[B! deeplearning] rin1024ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

rin1024 id:rin1024

é–¢é€£ã‚¿ã‚°ã§çµžã‚Šè¾¼ã‚€ (6)

ã‚¿ã‚°ã®çµžã‚Šè¾¼ã¿ã‚’è§£é™¤

deeplearningã«é–¢ã™ã‚‹rin1024ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (10)

${{author_name}}$

{{{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}}

æ‚ªå¤¢ç”»åƒã®è¡æ’ƒï¼ Googleã®äººå·¥çŸ¥èƒ½ã€ŒDeep Dreamã€ãŒè¦‹ã‚‹å¤¢ãŒãƒ¤ãƒã™ãŽ - ã¾ãã¾ããƒ‹ãƒ¥ãƒ¼ã‚¹ï¼
GoogleãŒé–‹ç™ºã—ãŸäººå·¥çŸ¥èƒ½ãŒè¦‹ã‚‹å¤¢ã€ŒDeep Dreamã€ãŒä¸–ç•Œä¸ã§è©±é¡Œã¨ãªã£ã¦ã„ã¾ã™ã€‚AIã«ã‚ˆã£ã¦äººå·¥çš„ã«ç”Ÿæˆã•ã‚ŒãŸç”»åƒã¨ã„ã†ã®ãŒã€ã¾ã•ã«æ‚ªå¤¢ãã®ã‚‚ã®ã€‚ã€Žã‚¯ãƒªã‚¨ã‚¤ã‚¿ãƒ¼ã¸ã€æ—¥åˆŠãƒ‡ã‚¸ã‚¿ãƒ«ã‚¯ãƒªã‚¨ã‚¤ã‚¿ãƒ¼ã‚ºã€‘ã€ã®ã‚¤ãƒ©ã‚¹ãƒˆãƒ¬ãƒ¼ã‚¿ãƒ¼ãƒ»å‰äº•å®ã•ã‚“ãŒã€å®Ÿéš›ã«ã€ŒDeep Dreamã€ã‚’ä½¿ç”¨ã—ãŸãƒ¬ãƒãƒ¼ãƒˆã‚’å ±å‘Šã—ã¦ã„ã¾ã™ã€‚â€¦â€¦äººã«ã‚ˆã£ã¦ã¯â€œé–²è¦§æ³¨æ„â€ãªç”»åƒã‚‚ã‚ã‚‹ã‹ã‚‚ã€‚ å…ˆæœˆã‹ã‚‰è©±é¡Œã«ãªã£ã¦ã„ã‚‹ã€GoogleãŒé–‹ç™ºã—ãŸäººå·¥çŸ¥èƒ½ãŒè¦‹ã‚‹æ‚ªå¤¢ã€ŒDeep Dreamã€ã€‚ãã‚ã—ã„è§£èª¬ã¯ãƒªãƒ³ã‚¯ã‚’èªã‚“ã§ã„ãŸã ãã¨ã—ã¦ã€ç°¡å˜ã«è¨€ã†ã¨ã€Œäººå·¥çŸ¥èƒ½ã«ç”»åƒèªè˜ã•ã›ã€ä»Šã¾ã§å¦ç¿’ã—ãŸè†¨å¤§ãªç”»åƒã‚’å‚è€ƒã«ã€ä½•ãŒè¦‹ãˆãŸã‹æã‹ã›ã‚‹ã‚‚ã®ï¼ˆãŸã¶ã‚“wï¼‰ã€ã€‚ æ™®é€šã¯å†™çœŸã®ä¸æ˜Žçžãªéƒ¨åˆ†ã¯ç”»åƒèªè˜ã§ããªã„ã‚“ã ã‘ã©ã€ã€Œä½•ã§ã‚‚ã„ã„ã‹ã‚‰ä½•ã«è¦‹ãˆãŸã‹æã‘ã£ï¼ã€ã£ã¦æ„Ÿã˜ã§ãƒ ãƒªãƒ¤ãƒªèªè˜ã•ã›ã‚‹æ„Ÿã˜ã‹ãªï¼Ÿã€€ä¸æ˜Žçžãªå†™çœŸã©ã“ã‚ã‹ã€ç ‚ã®åµã®ã‚ˆã†ãªãƒŽã‚¤ã‚ºç”»åƒã§ã•ãˆã€Œä½•ãŒè¦‹ãˆãŸï¼Ÿ
rin1024 2015/12/25
deeplearning
ãƒªãƒ³ã‚¯
ConvNetJS: Deep Learning in your browser
Description The library allows you to formulate and solve Neural Networks in Javascript, and was originally written by @karpathy (I am a PhD student at Stanford). However, the library has since been extended by contributions from the community and more are warmly welcome. Current support includes: Common Neural Network modules (fully connected layers, non-linearities) Classification (SVM/Softmax)
rin1024 2015/12/24
deeplearning
ãƒªãƒ³ã‚¯
Just a moment...
rin1024 2015/12/24
deeplearning
ãƒªãƒ³ã‚¯
ConvNetJS: Deep Learning in your browser
Description The library allows you to formulate and solve Neural Networks in Javascript, and was originally written by @karpathy (I am a PhD student at Stanford). However, the library has since been extended by contributions from the community and more are warmly welcome. Current support includes: Common Neural Network modules (fully connected layers, non-linearities) Classification (SVM/Softmax)
rin1024 2015/12/24
research

deeplearning
ãƒªãƒ³ã‚¯
Macã§TensorFlowã®TypeError: __init__() got an unexpected keyword argument 'syntax'ã‚’è§£æ±ºã™ã‚‹ - Qiita
>>> import tensorflow as tf Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/lib/python2.7/site-packages/tensorflow/__init__.py", line 4, in <module> from tensorflow.python import * File "/usr/local/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 13, in <module> from tensorflow.core.framework.graph_pb2 import * File "/usr/local/lib/python2.7/
rin1024 2015/12/24
mac

deeplearning

tensorflow
ãƒªãƒ³ã‚¯
TensorFlowã§ã‚¢ãƒ‹ãƒ¡ã‚†ã‚‹ã‚†ã‚Šã®åˆ¶ä½œä¼šç¤¾ã‚’è˜åˆ¥ã™ã‚‹ - kivantiumæ´»å‹•æ—¥è¨˜
ã¤ã„ã«ã‚¬ãƒã‚¢ãŒæ»ã‚“ã å¹´é½¢ã‚’è¶…ãˆã¦ã—ã¾ã£ãŸã“ã¨ã«æ°—ãŒã¤ãã€è‡ªåˆ†ãŒã¾ã ä½•ã‚‚æˆã—é‚ã’ã¦ã„ãªã„ã“ã¨ã‚’æ‚²ã—ãæ€ã†ä»Šæ—¥ã“ã®ã”ã‚ã§ã™ã€‚ ã•ã¦ã€ä»Šæ—¥ã¯GoogleãŒå‡ºã—ãŸæ©Ÿæ¢°å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã®TensorFlowã®ä½¿ã„æ–¹ã«ã¤ã„ã¦è»½ãèª¬æ˜Žã—ã¤ã¤ã€ã‚†ã‚‹ã‚†ã‚Šã®åˆ¶ä½œä¼šç¤¾ã®è˜åˆ¥ã‚’è¡ŒãŠã†ã¨æ€ã„ã¾ã™ã€‚ TensorFlowã¨ã¯ TensorFlowã¯GoogleãŒ11/9ã«å…¬é–‹ã—ãŸApache 2.0ãƒ©ã‚¤ã‚»ãƒ³ã‚¹ã§ä½¿ãˆã‚‹æ©Ÿæ¢°å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã§ã™ã€‚Googleã¯æ§˜ã€…ãªã¨ã“ã‚ã§ãƒ—ãƒãƒ€ã‚¯ãƒˆã«æ©Ÿæ¢°å¦ç¿’ã‚’æ´»ç”¨ã—ã¦ã„ã¾ã™ãŒã€TensorFlowã¯å®Ÿéš›ã«Googleå†…éƒ¨ã®ç ”ç©¶ã§ä½¿ã‚ã‚Œã¦ã„ã‚‹ãã†ã§ã™ï¼ˆTensorFlow: Google æœ€æ–°ã®æ©Ÿæ¢°å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã‚’ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹å…¬é–‹ - Google Developer Japan Blogï¼‰ã€‚ Googleã®ãƒãƒ¼ãƒ ãƒãƒªãƒ¥ãƒ¼ã¯æã‚ã—ã„ã‚‚ã®ã§ã€GitHubã®Staræ•°ã¯ã™ã§ã«Chainerã‚„C
rin1024 2015/12/24
deeplearning
ãƒªãƒ³ã‚¯
Googleç™ºã®æ·±å±¤å¦ç¿’ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã€ŒTensorFlowã€ãŒä¸€èˆ¬ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã«ä¸Žãˆã‚‹å¯èƒ½æ€§
ã“ã‚“ã«ã¡ã¯ã€PyData.Tokyoã‚ªãƒ¼ã‚¬ãƒŠã‚¤ã‚¶ãƒ¼ã®ã‚·ãƒã‚¿ã‚¢ã‚ãƒ©ã§ã™ã€‚å…ˆæ—¥å…¬é–‹ã•ã‚Œå¤§ããªåéŸ¿ã‚’å‘¼ã‚“ã Googleç¤¾ã®æ·±å±¤å¦ç¿’ï¼ˆãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°ï¼‰ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã€TensorFlowï¼ˆãƒ†ãƒ³ã‚½ãƒ«ãƒ•ãƒãƒ¼ï¼‰ã€‚ä»Šå›žã¯PyData.Tokyoã®ãƒãƒƒã‚«ã‚½ãƒ³ã‚¤ãƒ™ãƒ³ãƒˆã§ã€å‚åŠ è€…ã®çš†ã•ã‚“ã¨TensorFlowã«ã¤ã„ã¦å¦ã³ã€ã‚³ãƒ¼ãƒ‡ã‚£ãƒ³ã‚°ã—ã€ãƒ‡ã‚£ã‚¹ã‚«ãƒƒã‚·ãƒ§ãƒ³ã—ãŸå†…å®¹ã‹ã‚‰å¾—ãŸæ‰€æ„Ÿã‚’å…±æœ‰ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚æ·±å±¤å¦ç¿’ã®ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯è¨è¨ˆã¨ã„ã†éžå¸¸ã«é«˜åº¦ãªèª²é¡Œã¯ã€ä¸€èˆ¬ã®ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã‚„ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã«ã¯ã¨ã£ã¤ãã«ãã„ã‚‚ã®ã§ã™ãŒã€æ—¢å˜ã®ã•ã¾ã–ã¾ãªæ‰‹æ³•ã‚„ãƒ„ãƒ¼ãƒ«ãŒTensorFlowã«ç§»æ¤ã•ã‚Œã¤ã¤ã‚ã‚Šã€ä»Šå¾Œã“ã“ãŒæ·±å±¤å¦ç¿’é–‹ç™ºã®æ‹ ç‚¹ã«ãªã‚‹äºˆæ„ŸãŒã—ã¾ã—ãŸã€‚ ã‚·ãƒã‚¿ã‚¢ã‚ãƒ© PyData.Tokyoã®ãƒãƒƒã‚«ã‚½ãƒ³ã‚¤ãƒ™ãƒ³ãƒˆ å‰ç½®ãï½žä»Šå›žã®ãƒãƒƒã‚«ã‚½ãƒ³ã‚’è¡Œã†ã«è‡³ã£ãŸçµŒç·¯ PyData.Tokyoã¯æ˜¨å¹´ç§‹ã«é–‹å§‹ã—ã€ç¬¬1å›žã®ãƒ†ãƒ¼ãƒžã¨ã—ã¦æ·±å±¤å¦ç¿’
rin1024 2015/12/24
deeplearning
ãƒªãƒ³ã‚¯
Sideswipe
ãƒ•ã‚©ãƒãƒ¯ãƒ¼ãŒè©±é¡Œã«ã—ã¦ã„ãŸã®ã§è¨˜äº‹ã«ã—ã¾ã—ãŸã€‚ å½“ç„¶ã€Œå¤‰ã‚ã‚‰ãªã„ã€ã¨ã„ã†è©±ãªã®ã§ã™ãŒã€ãã‚Œã ã¨é¢ç™½ãã‚ã‚Šã¾ã›ã‚“ã€‚ ã¾ã£ã¨ã†ãªç”ãˆ ã„ã„ãˆã€‚ ãƒãƒ¼ãƒ‰ãƒ‡ã‚£ã‚¹ã‚¯ã¯ãªã«ã‹ã®ç‰©è³ªã‚’å…¥ã‚ŒãŸã‚Šå‡ºã—ãŸã‚Šã™ã‚‹ã‚ã‘ã§ã¯ãªãã€ç£æ°—ã‚’å¤‰åŒ–ã•ã›ã¦ãƒ‡ãƒ¼ã‚¿ã‚’è¨˜éŒ²ã—ã¾ã™ã€‚ ç£æ°—ãŒå¤‰ã‚ã‚‹ã ã‘ãªã®ã§ã€é‡ã•ã¯å¤‰ã‚ã‚Šã¾ã›ã‚“ã€‚æƒ…å ±ã«è³ªé‡ã¯ã‚ã‚Šã¾ã›ã‚“ã€‚ ã™ã“ã—å‡ã£ãŸç”ãˆ è»½ããªã‚Šã¾ã™ã€‚ ãƒãƒ¼ãƒ‰ãƒ‡ã‚£ã‚¹ã‚¯ã¯å›žè»¢éƒ¨å“ãŒå«ã¾ã‚Œã¦ã„ã‚‹ã®ã§ã€æ¬¡ç¬¬ã«æ‘©è€—ã—ã¦ã„ãã¶ã‚“ã€ãã®æ‘©è€—ã—ãŸéƒ¨å“ãŒHDDã‹ã‚‰å¤±ã‚ã‚Œã‚‹ã¨è€ƒãˆã‚Œã°ã€ã‚ãšã‹ãšã¤è»½ããªã‚Šã¾ã™*1ã€‚ ã‚ˆã‚Šèˆˆå‘³æ·±ã„ç”ãˆ é‡ããªã‚Šã¾ã™ã€‚ ã¿ãªã•ã‚“ãŒæœŸå¾…ã—ã¦ã„ãŸå›žç”ã§ã™ï¼ã©ã†ã«ã‹ã—ã¦ã€Œãƒ‡ãƒ¼ã‚¿ã‚’ä¿å˜ã™ã‚‹ã»ã©é‡ããªã‚‹ã€ã¨ã„ã†çµè«–ã‚’å°Žãã¾ã—ã‚‡ã†ã€‚ é‡ããªã‚‹ç†ç”± è‰²ã€…ãªè€ƒãˆæ–¹ãŒã‚ã‚‹æ°—ãŒã—ã¾ã™ãŒã€ã¾ãšç›´æ„Ÿçš„ã«è€ƒãˆã‚‰ã‚Œã‚‹ã®ã¯ ãƒ©ãƒ³ãƒ€ã‚¦ã‚¢ãƒ¼ã®åŽŸç† ã¨ ç›¸å¯¾æ€§ç†è«– ã‚’ä½¿ã†ã‚‚ã®ã§ã—ã‚‡ã†ã€‚ã“ã“ã§ã‚‚ã“ã‚Œã‚’ä½¿ã£ã¦ã¿ã¾ã™ã€‚ ãƒ©ãƒ³ãƒ€ã‚¦ã‚¢
rin1024 2015/12/16
research

deeplearning

blog

*
ãƒªãƒ³ã‚¯
äººå·¥çŸ¥èƒ½ã¯ Deep Learning ã«ã‚ˆã£ã¦æˆã•ã‚Œã‚‹ã®ã‹ï¼Ÿ - Sideswipe
æœ€è¿‘ã¯äººå·¥çŸ¥èƒ½åˆ†é‡Žã®è©±é¡Œã«äº‹æ¬ ã‹ãªã„ã®ã§ã€ITç³»ã«è©³ã—ããªã„äººã§ã‚‚ã€Deep Learning ãŒã©ã†ã¨ã‹ã€äººå·¥çŸ¥èƒ½ãŒã©ã†ã¨ã‹ã¨ã„ã†è©±é¡Œã‚’è€³ã«ã™ã‚‹ã“ã¨ãŒå¤šã„ã¨æ€ã„ã¾ã™ã€‚ çŒ«ã‚‚æ“åã‚‚ Deep Learning ãªä¸–ã®ä¸ã§ã™ãŒã€ãã‚‚ãã‚‚äººå·¥çŸ¥èƒ½ã¨ã‹ Deep Learning ã£ã¦ãªã‚“ãªã‚“ã ã£ã‘ï¼Ÿ ã¨ã„ã†ç–‘å•ã«ç”ãˆã‚‰ã‚Œã‚‹äººã¯å¤šããªã„ã¯ãšã§ã™ã€‚ ä»Šå›žã¯ã€åºƒãæµ…ãã€äººå·¥çŸ¥èƒ½ã¨ Deep Learning ã«ã¤ã„ã¦æ›¸ãã¾ã™ (ã“ã®è¨˜äº‹ã‚’ã”è¦§ã«ãªã‚Œã°ã‚ã‹ã‚‹ã‚ˆã†ã«ã€äººå·¥çŸ¥èƒ½ = Deep Learning ã§ã¯æ±ºã—ã¦ç„¡ã„ã®ã§ã™ãŒã€ä¸¡è€…ã¯ã‚ˆãä¸¦ã‚“ã§ç´¹ä»‹ã•ã‚Œã‚‹ã®ã§ã€ã“ã“ã§ã‚‚åŒåˆ—ã«æ›¸ã„ã¦ã„ã¾ã™)ã€‚ æœ€åˆã«çµè«– Deep Learning ã¯(çœŸã®)äººå·¥çŸ¥èƒ½ã§ã¯ãªã„ã€‚ãªã‚“ã§ã‚‚ã‹ã‚“ã§ã‚‚äººå·¥çŸ¥èƒ½ã£ã¦å‘¼ã°ãªã„ã€‚ ã€ŒDeep Learningã€ã€ã€Œäººå·¥çŸ¥èƒ½ã€ã¨ã‚‚ã«ãƒã‚ºãƒ¯ãƒ¼ãƒ‰*1ã«ãªã‚Šã¤ã¤ã‚ã‚‹ã®ã§æ°—ã‚’ã¤ã‘ã‚ˆã†ã€‚ ã‚³ãƒ³ãƒ”
rin1024 2015/12/16
research

deeplearning
ãƒªãƒ³ã‚¯
Caffeã§Deep Learning ã¤ã¾ãšãã‚„ã™ã„ã¨ã“ã‚ã‚’ä¸å¿ƒã« - Qiita
ã¯ã˜ã‚ã« è¿‘å¹´Deep Learningã¸ã®æ³¨ç›®ãŒé«˜ã¾ã£ã¦ã„ã¾ã™ãŒã€å¤šãã®å ´åˆè†¨å¤§ãªãƒ‡ãƒ¼ã‚¿ã‚’å¿…è¦ã¨ã™ã‚‹ã“ã¨ã€å¦ç¿’ã«ã¯GPUè¨ˆç®—ç’°å¢ƒãŒå¿…è¦ã§ã‚ã£ãŸã‚Šãªã©ã€ç‹¬ç‰¹ã®æ•·å±…ã®é«˜ã•ãŒã‚ã‚Šã¾ã™ã€‚ã“ã®è¨˜äº‹ã§ã¯ã€ã“ã®æ•·å±…ã‚’å¤§ããä¸‹ã’ã‚‹ã§ã‚ã‚ã†Caffeã«ã¤ã„ã¦ç´¹ä»‹ã—ã¾ã™ã€‚ãŸã ã€Caffeã‚’ç´¹ä»‹ã™ã‚‹è¨˜äº‹ã¯ã™ã§ã«è‰¯ã„ã‚‚ã®ãŒãŸãã•ã‚“ã‚ã‚Šã€ãã‚‚ãã‚‚å…¬å¼documentãŒã‹ãªã‚Šå……å®Ÿã—ã¦ã„ã‚‹ã®ã§ã€ä»Šå›žã¯èº“ãã‚„ã™ã„éƒ¨åˆ†ã‚„ä»–ã®è¨˜äº‹ãŒã‚ã¾ã‚Šè§¦ã‚Œã¦ã„ãªã„éƒ¨åˆ†ã‚’ä¸å¿ƒã«ç´¹ä»‹ã—ã¦ã„ãã¾ã™ã€‚ Caffeã£ã¦ä½•ï¼Ÿ Caffeã¯Deep Learningã®ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã®ä¸€ã¤ã§ã™ã€‚Deep Learningã¯ä¸€èˆ¬ã«å®Ÿè£…ãŒé›£ã—ã„ã¨ã•ã‚Œã¦ã„ã¾ã™ãŒã€ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã‚’ä½¿ãˆã°ã‹ãªã‚Šæ‰‹è»½ã«æ‰±ã†ã“ã¨ãŒã§ãã¾ã™ã€‚ ä»£è¡¨çš„ãªãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã«ã¯ã€ Caffe theano/Pylearn2 Cuda-convnet2 Torch7 ãªã©ãŒã‚ã‚Šã¾ã™ã€‚ã“ã®ä¸ã§ã‚‚
rin1024 2015/12/16
deeplearning

caffe
ãƒªãƒ³ã‚¯
1