CS231n Convolutional Neural Networks for Visual Recognition
Table of Contents: Gradient checks Sanity checks Babysitting the learning process Loss function Train/val accuracy Weights:Updates ratio Activation/Gradient distributions per layer Visualization Parameter updates First-order (SGD), momentum, Nesterov momentum Annealing the learning rate Second-order methods Per-parameter adaptive learning rates (Adagrad, RMSProp) Hyperparameter Optimization Evalua
Neural Style Transfer: Prismaã®èƒŒæ™¯æŠ€è¡“を解説ã™ã‚‹
写真をピカソやゴッホã®ã‚ˆã†ãªã‚¹ã‚¿ã‚¤ãƒ«ã«å¤‰æ›ã§ãるアプリPrismaãŒè©±é¡Œã«ãªã‚Šã¾ã—ãŸã€‚多ãã®äººã¯ã€ãƒ‡ã‚£ãƒ¼ãƒ—ラーニングãŒä½¿ã‚ã‚Œã¦ã„ã‚‹ã‹ã©ã†ã‹ã¨ã¯é–¢ä¿‚ãªãã€ç´”粋ã«ã‚¢ãƒ—リを楽ã—ã‚“ã§ã„ã‚‹ã®ã ã¨æ€ã„ã¾ã™ã€‚ ã“ã®ã‚ˆã†ã«ãƒ‡ã‚£ãƒ¼ãƒ—ラーニングを使ã£ãŸäººæ°—アプリãŒå‡ºã¦ãã‚‹ã¨ã„ã†ã“ã¨ã¯éžå¸¸ã«è‰¯ã„ã“ã¨ã§ã¯ãªã„ã‹ã¨æ€ã„ã¾ã™ã€‚今回ã¯ã€Prismaã®èƒŒæ™¯æŠ€è¡“（ã¨æ€ã‚れるもã®ï¼‰ã‚’解説ã—ã¦ã„ãã¾ã™ã€‚ 目次 基礎ç†è«– 実装 改善 高速化 ã¾ã¨ã‚ 基礎ç†è«– ディープラーニングを使ã£ãŸã‚¢ãƒ¼ãƒˆç³»ã®è«–æ–‡ã¯è‰²ã€…ã¨å‡ºã¦ã„ã¾ã™ãŒã€ä¸€ç•ªåŸºç¤Žã¨ãªã‚‹è«–æ–‡ã¯Gatys et al. 2016ã§ã¯ãªã„ã‹ã¨æ€ã„ã¾ã™ã€‚プレプリント版ã¯2015å¹´8月ã«å‡ºã¦ã„ã¾ã™ã€‚ ã“ã®è«–æ–‡ã¯è¨˜äº‹ã¨ã—ã¦å–り上ã’られã¦è©±é¡Œã«ãªã£ã¦ã„ãŸã‚Šã‚‚ã—ãŸã®ã§ã€çŸ¥ã£ã¦ã„る人も多ã„ã®ã§ã¯ãªã„ã‹ã¨æ€ã„ã¾ã™ã€‚ã“ã®ç« ã§ã¯ã€ã‚¹ã‚¿ã‚¤ãƒ«å¤‰æ›ã®åŸºç¤Žã¨ãªã‚‹ã“ã®è«–文を解説ã—ã¦ã„ãã¾ã™ã€‚ Gatys et a
ã¯ã˜ã‚ã¦ã®GAN
今回ã¯GAN（Generative Adversarial Network）を解説ã—ã¦ã„ãã¾ã™ã€‚ GANã¯â€œDeep Learningâ€ã¨ã„ã†æœ¬ã®è‘—者ã§ã‚‚ã‚ã‚‹Ian GoodfellowãŒè€ƒæ¡ˆã—ãŸãƒ¢ãƒ‡ãƒ«ã§ã™ã€‚NIPS 2016ã§ã‚‚GANã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ãŒè¡Œã‚れるãªã©éžå¸¸ã«æ³¨ç›®ã‚’集ã‚ã¦ã„る分野ã§ã€æ¬¡ã€…ã«è«–æ–‡ãŒå‡ºã¦ãã¦ã„ã¾ã™ã€‚ ã¾ãŸã€Quoraã®ã‚»ãƒƒã‚·ãƒ§ãƒ³ã§Yann LeCunãŒã€ã“ã®10å¹´ã®æ©Ÿæ¢°å¦ç¿’ã§æœ€ã‚‚é¢ç™½ã„アイディアã¨è¿°ã¹ã¦ã„ãŸã‚Šã‚‚ã—ã¾ã™ã€‚ “The most interesting idea in the last 10 years in ML, in my opinion.†–Yann LeCun GANã¯èžã„ãŸã“ã¨ã¯ã‚ã‚‹ã‘ã‚Œã©ã‚ã¾ã‚Šè¿½ãˆã¦ãªã„ã¨ã„ã†äººå‘ã‘ã«åŸºç¤Žã‹ã‚‰è§£èª¬ã—ã¦ã„ããŸã„ã¨æ€ã„ã¾ã™ã€‚ãã‚Œã§ã¯é †ã«è¦‹ã¦ã„ãã¾ã—ょã†ã€‚ 目次 基礎ç†è«– DCGAN 実装 論文紹介 ã¾ã¨ã‚ 基礎ç†
GitHub - alexjc/neural-doodle: Turn your two-bit doodles into fine artworks with deep neural networks, generate seamless textures from photos, transfer style from one image to another, perform example-based upscaling, but wait... there's more! (An impleme
Use a deep neural network to borrow the skills of real artists and turn your two-bit doodles into masterpieces! This project is an implementation of Semantic Style Transfer (Champandard, 2016), based on the Neural Patches algorithm (Li, 2016). Read more about the motivation in this in-depth article and watch this workflow video for inspiration. The doodle.py script generates a new image by using o
1
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
{{#tags}}- {{label}}
{{/tags}}