PyTorch (12) Generative Adversarial Networks (MNIST) - 人工知能ã«é–¢ã™ã‚‹æ–創録å‰å›žï¼ˆ2018/2/28）ã®æœ€å¾Œã§æ¬¡ã¯Conditional VAEã ã¨è¨€ã£ã¦ã„ãŸã‘ã©æ€ã„ã£ãり無視ã—㦠(^^;) 今回ã¯Generative Adversarial Networks (GAN) ã‚„ã‚ã†ã€‚ã„ãã¤ã‹ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã§å®Ÿé¨“ã—よã†ã¨æ€ã£ã¦ã„ã‚‹ã‘ã©ä»Šå›žã¯æœ€åˆã¨ã„ã†ã“ã¨ã§MNISTã‹ã‚‰ã€‚ 今回ã®å®Ÿè£…ã¯æ£ç¢ºã«è¨€ã†ã¨Generatorã¨Discriminatorã«ç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã‚’使ã£ã¦ã„ã‚‹ã®ã§ DCGAN（Deep Convolutional Generative Adversarial Networks） ã¨å‘¼ã°ã‚Œã‚‹GANã«ã‚ãŸã‚‹ã€‚è«–æ–‡ã®è¨å®šã¨ã¯å¾®å¦™ã«é•ã†ã¨ã“ã‚ã‚ã‚‹ã‘ã©ã€‚ 180303-gan-mnist.ipynb - Google ドライブ ã¾ãšã¯ã„ã¤ã‚‚ã®importã‹ã‚‰ã€‚ import os import pickle import numpy as np impor
[最新論文/GAN] ラベルãŒãªã„ãªã‚‰æŽ¨æ¸¬ã™ã‚Œã°ã„ã„。 - Qiita![[最新論文/GAN] ラベルãŒãªã„ãªã‚‰æŽ¨æ¸¬ã™ã‚Œã°ã„ã„。 - Qiita](https://cdn-ak-scissors.b.st-hatena.com/image/square/89792d24bfcbaed2f7f5f8a9a554f92748054a07/height=288;version=1;width=512/https%3A%2F%2Fqiita-user-contents.imgix.net%2Fhttps%253A%252F%252Fcdn.qiita.com%252Fassets%252Fpublic%252Farticle-ogp-background-412672c5f0600ab9a64263b751f1bc81.png%3Fixlib%3Drb-4.0.0%26w%3D1200%26mark64%3DaHR0cHM6Ly9xaWl0YS11c2VyLWNvbnRlbnRzLmltZ2l4Lm5ldC9-dGV4dD9peGxpYj1yYi00LjAuMCZ3PTk3MiZoPTM3OCZ0eHQ9JTVCJUU2JTlDJTgwJUU2JTk2JUIwJUU4JUFCJTk2JUU2JTk2JTg3JTJGR0FOJTVEJTIwJUUzJTgzJUE5JUUzJTgzJTk5JUUzJTgzJUFCJUUzJTgxJThDJUUzJTgxJUFBJUUzJTgxJTg0JUUzJTgxJUFBJUUzJTgyJTg5JUU2JThFJUE4JUU2JUI4JUFDJUUzJTgxJTk5JUUzJTgyJThDJUUzJTgxJUIwJUUzJTgxJTg0JUUzJTgxJTg0JUUzJTgwJTgyJnR4dC1hbGlnbj1sZWZ0JTJDdG9wJnR4dC1jb2xvcj0lMjMxRTIxMjEmdHh0LWZvbnQ9SGlyYWdpbm8lMjBTYW5zJTIwVzYmdHh0LXNpemU9NTYmcz00N2ZkZDNiNmQ5YjA0OGRkMGE4YjdmOWRiMmQ4NTFkNQ%26mark-x%3D142%26mark-y%3D57%26blend64%3DaHR0cHM6Ly9xaWl0YS11c2VyLWNvbnRlbnRzLmltZ2l4Lm5ldC9-dGV4dD9peGxpYj1yYi00LjAuMCZoPTc2Jnc9NzcwJnR4dD0lNDBQaG9lYm9vb28mdHh0LWNvbG9yPSUyMzFFMjEyMSZ0eHQtZm9udD1IaXJhZ2lubyUyMFNhbnMlMjBXNiZ0eHQtc2l6ZT0zNiZ0eHQtYWxpZ249bGVmdCUyQ3RvcCZzPWY5MTEwNzA0OWVjNTk4ZWYyY2FkMTY5MzI1NzgzNzY5%26blend-x%3D142%26blend-y%3D486%26blend-mode%3Dnormal%26s%3D911774cdadc5ddf53bf781acc1d8e6fd)
奈良先端科å¦æŠ€è¡“大å¦é™¢å¤§å¦ãªã©ã€ãã†ã‚んをラーメンã«éŒ¯è¦šã•ã›ã‚‹ARã¨GANを組ã¿åˆã‚ã›ãŸãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ 味覚æ“作システムを発表。白ã”飯ãŒç„¼é£¯ã«ã‚‚
Jesse Engel on Twitter: "Make music with GANs! GANSynth is a new method for fast generation of high-fidelity audio. 🎵 Examples:… https://t.co/MwxjZTO3g9"
examples/dcgan at main · pytorch/examples
SRGANã‚’pytorchã§å®Ÿè£…ã—ã¦ã¿ãŸ - Qiita
keras GAN 自å‰ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã§ç”»åƒç”Ÿæˆ - Qiita
Reddit - Dive into anything
GitHub - akanimax/pro_gan_pytorch-examples: Examples trained using the python pytorch package pro-gan-pth
GitHub - akanimax/pro_gan_pytorch: Unofficial PyTorch implementation of the paper titled "Progressive growing of GANs for improved Quality, Stability, and Variation"
GitHub - daigo0927/WGAN_GP: Keras model and tensorflow optimization of 'improved Training of Wasserstein GANs'
https://githubja.com/eriklindernoren/keras-gan
Kerasã®ãƒ¢ãƒ‡ãƒ«ã¨TensorFlowã®æœ€é©åŒ–ã§Wasserstein GANã‚’å¦ç¿’ã™ã‚‹ - Qiita
Ryobot on Twitter: "StackGANã‚„PGGANã¯å¤šæ®µéšŽã®Dã§é«˜ç”»è³ªåŒ–ã™ã‚‹ãŒå¦ç¿’ã¨ãƒ¢ãƒ‡ãƒ«ãŒè¤‡é›‘ã«ãªã‚‹ï¼Ž IntroVAEã¯ä¸€æ®µéšŽã§1024x1024ç”»åƒã‚’生æˆã§ãる．Dã‚‚ä¸è¦ï¼Ž VAEをベースã«GANã‚’Z, Zr, Zpé–“ã®KLæƒ…å ±é‡ã®æœ€å°åŒ–/最大化… https://t.co/c3OJRr6UfE"
Python + tensorflow ã§GANを試ã—ã¦ã¿ãŸï¼Ž
GANã«ã‚ˆã‚‹è¶…解åƒãŒã‚‚ãŸã‚‰ã™æ–°ã—ã„æ–¹å‘性ã¨ã€ŒThe Perception-Distortion Tradeoffã€ã®è©± - Think more, try less
Conditional DCGANã§ç”»åƒç”Ÿæˆ - kumilog.net
GANã«ã¤ã„ã¦æ¦‚念ã‹ã‚‰å®Ÿè£…ã¾ã§ã€€ï½žDCGANã«ã‚ˆã‚‹ã‚ルミーベイベー生æˆï½ž - Qiita
今ã•ã‚‰èžã‘ãªã„GAN（1ï¼‰ã€€åŸºæœ¬æ§‹é€ ã®ç†è§£ - Qiita
ai-topics.com
{{#tags}}- {{label}}
{{/tags}}