[B! deep learning] motowotaã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

motowota id:motowota

deep learningã«é–¢ã™ã‚‹motowotaã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (5)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

97%ã®ç²¾åº¦ï¼šæ•™å¸«ãªã—å¦ç¿’ã§MNISTãƒ‡ãƒ¼ã‚¿ã‚’åˆ†é¡žã™ã‚‹æ‰‹æ³• - Qiita
ã¯ã˜ã‚ã« ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°ã¨ãƒãƒƒã‚·ãƒ¥ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ã®éƒ¨åˆ†ã‚’å®Œå…¨ã«ç†è§£ã™ã‚‹ã“ã¨ã¯ã§ãã¾ã›ã‚“ã§ã—ãŸã€‚æœ‰è˜è€…ã®æ–¹ãŒã„ã‚Œã°æ•™ãˆã¦æ¬²ã—ã„ãƒ»ãƒ»ãƒ» å¯¾è±¡èªè€… æ•™å¸«ãªã—å¦ç¿’ã§åˆ†é¡žã‚’è¡Œã„ãŸã„æ–¹ã€‚ä¸€èˆ¬çš„ãªæ©Ÿæ¢°å¦ç¿’ã®çŸ¥è˜ã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ãƒ¼ã‚µã‚¤ã‚¨ãƒ³ã‚¹ã®çŸ¥è˜ãŒã‚ã‚‹æ–¹ãŒå¯¾è±¡ã«ãªã‚Šã¾ã™ã€‚ èƒŒæ™¯ æ•™å¸«ãƒ‡ãƒ¼ã‚¿ãŒç”¨æ„ã§ããªã„ã‚¿ã‚¹ã‚¯ãŒã‚ã‚‹ã€‚ æ•™å¸«ãƒ‡ãƒ¼ã‚¿ãŒå°‘é‡ã—ã‹ç”¨æ„ã§ããªã„ã‚¿ã‚¹ã‚¯ãŒã‚ã‚‹ã€‚ æ•™å¸«ãªã—ãƒ‡ãƒ¼ã‚¿ã¯ã‚ã‚‹ã€‚ ä¸Šè¨˜ã®çŠ¶æ³ã§æœ‰åŠ¹ãªæ‰‹æ³•ã‚’æŽ¢ã™ãŸã‚ã«ä¸‹è¨˜ã®è«–æ–‡ã‚’èªã¿ã¾ã—ãŸã€‚ Hu, Weihua, et al. "Learning Discrete Representations via Information Maximizing Self Augmented Training." arXiv preprint arXiv:1702.08720 (2017). é¸ã‚“ã ç†ç”± ç²¾åº¦ãŒé«˜ã„ ã‚³ãƒ¼ãƒ‰ãŒå…¬é–‹ã•ã‚Œã¦ã„ã¦å†ç¾ã§ããã† ãƒ™ãƒ¼ã‚¹ã¨ãªã‚‹æ‰‹æ³•
motowota 2017/08/19
å¦ç¿’

Deep Learning
ãƒªãƒ³ã‚¯
LightNet: A Versatile, Standalone Matlab-based Environment for Deep Learning
LightNet is a lightweight, versatile and purely Matlab-based deep learning framework. The idea underlying its design is to provide an easy-to-understand, easy-to-use and efficient computational platform for deep learning research. The implemented framework supports major deep learning architectures such as Multilayer Perceptron Networks (MLP), Convolutional Neural Networks (CNN) and Recurrent Neur
motowota 2016/05/12
deep learning
ãƒªãƒ³ã‚¯
SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
View PDF Abstract:We present a novel and practical deep fully convolutional neural network architecture for semantic pixel-wise segmentation termed SegNet. This core trainable segmentation engine consists of an encoder network, a corresponding decoder network followed by a pixel-wise classification layer. The architecture of the encoder network is topologically identical to the 13 convolutional la
motowota 2016/05/12
deep learning
ãƒªãƒ³ã‚¯
Transport Analysis of Infinitely Deep Neural Network
We investigated the feature map inside deep neural networks (DNNs) by tracking the transport map. We are interested in the role of depth (why do DNNs perform better than shallow models?) and the interpretation of DNNs (what do intermediate layers do?) Despite the rapid development in their application, DNNs rem ain analytically unexplained because the hidden layers are nested and the parameters are
motowota 2016/05/12
deep learning
ãƒªãƒ³ã‚¯
NVIDIA Dockerã§ç°¡å˜ã«GPUå¯¾å¿œã®TensorFlowå…¥ã‚Šã‚³ãƒ³ãƒ†ãƒŠã‚’ä½œã‚‹æ–¹æ³•
ã“ã‚“ã«ã¡ã¯ã€‚ å°èª¬å®¶ã«ãªã‚ã†ã®ä½œå“ç¾¤ã‚’èªã‚“ã§ã„ãŸã‚‰é€£ä¼‘ãŒçµ‚ã‚ã‚Šãã†ã§ã™ã€‚ é€£ä¼‘ã®çµ‚ã‚ã‚Šã®ã‚¢ãƒ³ãƒŸãƒ©å¸°ã‚Šã«è¡æ’ƒä½œTensorFlowã§GPUãŒä½¿ãˆãªã„ã¨ãã®ã‚¢ãƒ³ã‚µãƒ¼ã‚¨ãƒ³ãƒˆãƒªã§ã‚ã‚‹TensorFlow on Dockerã§GPUã‚’ä½¿ãˆã‚‹ã‚ˆã†ã«ã™ã‚‹æ–¹æ³•ã‚’èªã¿ã€æ›¸ã„ã¦ã¿ãŸã‚¨ãƒ³ãƒˆãƒªã§ã™ã€‚å…ˆæœˆè³¼å…¥ã—ãŸæœ€ä½Žé™Maxwellçš„ãªã‚°ãƒ©ãƒ•ã‚£ãƒƒã‚¯ã‚«ãƒ¼ãƒ‰(å›³1)çµ¡ã¿ã§è¦‹ã¤ã‘ãŸNVIDIA Dockerã€ãªã‹ãªã‹è‰¯ã„ã‚‚ã®ãªã®ã§ã‚‚ã£ã¨å¤šãã®æ–¹ã«çŸ¥ã£ã¦ã‚‚ã‚‰ã„ãŸã„ãªã¨æ€ã„ã€‚ ä»Šå›žã¯NVIDIA Docker + TensorFlowã§GPUã‚’æœ‰åŠ¹æ´»ç”¨ã™ã‚‹æ‰‹é †ã‚’ç´¹ä»‹ã—ã¾ã™ã€‚ä»–ã®æ–¹ã«ã‚ˆã‚‹é–¢é€£è¨˜äº‹ã¨ã—ã¦ã€æ—¥æœ¬èªžã§ã®NVIDIA Docker + Caffeè§£èª¬ã¯ã™ã§ã«Ubuntu14.04.3ã§nvidia-dockerä½¿ã£ã¦Caffeã‚’ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«ã—ã¦ã¿ãŸãŒã‚ã‚Šã¾ã™ã€‚Caffeãªæ–¹ã¯ãã¡ã‚‰ã‚’å‚ç…§ãã ã•ã„*1ã€‚ä»Šå›žã¯ãŠé¡ŒãŒTen
motowota 2016/05/09
deep learning

GPU

NVIDIA
ãƒªãƒ³ã‚¯
1