CrypkoHigh-quality Anime Character Generation and Design powered by GAN (Generative Adversarial Networks).
AnimeGANv2 - a Hugging Face Space by akhaliq
Discover amazing ML apps made by the community
This Anime Does Not Exist
For significantly better and customizable anime image generation, check out Holara AI Creativity Slider       0.5 Higher creativity values tell the AI to be more creative and detailed, but also messy and weird Speed Slider    1.5 Space: pause grid, Drag: pan grid, Click: open image in new tab F: fullscreen mode, Z: toggle zoom on hover, V: video mode You can find updates about anime and AI on Twi
Making Anime Faces With StyleGAN · Gwern.net
A tutorial explaining how to train and generate high-quality anime faces with StyleGAN 1+2 neural networks, and tips/scripts for effective StyleGAN use. Generative neural networks, such as GANs, have struggled for years to generate decent-quality anime faces, despite their great success with photographic imagery such as real human faces. The task has now been effectively solved, for anime faces as
「生æˆDeep Learningã€ã¯ä¹…ã—ã¶ã‚Šã«ãƒ¯ã‚¯ãƒ¯ã‚¯ã™ã‚‹AI関係ã®æœ¬ã ã£ãŸ - karaage. [ã‹ã‚‰ã‚ã’]
「生æˆDeep Learningã€ã‚’èªã¿ã¾ã—㟠生æˆã«ç‰¹åŒ–ã—ãŸAIã®æœ¬ã€Œç”ŸæˆDeep Learningã€ã‚’èªã¿ã¾ã—ãŸã€‚ ç”Ÿæˆ Deep Learning ―絵をæãã€ç‰©èªžã‚„音楽を作りã€ã‚²ãƒ¼ãƒ をプレイã™ã‚‹ 作者:David Foster発売日: 2020/10/05メディア: å˜è¡Œæœ¬ï¼ˆã‚½ãƒ•ãƒˆã‚«ãƒãƒ¼ï¼‰ 独創力ã¨ã¯ã€æ€æ…®æ·±ã„模倣以外ã®ä½•ã‚‚ã®ã§ã‚‚ãªã„ ã¨ã¯18世紀フランスã®å“²å¦è€…ヴォルテールã®è¨€è‘‰ã¨ã®ã“ã¨ã§ã™ãŒã€AIã«ã‚ˆã‚‹æ¨¡å€£ã¨å‰µé€ ã®å¢ƒç•Œã«é–¢ã—ã¦è€ƒãˆã•ã›ã‚‰ã‚Œã‚‹ã‚ˆã†ãªæœ¬ã§ã—ãŸã€‚ 本ã®å†…容ã¯ã€ã‚ªãƒ¼ãƒˆã‚¨ãƒ³ã‚³ãƒ¼ãƒ€ã‹ã‚‰ã¯ã˜ã¾ã‚ŠGANã§ã®ç”»åƒç”Ÿæˆã€æ–‡ç« 生æˆã€éŸ³æ¥½ç”Ÿæˆã€‚アテンションã€å¼·åŒ–å¦ç¿’ã®ä¸–界モデルã¾ã§ã¨ã‚¿ã‚¤ãƒˆãƒ«é€šã‚Šç”Ÿæˆç³»ã®AI技術を幅広ãã‚«ãƒãƒ¼ã—ã¦ã„ã¾ã™ã€‚ çµæ§‹éª¨å¤ªã§ã€AIã®æœ¬ã¯1冊目ã¨ã„ã†äººãŒèªã‚€æœ¬ã§ã¯ãªãã¦ã€ä½•å†Šã‹AI関係ã®æ›¸ç±ã‚’èªã‚“ã§ã€ã‚る程度自分ã§å®Ÿè·µã—ã¦ã€ç”Ÿæˆç³»ã®AI（Auto Encoderã¨
![「生æˆDeep Learningã€ã¯ä¹…ã—ã¶ã‚Šã«ãƒ¯ã‚¯ãƒ¯ã‚¯ã™ã‚‹AI関係ã®æœ¬ã ã£ãŸ - karaage. [ã‹ã‚‰ã‚ã’]](https://cdn-ak-scissors.b.st-hatena.com/image/square/0e716f3414d0fb9918c8f114d6308be989b9ecd8/height=288;version=1;width=512/https%3A%2F%2Fcdn-ak.f.st-hatena.com%2Fimages%2Ffotolife%2Fk%2Fkaraage%2F20201030%2F20201030014348.jpg)
StyleGAN can generates realistic photos of Japanese Idols
Japanese Idol GeneratorJapanese idols are really cute and makes some people happier. So I’m developing a program to generate Japanese idols. I also developed face generator with face characteristic controllers (like more smiling, more pointy nose, more straight hair…etc). This is the video of how I controll the characteristics of idol faces. If you are interested in how this system is built, pleas
StyleGANã‚’éŠã³å°½ãã›!! ~è¿½åŠ å¦ç¿’ä¸è¦ã®ç”»åƒç·¨é›†~ - Qiita
ã¯ã˜ã‚ã« AdventCalenderè«–æ–‡24日目担当ã®ã±ã—ãµãƒã‚“ã§ã™ã€‚ çªç„¶ã§ã™ãŒ,最近ã“ã®ã‚ˆã†ãªè«–æ–‡1ãŒå‡ºã¾ã—ãŸã€‚ テスト https://t.co/QoXamqHB9w — ã±ã—ãµãƒã‚“ (@pacifinapacific) December 21, 2019 ãªã‚“ã¨ãŸã ã®ï¼‘枚絵をVtuberã®ãƒ¢ãƒ‡ãƒ«ã«ã§ãã¡ã‚ƒã†ã®ã§ã™ã€‚ニコニコã«ä¸ŠãŒã£ã¦ã„る解説動画ã§ã¯è³žè³›ã®ã‚³ãƒ¡ãƒ³ãƒˆãŒå¤šæ•°å¯„ã›ã‚‰ã‚Œã¦ã„ã¾ã—ãŸã€‚ ã“ã‚Œã¯ã™ã”ã„ï¼ã¨ã„ã†ã“ã¨ã§ç§ã‚‚èªã‚“ã ã®ã§ã™ãŒã€ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã‚’作る段階ã§ã™ã”ã„労力を費やã—ã¦ã„るよã†ã§ã—ãŸã€‚３ｄモデル1ã¤1ã¤ã‚’目を閉ã˜ãŸã‚Šã€é–‹ã‘ãŸã‚Šã€é¡”を傾ã‘ãŸã‚Šã¨å·®åˆ†ã‚’ã¨ã‚Šã€ãƒ©ãƒ™ãƒ«ä»˜ã‘ã—ã¦ã„ãã®ã¯ã¨ã¦ã‚‚大変ã§ã™ã€‚ 「ãã‚“ãªãƒ©ãƒ™ãƒ«ä»˜ã‘ã®æ‰‹é–“を失ãã—ã¦ä¼¼ãŸã‚ˆã†ãªã“ã¨ãŒã‚„ã‚ŠãŸã„ï¼ï¼ã€ ãã®ä¸€ã¤ã®å¯èƒ½æ€§ã¨ã—ã¦ä»Šå›žStyleGANã«ç€ç›®ã—ã¦ã¿ã¾ã™ã€‚StyleGANã¯æ»…茶苦茶綺麗ãªç”»åƒã‚’生æˆã§ãるモデルã§
深層生æˆãƒ¢ãƒ‡ãƒ«ã‚’å·¡ã‚‹æ—…(3): GAN - Qiita
ã¯ã˜ã‚㫠「深層生æˆãƒ¢ãƒ‡ãƒ«ã‚’å·¡ã‚‹æ—…ã€ã‚·ãƒªãƒ¼ã‚ºç¬¬3回ã¯ã¿ãªã•ã‚“ãŠå¾…ã¡ã‹ãã®ï¼ˆï¼Ÿï¼‰GANã®ã¾ã¨ã‚ã§ã™. GANã¯ç¶ºéº—ãªç”»åƒã‚’生æˆã™ã‚‹ã“ã¨ã«é•·ã‘ã¦ãŠã‚Š, ãã®äººæ°—ã¯Flowã‚„VAEã¨æ¯”ã¹ã¦ã‚‚圧倒的ã§ã™. ãã®ä¸€æ–¹ã§, 世ã«ã¯GANã®ç ”究ãŒã‚ãµã‚Œã¦ã„ã¦, ç”»åƒç”Ÿæˆã«é™ã£ã¦ã‚‚把æ¡ã™ã‚‹ã®ãŒå›°é›£ãªçŠ¶æ…‹ã«ãªã£ã¦ã„ã¾ã™. 本記事ã§ã¯, 元祖ã‹ã‚‰æœ€æ–°ã®ç ”究ã«è‡³ã‚‹æ´å²ã®ä¸ã§é‡è¦ã¨æ€ã‚れるもã®ã‚’ジャンル別ã«ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™. 今回も画åƒç”Ÿæˆã®ã¿ã‚’扱ã„ã¾ã™. GANã®åŸºæœ¬ GANãã®ã‚‚ã®ã«ã¤ã„ã¦ã®è§£èª¬ã¯æ—¥æœ¬èªžã®ã‚‚ã®ã«é™ã£ã¦ã‚‚æ—¢ã«å¤šæ•°ã‚ã‚‹ã®ã§, ã“ã“ã§ã¯ç°¡å˜ã«è§¦ã‚Œã‚‹ã“ã¨ã¨ã—, 後ã«ç¶šãå„手法ã®ç´¹ä»‹ã«é›†ä¸ã—ãŸã„ã¨æ€ã„ã¾ã™. ã”å˜ã˜ã®æ–¹ã¯é£›ã°ã—ã¦ãã ã•ã„. å…¨ä½“åƒ A Beginner's Guide to Generative Adversarial Networks (GANs) | Skymind GANã¯ç”Ÿæˆ
AIãŒ100万種類以上をæã。世界最大ã®é¡”イラストæ供サービス「彩ã¡ã‚ƒã‚“（SAI）ã€ã‚’9月10日よりæ供開始
æ ªå¼ä¼šç¤¾ãƒ©ãƒ‡ã‚£ã‚¦ã‚¹ãƒ»ãƒ•ã‚¡ã‚¤ãƒ–（本社：æ±äº¬éƒ½æ–°å®¿åŒºã€ä»£è¡¨å–ç· å½¹ï¼šæ¼†åŽŸå¤§ä»‹ã€ä»¥ä¸‹ RADIUS5）ã¯ã€100万種類以上ã®å¤šç¨®å¤šæ§˜ãªé¡”イラストをAIã«æã„ã¦ã‚‚らãˆã‚‹æ–°ã‚µãƒ¼ãƒ“ス「彩ã¡ã‚ƒã‚“（SAI）ã€ã‚’cre8tiveAI ã§æ供開始ã—ã¾ã—ãŸã€‚ イラスト生æˆAI，彩ã¡ã‚ƒã‚“ã®ãƒˆãƒƒãƒ—イメージ 開発ã®èƒŒæ™¯ Deep learning 技術ã®é€²å±•ã«ã‚ˆã£ã¦ã€æ§˜ã€…ãªæ¥ç•Œã«æŠ€è¡“é©æ–°ãŒèµ·ã“ã£ã¦ã„ã¾ã™ã€‚ã“ã‚Œã¾ã§äººã®æ‰‹ã§ã¯è§£æ±ºã«è†¨å¤§ãªæ™‚é–“ãŒã‹ã‹ã£ã¦ã„ãŸä½œæ¥ãŒã€ä¸€çž¬ã§è§£æ±ºã§ãã¦ã—ã¾ã†é©æ–°çš„ãªã‚µãƒ¼ãƒ“ã‚¹ã‚‚ç™»å ´ã™ã‚‹ã‚ˆã†ã«ãªã£ã¦ãã¾ã—ãŸã€‚ ã—ã‹ã—ã€æŠ€è¡“é©æ–°ã®æ©æµã¯ã‚¢ãƒ‹ãƒ¡åˆ¶ä½œã‚„ゲーム制作をã¯ã˜ã‚ã¨ã—ãŸã‚¯ãƒªã‚¨ã‚¤ãƒ†ã‚£ãƒ–制作æ¥å‹™ã‚’è¡Œã†äººé”ã«ã¯ã€ã¾ã ã¾ã 届ã„ã¦ã„ãªã„状æ³ã§ã™ã€‚ RADIUS5ã§ã¯ã€Deep learning 技術を用ã„ãŸç ”究開発を行ã„ã€ã‚¯ãƒªã‚¨ã‚¤ãƒ†ã‚£ãƒ–ã®ç¾å ´ã§åˆ©ç”¨ã™ã‚‹ã“ã¨ãŒã§ãã‚‹AIサービスã®é–‹ç™ºã«å–り組んã§å‚ã‚Šã¾ã—ãŸã€‚20
StyleGAN「写真ãŒè¨¼æ‹ ã«ãªã‚‹æ™‚代ã¯çµ‚ã‚ã£ãŸã€‚〠- Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? ・å„convolution層後ã«styleã®èª¿æ•´ã‚’行ㆠ・細部ã®ç‰¹å¾´ï¼ˆé«ªè³ªã‚„ãã°ã‹ã™ï¼‰ã¯ãƒŽã‚¤ã‚ºã«ã‚ˆã£ã¦ç”Ÿæˆã•ã‚Œã‚‹ ・潜在変数$z$ã‚’ä¸é–“潜在変数$w$ã«ãƒžãƒƒãƒ”ングã™ã‚‹ ・ã“ã‚Œã¾ã§ã®GANã®ã‚ˆã†ã«Generatorã®å…¥åŠ›å±¤ã«æ½œåœ¨å¤‰æ•°$z$を入れるã“ã¨ã¯ã—ãªã„ Style-based generator A：$w$ã‚’style($y_s,y_b$)ã«å¤‰ãˆã‚‹ãŸã‚ã®ã‚¢ãƒ•ã‚£ãƒ³å¤‰æ› $y_s,y_b$ã¯ãƒãƒ£ãƒ³ãƒãƒ«ã”ã¨ã«å€¤ã‚’も㤠B：ノイズã¯1ãƒãƒ£ãƒ³ãƒãƒ«ç”»åƒã‹ã‚‰æˆã‚‹ convã®å‡ºåŠ›ã«è¶³ã—åˆã‚ã›ã‚‹å‰ã«ã€ãƒŽã‚¤ã‚ºã‚’ãƒãƒ£ãƒ³ãƒãƒ«ã”ã¨ã«ã‚¹ã‚±ãƒ¼ãƒªãƒ³ã‚°ã™ã‚‹ã“ã¨ã‚’æ„味
A Style-Based Generator Architecture for Generative Adversarial Networks - YouTube
Paper (PDF): http://stylegan.xyz/paper Authors: Tero Karras (NVIDIA) Samuli Laine (NVIDIA) Timo Aila (NVIDIA) Abstract: We propose an alternative generator architecture for generative adversarial networks, borrowing from style transfer literature. The new architecture leads to an automatically learned, unsupervised separation of high-level attributes (e.g., pose and identity when trained on huma
写真ã«å†™ã£ã¦ã„ãªã„ã¨ã“ã‚を復元ã™ã‚‹ - エクサウィザーズ Engineer Blog
ã“ã‚“ã«ã¡ã¯ã€‚エクサウィザーズAIエンジニアã®é ˆè—¤ã§ã™ã€‚ ã¿ãªã•ã‚“ã¯ãƒã‚¤ã‚ングã®å†™çœŸã§ã—ãšã¡ã‚ƒã‚“ã°ã‹ã‚Šå†™ã—ã¦ã€ã¾ã¨ã‚‚ã«æ’®ã‚‰ã‚Œãªã‹ã£ãŸã‚¸ãƒ£ã‚¤ã‚¢ãƒ³ã«æ®´ã‚‰ã‚Œã‹ã‘ãŸã“ã¨ã¯ç„¡ã„ã§ã—ょã†ã‹ã€‚ ãã‚“ãªã¨ã「万能プリンターã€ãŒã‚ã£ãŸã‚‰ä¾¿åˆ©ã§ã™ã。もã†æ’®ã£ã¦ã—ã¾ã£ãŸå†™çœŸã®ã€å‘ãやズームを後ã‹ã‚‰ä¿®æ£ã—ã¦ã€å†™ã£ã¦ãªã‹ã£ãŸã¨ã“ã‚を復元ã—ã¦å†ãƒ—リントã§ãã‚‹ã¨ã„ã†ã‚‚ã®ã§ã™ã€‚ ã—ã‹ã—æŒã¡ä¸»ã§ã‚るドラãˆã‚‚ã‚“ã¯ã€ã†ã¡ã«ã‚‚ã¾ã æ¥ã¦ã„ã¾ã›ã‚“。仕方ãŒãªã„ã®ã§AIã®åŠ›ã§ãªã‚“ã¨ã‹ã—ã¾ã—ょã†ã€‚ 目的 写真ã®å¤–å´ã«å†™ã£ã¦ã„ã‚‹ã‚‚ã®ã‚’推測ã—ã€è‡ªç„¶ãªå½¢ã§åˆæˆã—ã¾ã™ã€‚ 物体ã®éƒ¨åˆ†ç”»åƒã‹ã‚‰ãã®ç¨®é¡žãªã„ã—ä½ç½®ã‚’推測ã—ã€æ—¢å˜ç”»åƒã‚’本ã«å…¨ä½“を復元ã™ã‚‹ã“ã¨ãŒã€åŽŸç†çš„ã«ã¯å¯èƒ½ãªã¯ãšã§ã™ã€‚ ãã®éŽç¨‹ã‚’直接ã«ãƒ—ãƒã‚°ãƒ©ãƒ ã™ã‚‹ã“ã¨ã¯ç¾å®Ÿçš„ã§ã¯ã‚ã‚Šã¾ã›ã‚“。 代ã‚ã‚Šã«ç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯(CNN)ã«ã€ã‹ã„ã¤ã¾ã‚“ã§å¦ç¿’ã•ã›ã¾ã™ã€‚ å¦ç¿’モデル 敵対的生æˆãƒãƒƒãƒˆãƒ¯ãƒ¼
Generative Adversarial Networks (GAN) @ NIPS2017
NIPS2017本会è°ã§æŽ¡æŠžã•ã‚ŒãŸ Generative Adversarial Networks (GAN) 論文をã¾ã¨ã‚紹介ã—ã¦ã„ã¾ã™ã€‚å¦ç¿’ã®åŽæŸæ€§ãƒ»å®‰å®šæ€§ã€åŠæ•™å¸«ã‚ã‚Šå¦ç¿’ã€Mode Collapse回é¿ã€è§£ãã»ãã•ã‚ŒãŸè¡¨ç¾å¦ç¿’ã€æ§‹é€ çš„ãªç”Ÿæˆã€ç‰ã€‚NIPS2017èªã¿ä¼š@PFNã§ã®ç™ºè¡¨è³‡æ–™ã§ã™ã€‚Read less
Meow Generator
I experimented with generating faces of cats using Generative adversarial networks (GAN). I wanted to try DCGAN, WGAN and WGAN-GP in low and higher resolutions. I used the CAT dataset (yes this is a real thing!) for my training sample. This dataset has 10k pictures of cats. I centered the images on the kitty faces and I removed outliers (I did this from visual inspection, it took a couple of hours
Deep Learningã§ã€Œã„らã™ã¨ã‚„風人間画åƒç”Ÿæˆãƒ¢ãƒ‡ãƒ«ã€ã‚’作ã£ãŸè©±ï¼ˆDCGANã€Wasserstein GAN） - ã¬ã„ãã‚‹ã¿ãƒ©ã‚¤ãƒ•ï¼Ÿ
ãƒ•ãƒªãƒ¼ç´ æサイト「ã„らã™ã¨ã‚„ã€ã«å‡ºã¦ãる人間風ã®ç”»åƒã‚’自動生æˆã™ã‚‹ãƒ¢ãƒ‡ãƒ«ã‚’Deep Learningã§ä½œã‚Šã¾ã—ãŸã€‚実装ã«ã¯Google製ã®ãƒ©ã‚¤ãƒ–ラリ「TensorFlowã€ã¨æ©Ÿæ¢°å¦ç¿’アルゴリズムã®ã€ŒDCGANã€ã€ŒWasserstein GANã€ã‚’用ã„ã¦ã„ã¾ã™ã€‚ 以下ã¯ç”Ÿæˆã•ã‚ŒãŸäººé–“ç”»åƒã®ã†ã¡ãã‚Œãªã‚Šã«ãã‚Œã„ãªã‚‚ã®ã®ä¸€ä¾‹ã§ã™ã€‚é ¬ã®ã¨ã“ã‚ãŒèµ¤ããªã£ã¦ã„ã¦ä½•ã¨ãªã本家ã„らã™ã¨ã‚„ã®ç‰¹å¾´ã‚’æ‰ãˆã‚‰ã‚Œã¦ã„ã‚‹ã¨æ€ã„ã¾ã™ã€‚ 「ã„らã™ã¨ã‚„ã€ã¨ã¯ï¼Ÿ 実装ã—ãŸæ‰‹æ³•ã®æ¦‚è¦ DCGANã€Wasserstein GANã«ã¤ã„㦠Generator Discriminator Generatorã¨Discriminatorã®å¦ç¿’ å¦ç¿’や実装ã®è©³ç´° Generatorã€Discriminatorã®ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯æ§‹æˆã‚„パラメーター 訓練データ ãã®ä»– å¦ç¿’çµŒéŽ ãƒ¢ãƒ‡ãƒ«ã‚’æ¤œè¨¼ã™ã‚‹ 入力ã«ãƒã‚¤ã‚¢ã‚¹ã‚’掛ã‘ã¦ã„ã„ç”»åƒã‚’出やã™ãã™ã‚‹ ã¾
ランã‚ング
ランã‚ング
ランã‚ング
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
Artbreeder
CLOVA OCRã¯ãªãœã“ã‚“ãªã«èªè˜ç²¾åº¦ãŒé«˜ã„ã®ã‹ï¼ŸAIã§ã€Œæ‰‹æ›¸ãフォントã€ã‚’作ã£ãŸLINE CLOVAã®é–‹ç™ºç§˜è©±
Generating custom photo-realistic faces using AI – Insight Data
TFGAN: A Lightweight Library for Generative Adversarial Networks
MakeGirlsMoe
{{#tags}}- {{label}}
{{/tags}}