけど陳腐化した情報があまりに多いため増田にまとめることにした。

しかたらこ記事もすでに陳腐化しているかもしれないが…単純に間違ってたらトラバで教えてほしい。

もちろん自己責任。この記事を見て導入した結果何かあっても増田は何も保証しない。

英語がわかる人はこっちを見た方が早いと思う。今は導入RTAができるくらい導入は楽になっている。

https://rentry.org/nai-speedrun

推奨環境

VRAMが2GB以上あるNVIDIA製のグラフィックボードがあればローカル環境を構築できる。

GPU世代はGTX700シリーズ以降。なので一昔前のミドル級ボードでも動作するらしい。

IntelオンボードGPUでも実行する方法があるらしい(stable_diffusion.openvino)が今回は割愛する。自分で探してね。

その他の推奨環境は以下の通り。

対応OSWindows7以上(と言うがM1Macでも動作する方法があるとかなんとか)

必要な空きストレージ容量:20GB以上

インメモリ:16GB以上(VRAMもたくさん必要だが起動時にメインメモリも大量に食う。WebUI起動時にタスクマネージャを見ているとよくわかる)

スマホしか持ってないような人やこういうのがよくわからない人はNovelAIを使った方が良いと思う。

今は重いけど、きっとそのうちみんな飽きてサーバも軽くなるかもしれないし。

(追記)NovelAIリソースを確保してサーバが軽くなったかリスクを背負ってまで導入しなくても良いか

手順1:PythonGitを導入する

(追記)Pythonは当然3系。最新の奴を入れれば問題無い。

導入方法はいちいち書かないけど、「python --version」や「git -v」で

正常にバージョン情報が出る(パスがきちんと通っている)ことはちゃん確認しよう。

手順2:Stable Diffusion web UI(AUTOMATIC1111)を導入する

Stable Diffusion web UIはStable Diffusionやそれをベースとした画像生成AIを利用するためのフロントエンド

その中でも特に開発が活発でデファクトスタンダードとなっているのがAUTOMATIC1111版だ。

導入したい適当ディレクトリに対してPowerShellなどで

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

とやってやれば必要ファイルの導入が行われる。

なお、AUTOMATIC1111版は数時間単位コミットが行われるから

定期的に「git pull origin master」で更新しよう。

手順3:BitTorrent流出モデルダウンロードする

クライアントはqBitTorrentが一番楽だと思う。

ここにはさすがにmagnetリンクは書かないか各自ググって欲しい。

結構誤解されがちなことだが流出データ50GBを全部ダウンロードする必要は無い。

必要ファイルはanimefull-final-prunedディレクトリの中身とanimevae.ptから5GBちょっとくらいなんじゃないかな。

もし余裕があるならmoduleディレクトリの中身もダウンロードすればいいけど、ぶっちゃけ必要無いんじゃないか

手順4:ダウンロードした各ファイルリネーム・移動

まずはanimefull-final-prunedの中身のファイルリネーム

model.ckpt」を「animefinal-full-pruned.ckpt」のようなわかりやす名前にして、

「animevae.pt」を例えば「animefinal-full-pruned.vae.pt」のような拡張子以外は同じファイル名にする。

WebUI起動フォルダ配下の\models\Stable-diffusionリネームしたファイルを移動させれば配置はOK

ちなみにmoduleディレクトリの中身は\models\hypernetworksに移動させて使う。

それらのファイルを設定で適用させると画風が結構変わるがNovelAI再現とは関係無いみたいだ。

(追記)moduleディレクトリの中身の.ptファイルはhypernetworksという技術によって画風などを学習したものらしい。

すでに複数イラストレーターの画風を学習したptファイル作成されており議論を呼んでいる。

手順5:webui-user.batの中身に設定を追加する

自分グラボのVRAMが4GB未満の場合は「set COMMANDLINE_ARGS=」の後に

4GB未満の場合は「--medvram」、2GB未満の場合は「--lowvram」とパラメータを追加しておこう。

自分の持ってるグラボのVRAMがわからないときGPU-Zなどで調べよう。

またGTX1600系固有のバグ(単色の画像が出力される)があるらしいので

その場合は「--no-half-vae」もしくは「--no-half」や「--precision full」とパラメータを追加。

ちなみにパラメータに「--xformers」を追加してxformersを導入・使用すると

消費VRAMが減って画像生成処理時間も短縮されるので是非導入しよう。

画像からdanbooruタグAI調査するdeepdanbooruを利用する場合は「--deepdanbooru」を追加。

これらの設定は同時に複数適用させることもできる。例えば

set COMMANDLINE_ARGS=--medvram --xformers --deepdanbooru

のようになる。

手順6:webui-user.bat起動、設定変更

ターミナルPowerShellなどでwebui-user.batを起動しwebUIの初期導入と起動を行う。

過去には手動でCUDA等を導入する必要があったが、現在はこの初期導入でだいたいの導入が行われる。

ずいぶん楽にはなったがその分初期導入の時間結構長い。10分~20分くらいかかるかもしれない。

途中で導入処理がエラーで止まってしまった場合管理者権限で実行するなどして対応して欲しい。

起動ができたらSettingで以下の設定を変更してNovelAIに近づける。

Stop At last layers of CLIP modelを2に、

Eta noise seed deltaを31337にする。

これで設定は完了

おまけ:アスカテスト

設定を合わせて完全にNovelAIと同じ内容になったのかを確認するテストがある。

出力結果から海外じゃHallo Asuka Testなんて呼ばれている。

これは初期SEEDをはじめとする設定内容が完全に一致していれば同じ出力結果を得られる仕組みを利用している。

プロンプトの内容:masterpiece, best quality, masterpiece, asuka langley sitting cross legged on a chair

ネガティブプロンプトの内容:lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name

サンプリングステップ数:28

サンプリング形式:Euler

CFG Scale(プロンプトの強度):12

初期Seed2870305590

この内容で見事下の画像と全く同じ画像が出力されれば合格だ。

https://i.imgur.com/Bfl5qJB.jpg

なお、このテストはAUTOMATIC1111のバージョンやxformersの適用状態によっては微妙に違う画像が出力されることがあるらしい。

xformersを適用させている増田環境だと確かに二つ並べると間違い探しレベルの違いがあった。

正直このテストクリアしなくても十分だと個人的には思う。

おまけ2:その他便利になる設定や拡張機能

「Booru tag autocompletion for A1111」を導入すればNovelAIのように自動danbooruタグを保管してくれる。

注意

画像生成AIモデルはStable DiffusionOSSのため派生結構多い。

自前で追加学習もできるため自前で学習した追加AIモデル4chanのような掲示板などで共有する人もいるらしい。

しかしそのようなモデルの中にウィルスのような悪意のある動作を行うものもあるらしい。

FBIペドフィリア一網打尽にするためにIPアドレスなどの個人情報を抜き出す動作を行うロリ特化AIモデル掲示板で配布していて

しかもそれには本物の児童ポルノ教師データとして使われている…などという都市伝説的な話が今界隈を賑わせている。

それが本当の話かどうかはわからないが、とにかく変なところからモデルダウンロードするのは危険なのでやめよう。

自己矛盾溢れる注意喚起かもしれないが…

2022-09-27

https://anond.hatelabo.jp/20220914223600

日本人女性写真を生成するプロンプトについて私なりに試行錯誤した結果を示します。

画像は4つ以下に貼り付けました。

https://imgur.com/a/x8TOFla

約2週間経過しているため、既に何らかの結論を得られているかもしれませんが、どなたかの参考にでもなれば幸いです。

1.結論

(1)安定させるwordについて

actressを付けると安定すると書かれていましたが、正にその通りだと思います

actressだけでなく、2つの言葉を重ねると表情や顔の作りが安定するように思います

例として

actress×Japanese idol

AKB48 ×Japanese idol

AKB48 × a famous Japanese girl in Instagram

などです。

今回のpromptで、日本人を示す言葉はこの部分くらいだったのですがそれなりに日本人ぽくなっていると思います。(ちなみにAKB48とnogizaka46で素人目に大きな差は出ませんでした。)

また結果(1)と(2)に含まれているAya Kodaのように適当女性名前を入れると結果が安定する傾向も見られました。

(私は適当に、北条政子幸田文など、歴史上の女性名を使ってみましたが特にその名前の方々の特徴が取り入れられている訳ではなさそうです)

結果(3)は特定名前を使わない代わりに、女性らしさが出るかと思ってmuscular arm, cleavage of breasts, detailedなどを入れています

(2)CFG scale

CFG scaleは5~22で振ってみましたが、数値を大きくしすぎると光を強く当てたようなギラついた雰囲気になるようです。

5~7くらいが自然な気がしました。

(なお、自然光を当てた写真にするため、「on the beach」としましたので室内の設定だと異なると思います

(3)Step

STEPは100にしてもそれほど良い結果にならなかったため40~70にしています

2.条件と結果

使用しているツールは、下記のstable-diffusion-webui(by Automatic1111)です。

https://github.com/AUTOMATIC1111/stable-diffusion-webui

なお本稿では以降stable-diffusion-webuiと記載します。

「Restore faces」と「Highres,fix」はONにした状態とし、Samplerは「Euler a」を基本としています

また「Negative prompt」も使用しました。

stable-diffusion-webui(Automatic1111)の機能については下記をご参照ください。

https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features

美人なんだけどなんか複数人うつってる」みたいな変な写真がたまに出ること(これを解消できるワードが欲しい)

「Highres,fix」がその一助になると思いますのでもしお使いで無ければお試しください。

(1)grid-0015-18523803

◆prompt:

Photo of a young female wearing a T-shirt and jeans, Aya Koda, (((beauty bright eyes))), cute face, Angel face, short hair, small nose, on the beach, ((perfect face)), detailed, photorealistic, cute slight smile, Medium Shot, actress, Japanese idol,

Negative prompt:

ugly, fat, obese, chubby, (((deformed))), [blurry], bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), messy drawing, penis, nose, eyes, lips, eyelashes, text, red_eyes, (((nipples))),(skinny), (((underbite))), long jaw, long hair,
Steps: 70, Sampler: Euler a, CFG scale: 6.5, Seed: 18523803, Face restoration: GFPGAN, Size: 576x768, Denoising strength: 0.4
(2)grid-0009-1585211664

◆prompt:

Photo of a young female wearing a T-shirt and jeans, Aya Koda, (((beauty bright eyes))), cute face, Angel face, short hair, small nose, on the beach, ((perfect face)), detailed, photorealistic, cute slight smile, Medium Shot, actress, Japanese idol,

Negative prompt:

ugly, fat, obese, chubby, (((deformed))), [blurry], bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), messy drawing, penis, nose, eyes, lips, eyelashes, text, red_eyes, (((nipples))),(skinny), (((underbite))), long jaw, long hair,
Steps: 50, Sampler: Euler a, CFG scale: 16, Seed: 1585211664, Face restoration: GFPGAN, Size: 576x768, Denoising strength: 0.4
(3)grid-0001-3339383866

◆prompt:

portrait of an attractive talent wearing a T-shirt and jeans,  (((beauty bright eyes))), cute face with elegant hairstyle, face to the left, on the beach, ((perfect face)), muscular arm, cleavage of breasts, detailed, intricate, photorealistic, a famous Japanese girl in Instagram, cute slight smile, Medium Shot, akb48, Japanese idol,

Negative prompt:

ugly, fat, obese, chubby, (((deformed))), [blurry], bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), messy drawing, penis, nose, eyes, lips, eyelashes, text, red_eyes, (((nipples))),(skinny)
Steps: 55, Sampler: Euler a, CFG scale: 9, Seed: 3339383866, Face restoration: GFPGAN, Size: 576x768, Denoising strength: 0.25
(4)xy_grid-0401-1729711779

※これはsamplerとCFG Scale についての比較用です。

◆prompt:

Photo of a young female wearing a T-shirt and jeans, (((beauty bright eyes))), cute face, Angel face, short hair, small nose, on the beach, ((perfect face)), detailed, photorealistic, cute slight smile, Medium Shot, actress, Japanese idol,

Negative prompt:

ugly, fat, obese, chubby, (((deformed))), [blurry], bad anatomy, disfigured, poorly drawn face, mutation, mutated, (extra_limb), (ugly), (poorly drawn hands), messy drawing, penis, nose, eyes, lips, eyelashes, text, red_eyes, (((nipples))),(skinny), (((underbite))),
Steps: 70, Sampler: Euler a, CFG scale: 5.0, Seed: 1729711779, Face restoration: GFPGAN, Size: 512x768

3.補足・雑感

Negative promptは以下くらいに省略しても問題は無さそうです。

ugly, fat, obese, chubby, (((deformed))), bad anatomy, disfigured, poorly drawn face, mutation, mutated

ただしpromptもNegative promptは、あまりシンプルにし過ぎるとハズレ率が高くなるかもしれません。

またseedを固定してCFG Scaleを変えた方が効果が出るケースも多いような気がしています

結果(3)の「talent」の部分は、適当職業名に変えても成立しました。

例:swimmer、fireman、golfer

4.終わりに

面白い課題を与えていただきありがとうございました。

もう少し早く投稿たかったのですが、色々やっているうちにあっという間に時間が経ってしまいました。

また、日本人(というかモンゴロイド)に多い頬骨が張っていない丸顔を作るのは結構難しいと感じました。

以上

2022-08-28

[] 右脳左脳かい都市伝説

右脳左脳というのがそもそもないと思うの

人間右脳派・左脳派で分類をするのガーとか、クリエイティブ右脳左脳ガーとか、そう言うのは非科学的以前に、

今時は局在論ではなく全体論だと思うの。ついでに脳腸相関など単一でどうこうではないし、

割と脳奇形や無脳で生まれてくる子どもたちすらもいる

殆ど脳がないけど、公務員(ホワイトカラー)職に就いて、結婚子どももいて、44歳まで過ごしてきた男性もいるし、

脳がなく脳幹しかなかった少年自発呼吸をしつつ12歳まで家族と過ごしていたよ

Meet The Healthy, Functioning Man Who Survived With Almost No Brain. | IFLScience

https://www.iflscience.com/man-tiny-brain-lived-normal-life-31083 

 

 

When it comes to our brains, does size really matter? One of the biggest myths about the brain is that bigger is always better. But what about those who sit on the extreme end of that scale? How much of our brain do we actually need to survive? Looking through the archives of medical history, there are a number of people with tiny brains, or brains with huge chunks missing entirely, which defy all odds.

 

In a 2007 Lancet study, doctors described an incredible medical oddity – the 44-year-old civil servant who had lived a normal life despite having an incredibly tiny brain. The French man went into hospital after he experienced weakness in his left leg for two weeks. Doctors were quite surprised when they took scans of his brain and found a huge fluid-filled chamber.

 

The scans showed that the man had a “massive enlargement of the lateral, third, and fourth ventricles, a very thin cortical mantle and a posterior fossa cyst,” researchers noted in the study. In short, while fluid normally circulates throughout the brain, it’s regularly drained. But instead of draining the fluid into the circulatory system, the fluid in this man’s brain built up. Eventually, the accumulation of fluid resulted in only a tiny amount of actual brain material.

 

 

The man’s medical history showed that he had to get a shunt inserted into his head as an infant to get rid of the buildup of fluid on the brain, known as hydrocephalus. The shunt was eventually removed when at age 14, he complained of left leg weakness and some unsteadiness. The man went on to live a normal life and he got married and had two children. Tests showed that he had an IQ of 75 which, though below the average of 100, is not considered a mental disability.

 

“What I find amazing to this day is how the brain can deal with something which you think should not be compatible with life,” Dr. Max Muenke, from the National Human Genome Research Institute, told Reuters.

 

Earlier last year, IFLScience reported on the ninth known case of someone living without a cerebellum. This is the part of the brain that controls a number of important functions such as balance, motor movements and motor learning. The 24-year-old Chinese woman went into a hospital complaining of nausea and vertigo, and doctors discovered that she suffered from a rare condition known as cerebellar agenesis.

 

In another case, 12-year-old Trevor Judge Waltrip shocked medical professionals when he survived as long as he did with only his brain stem. Trevor passed away last year after going his entire life without a brain. He suffered from a rare condition called hydranencephaly, whereby the cerebral hemispheres are replaced entirely with cerebrospinal fluid. People with hydranencephaly usually survive for up to 12 weeks, which made Trevor’s case so remarkable. He was able to breathe on his own and respond to stimuli, but was blind and unable to communicate.

 

These cases show not only the adaptability and resilience of the human brain, but also how little we know about one of our most important organs. Cases like this force neuroscientists to rethink how we view the brain, particularly what functions different regions have and how the brain adapts when these regions become damaged.

2022-08-24

anond:20220824182342

これ優秀やね。元増田のやつよりお手軽。

あとせっかくだからサムネイルを並べて表示するコードを紹介しとこう。

import matplotlib.pyplot as plt

num_inference_steps = 10            # Number of denoising steps
guidance_scale = 7.5                # Scale for classifier-free guidance
batch_size = 1

def show_images(images, figsize=(20,10), columns = 5):
  plt.figure(figsize=figsize)
  for i, image in enumerate(images):
      plt.subplot(len(images) / columns + 1, columns, i + 1)
      plt.imshow(image)

#入力文字 ここに好きな禁則文字をいれてください
prompt = ["hatena anonymous diary"]

# 画像サイズ
height = 512                        # default height of Stable Diffusion
width = 512                         # default width of Stable Diffusion

# SEED値、ここをかえると 同じ入力文字でも別の画像がでます
seedId = 1

images = []
for i in range(5):
  generator = torch.manual_seed(seedId + i)
  print(seedId + i)
  image = run(prompt,generator,height,width,num_inference_steps,guidance_scale,batch_size)
  images.append(image)

show_images(images)

2022-07-26

If you could only eat one meal for the rest of your life, what would it be?

Who is your hero?

If you could live anywhere, where would it be?

What is your biggest fear?

What is your favorite family vacation?

What would you change about yourself if you could?

What really makes you angry?

What motivates you to work hard?

What is your biggest complaint about your job?

What is your favorite book to read?

What makes you laugh the most?

What was the last movie you went to? What did you think?

What did you want to be when you were small?

If you could choose to do anything for a day, what would it be?

What would you sing at Karaoke night?

How would your friends describe you?

What are your hobbies?

36. What is the best gift you have been given?

37. What is the worst gift you have received?

40. Where do you see yourself in five years?

42. If you were a super-hero, what powers would you have?

43. What would you do if you won the lottery?

44. What form of public transportation do you prefer? (air, boat, train, bus, car, etc.)

45. What's your favorite zoo animal?

46. If you could go back in time to change one thing, what would it be?

48. How many pillows do you sleep with?

49. What's the longest you've gone without sleep (and why)?

52. How often do you buy clothes?

53. Have you ever had a secret admirer?

54. What's your favorite holiday?

55. What's the most daring thing you've ever done?

56. What was the last thing you recorded on TV?

57. What was the last book you read?

58. What's your favorite type of foreign food?

59. Are you a clean or messy person?

60. Who would you want to play you in a movie of your life?

61. How long does it take you to get ready in the morning?

62. What kitchen appliance do you use every day?

63. What's your favorite fast food chain?

64. What's your favorite family recipe?

65. Do you love or hate rollercoasters?

66. What's your favorite family tradition?

67. What is your favorite childhood memory?

68. What's your favorite movie?

69. How old were you when you learned Santa wasn't real? How did you find out?

70. Is your glass half full or half empty?

71. What's the craziest thing you’ve done in the name of love?

72. What three items would you take with you on a deserted island?

73. What was your favorite subject in school?

74. What's the most unusual thing you've ever eaten?

75. Do you collect anything?

76. Is there anything you wished would come back into fashion?

77. Are you an introvert or an extrovert?

78. Which of the five senses would you say is your strongest?

79. Have you ever had a surprise party? (that was an actual surprise)

80. Are you related or distantly related to anyone famous?

81. What do you do to keep fit?

82. Does your family have a “motto” – spoken or unspoken?

83. If you were ruler of your own country what would be the first law you would introduce?

84. Who was your favorite teacher in school and why?

85. What three things do you think of the most each day?

86. If you had a warning label, what would yours say?

87. What song would you say best sums you up?

88. What celebrity would you like to meet at Starbucks for a cup of coffee?

89. Who was your first crush?

time clocks alarms appointments schedules tan sign up form team volunteers meeting black recruitment recruiting need sign up form 5k marathon running race marathon green shoes sign up form

90. What's the most interesting thing you can see out of your office or kitchen window?

91. On a scale of 1-10 how funny would you say you are?

92. Where do you see yourself in 10 years?

93. What was your first job?

94. If you could join any past or current music group which would you want to join?

95. How many languages do you speak?

96. What is your favorite family holiday tradition?

97. Who is the most intelligent person you know?

98. If you had to describe yourself as an animal, which one would it be?

99. What is one thing you will never do again?

100. Who knows you the best?

ログイン ユーザー登録
ようこそ ゲスト さん