[B! TensorFlow] penaltyã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

penalty id:penalty

TensorFlowã«é–¢ã™ã‚‹penaltyã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (7)

${{author_name}}$

{{{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}}

ç¬¬10å›žã€€å¦ç¿’æ¸ˆã¿Inception-v3ãƒ¢ãƒ‡ãƒ«ã‚’ä½¿ã£ãŸç”»åƒèªè˜ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã‚’ä½œã‚ã†|Tech Book Zone Manatee
ä»Šå›žã¯ç¬¬9å›žã«ç¶šã„ã¦ã€å…¬é–‹ã•ã‚Œã¦ã„ã‚‹ã€ŒInception-v3ã€ã®ãƒ¢ãƒ‡ãƒ«ã‚’ç”¨ã„ã¦ã€ç”»åƒè˜åˆ¥ã®Webã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã‚’ä½œæˆã—ã¾ã™ã€‚ 1. ç”»åƒè˜åˆ¥ã‚¿ã‚¹ã‚¯ã¨åˆ†é¡žãƒ¢ãƒ‡ãƒ« ç¬¬9å›žã§ã¯ã€ç”»åƒå†…ã‹ã‚‰ã®ç‰©ä½“èªè˜ã«ã¤ã„ã¦ç´¹ä»‹ã—ã¾ã—ãŸã€‚ ã•ã¾ã–ã¾ãªç‰©ä½“ãŒå«ã¾ã‚ŒãŸç”»åƒã«å¯¾ã—ã¦ã€ç”»åƒã®ã©ã®ã‚ãŸã‚Šã«ç‰©ä½“ãŒå†™ã£ã¦ã„ã‚‹ã‹ã‚’æ¤œå‡ºã—ã€ãã®ç¨®é¡žã«ã¤ã„ã¦ã‚‚åˆ¤æ–ã—ã¦ã„ã¾ã—ãŸã€‚ä¸€æ–¹ã€ä»Šå›žæ‰±ã†ç”»åƒè˜åˆ¥ã¨ã¯ã€ä¸Žãˆã‚‰ã‚ŒãŸç”»åƒã«å¯¾ã—ã€Œãã®ç”»åƒã«ã¯ä½•ãŒå†™ã£ã¦ã„ã‚‹ã‹ã€ã‚’åˆ¤åˆ¥ã™ã‚‹ã‚¿ã‚¹ã‚¯ã§ã™ã€‚ç”»åƒè˜åˆ¥ã§ã¯ã€ã€Œä¸Žãˆã‚‰ã‚ŒãŸç”»åƒãã®ã‚‚ã®ã€ã«å¯¾ã—ã¦ã€ã€Œå†™ã£ã¦ã„ã‚‹ã‚‚ã®ãŒä½•ã‹ã€ã‚’åˆ†é¡žã™ã‚‹ã¨ã„ã†ã“ã¨ã«ãªã‚Šã¾ã™ã€‚ ã€Œä½•ãŒå†™ã£ã¦ã„ã‚‹ã‹ã€ã¨ã„ã£ã¦ã‚‚ãã®ç”ãˆã¯ç„¡é™ã«ã‚ã‚Šã€å®Ÿéš›ã«ã¯è§£æ±ºã—ãŸã„èª²é¡Œã«ã‚ˆã£ã¦ç¯„å›²ã‚„å¯¾è±¡ã‚’çµžã‚‹ã“ã¨ã«ãªã‚Šã¾ã™ã€‚ ä¾‹ãˆã°ã€Œã“ã®æ‰‹æ›¸ãæ–‡å—ã¯ã©ã®æ•°å—ã‹ã€ã§ã‚ã£ãŸã‚Šã€Œã“ã®é¡”ç”»åƒã¯ã©ã®ã‚¢ã‚¤ãƒ‰ãƒ«ã‹ã€ã§ã‚ã£ãŸã‚Šã€Œã“ã‚Œã¯æœ‰æ¯’ãªã‚¢ãƒªã‹ ãã†ã§ã¯ãªã„ã‹ã€ã§ã‚ã£ãŸã‚Š
penalty 2017/08/03
TensorFlow
ãƒªãƒ³ã‚¯
Image-to-Image Demo - Affine Layer
Recently, I made a Tensorflow port of pix2pix by Isola et al., covered in the article Image-to-Image Translation in Tensorflow. I've taken a few pre-trained models and made an interactive web thing for trying them out. Chrome is recommended. The pix2pix model works by training on pairs of images such as building facade labels to building facades, and then attempts to generate the corresponding out
penalty 2017/02/22
TensorFlow

deep learning
ãƒªãƒ³ã‚¯
TensorFlowã‚’ä½¿ã£ãŸç‚ºæ›¿(FX)ã®ãƒˆãƒ¬ãƒ¼ãƒ‰ã‚·ã‚¹ãƒ†ãƒ ã‚’ä½œã‚‹ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ« ï½žã‚·ã‚¹ãƒ†ãƒ ã®ã‚»ãƒƒãƒˆã‚¢ãƒƒãƒ—ã‹ã‚‰ãƒˆãƒ¬ãƒ¼ãƒ‰ã¾ã§ï½ž - Qiita
æ©Ÿæ¢°å¦ç¿’ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã€ŒTensorFlowã€ã¨ã€ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã®ã‚·ã‚¹ãƒ†ãƒ ãƒˆãƒ¬ãƒ¼ãƒ‰ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã€ŒJijiã€ã‚’çµ„ã¿åˆã‚ã›ã¦ã€æ©Ÿæ¢°å¦ç¿’ã‚’ä½¿ã£ãŸç‚ºæ›¿(FX)ã®ãƒˆãƒ¬ãƒ¼ãƒ‰ã‚·ã‚¹ãƒ†ãƒ ã‚’ä½œã‚‹ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã§ã™ã€‚ ã‚·ã‚¹ãƒ†ãƒ ã®ã‚»ãƒƒãƒˆã‚¢ãƒƒãƒ—ã‹ã‚‰ã¯ã˜ã‚ã¦ã€æ©Ÿæ¢°å¦ç¿’ãƒ¢ãƒ‡ãƒ«ã®ä½œæˆã€ãƒˆãƒ¬ãƒ¼ãƒ‹ãƒ³ã‚°ã€ãã‚Œã‚’ä½¿ã£ã¦å®Ÿéš›ã«ãƒˆãƒ¬ãƒ¼ãƒ‰ã‚’è¡Œã†ã¨ã“ã‚ã¾ã§ã€å…·ä½“ä¾‹ã‚’äº¤ãˆã¦è§£èª¬ã—ã¾ã™ã€‚ ã‚·ã‚¹ãƒ†ãƒ æ§‹æˆ æ¬¡ã®ã‚ˆã†ãªã‚·ã‚¹ãƒ†ãƒ ã‚’ä½œã‚Šã¾ã™ã€‚ Jijiã®ãƒãƒƒã‚¯ãƒ†ã‚¹ãƒˆæ©Ÿèƒ½ã‚’ä½¿ã£ã¦ãƒˆãƒ¬ãƒ¼ãƒ‰ãƒ‡ãƒ¼ã‚¿ã‚’åŽé›†ã€‚ã“ã‚Œã‚’TensorFlowã«å…¥åŠ›ã—ã¦ãƒ¢ãƒ‡ãƒ«ã‚’ãƒˆãƒ¬ãƒ¼ãƒ‹ãƒ³ã‚°ã—ã¾ã™ã€‚ äºˆæ¸¬ã™ã‚‹å†…å®¹ã«ã¤ã„ã¦ã¯å¾Œè¿°ã€‚ è¨“ç·´ã—ãŸãƒ¢ãƒ‡ãƒ«ã‚’ä½¿ã£ã¦äºˆæ¸¬çµæžœã‚’è¿”ã™REST APIã‚’ä½œã‚Šã€ãƒˆãƒ¬ãƒ¼ãƒ‰æ™‚ã«Jijiã‹ã‚‰å‘¼ã³å‡ºã—ã¦ä½¿ã„ã¾ã™ã€‚ ãƒ¬ãƒ¼ãƒˆæƒ…å ±ã®å–å¾—ã‚„ãƒˆãƒ¬ãƒ¼ãƒ‰ã«ã¯ã€OANDA REST API ã‚’åˆ©ç”¨ ãƒˆãƒ¬ãƒ¼ãƒ‰çŠ¶æ³ã®ç¢ºèªã‚„ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã®ç®¡ç†ã¯ã€ãƒ–ãƒ©ã‚¦ã‚¶ or ã‚¹ãƒžãƒ›ã‚¢ãƒ—ãƒªã§ å¤–å‡ºå…ˆã§ã‚‚çŠ¶æ³
penalty 2016/08/17
tensorFlow
ãƒªãƒ³ã‚¯
TensorFlowã«ã‚ˆã‚‹Deep Learningã§ã®ã‚¢ã‚¤ãƒ‰ãƒ«é¡”è˜åˆ¥ãƒ¢ãƒ‡ãƒ«ã®æ€§èƒ½è©•ä¾¡ã¨å®Ÿé¨“ - ã™ãŽã‚ƒãƒ¼ã‚“ãƒ¡ãƒ¢
ä»¥å‰ã‹ã‚‰æ›¸ã„ã¦ã„ã‚‹Deep Learningã«ã‚ˆã‚‹ã‚¢ã‚¤ãƒ‰ãƒ«é¡”è˜åˆ¥ã®è©±ã®ç¶šãã€‚ ã‚³ãƒ„ã‚³ãƒ„ã¨é¡”ç”»åƒåŽé›†ã¨ãƒ©ãƒ™ãƒ«ä»˜ã‘ã‚’ç¶šã‘ã¦ã€ãã“ãã“ã«ãƒ‡ãƒ¼ã‚¿ãŒé›†ã¾ã£ã¦ããŸã®ã§ ã“ã“ã‚‰ã§ã¡ã‚ƒã‚“ã¨æ€§èƒ½è©•ä¾¡ã‚’ã—ã¦ã¿ã‚ˆã†ã€ã¨ã€‚ ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã®ä½œæˆ ä»Šå›žã¯ã€ç¾æ™‚ç‚¹ã§é‡è¤‡ãªã180ä»¶ä»¥ä¸Šã®é¡”ç”»åƒãŒé›†ã¾ã£ã¦ã„ã‚‹40äººã®ã‚¢ã‚¤ãƒ‰ãƒ«ã‚’åˆ†é¡žå¯¾è±¡ã¨ã—ãŸã€‚ å¯¾è±¡ã‚¢ã‚¤ãƒ‰ãƒ«ä¸€è¦§ ã“ã‚Œã‚‰ã®ã‚¢ã‚¤ãƒ‰ãƒ«ã«åˆ†é¡žã®ãƒ©ãƒ™ãƒ«indexã‚’æŒ¯ã‚Š(æŽ¨ã—ã¦ã‚‹é †ã¨ã‹ã˜ã‚ƒãªãã¦ãƒ©ãƒ³ãƒ€ãƒ ã«ã)ã€ãã‚Œãžã‚Œã‹ã‚‰ç„¡ä½œç‚ºã«æŠ½å‡ºã—ãŸ180ä»¶ã®é¡”ç”»åƒã‚’ãã‚Œãžã‚Œãƒ©ãƒ™ãƒ«ã¨ã‚»ãƒƒãƒˆã§ãƒ¬ã‚³ãƒ¼ãƒ‰ã‚’ä½œã‚Šã€ã‚·ãƒ£ãƒƒãƒ•ãƒ«ã—ã¦30ä»¶ãšã¤6ã¤ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã«åˆ†ã‘ã¦ä¿å˜ã€‚ data-00.tfrecords data-01.tfrecords data-02.tfrecords data-03.tfrecords data-04.tfrecords data-05.tfrecords ãƒ¬ã‚³ãƒ¼ãƒ‰ã¯ã€ä»¥å‰ã®è¨˜
penalty 2016/06/15
tensorflow
ãƒªãƒ³ã‚¯
ç‰¹ã«ãƒ—ãƒã‚°ãƒ©ãƒžãƒ¼ã§ã‚‚ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã§ã‚‚ï¿½ãªã„ã‘ã©ã€Tensorflowã‚’1ãƒ¶æœˆè§¦ã£ãŸã®ã§è¶…åˆ†ã‹ã‚Šã‚„ã™ãè§£èª¬ - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? ã“ã‚Œæ›¸ãã ã‘ã§åœŸæ—¥ï¼’æ—¥é–“ã¾ã‚‹ã¾ã‚‹æ½°ã‚Œã¦ã—ã¾ã£ãŸã€‚ å¦ã‚“ã å†…å®¹ã«æ²¿ã£ã¦ã„ã‚‹ã®ã§ã€é †ã«èªã¿é€²ã‚ã‚‹ã«å¾“ã£ã¦ã‚³ãƒ¼ãƒ‰ã®è©±ã«ãªã£ã¦ã„ãã¾ã™ã€‚ Tensorflowè§¦ã£ã¦ã¿ãŸã„/ã¿ãŸã‘ã©ã€ã„ã‚ã„ã‚ã¾ã ç†è§£ã§ãã¦ãªã„ï¼ã¨ã„ã†æ–¹å‘ã‘ã«æ›¸ãã¾ã—ãŸã€‚ â€»2018å¹´10æœˆ4æ—¥è¿½è¨˜ å¤§åˆ†å¤ã„è¨˜äº‹ãªã®ã§ãƒªãƒ³ã‚¯åˆ‡ã‚Œã‚„å…¬å¼ãƒ‰ã‚ãƒ¥ãƒ¡ãƒ³ãƒˆãŒå¤§åˆ†å¤‰æ›´ã•ã‚Œã¦ã„ã‚‹å¯èƒ½æ€§ãŒé«˜ã„ã§ã™ã€‚ ã“ã®è¨˜äº‹ã®Tensorflowã¯ ver0.4~0.7ãã‚‰ã„ã ã£ãŸæ°—ãŒã™ã‚‹ã®ã§ ver2.0~ã¨ãªã‚Šãã†ãªç¾åœ¨ã¯æ–‡ç« ã®å¤§åŠãŒä½•ã‚’å‚è€ƒã«ã—ã¦ã„ã‚‹ã®ã‹åˆ†ã‹ã‚‰ãªã„ã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ã€‚ #1: Deep Lea
penalty 2016/05/24
tensorflow
ãƒªãƒ³ã‚¯
TensorFlowã«ã‚ˆã‚‹DCGANã§ã‚¢ã‚¤ãƒ‰ãƒ«ã®é¡”ç”»åƒç”Ÿæˆ - ã™ãŽã‚ƒãƒ¼ã‚“ãƒ¡ãƒ¢
ã‚¢ã‚¤ãƒ‰ãƒ«é¡”è˜åˆ¥ã®ãŸã‚ã®ãƒ‡ãƒ¼ã‚¿åŽé›† ã‚’ã‚³ãƒ„ã‚³ãƒ„ç¶šã‘ã¦ ãã‚Œãªã‚Šã«é›†ã¾ã£ã¦ããŸã—ã€ã“ã‚Œã‚’ä½¿ã£ã¦åˆ¥ã®ã“ã¨ã‚‚â€¦ã¨ã„ã†ã“ã¨ã§DCGANã‚’ä½¿ã£ãŸDeep Learningã«ã‚ˆã‚‹ã‚¢ã‚¤ãƒ‰ãƒ«ã®é¡”ç”»åƒã®ã€Œç”Ÿæˆã€ã‚’ã‚„ã£ã¦ã¿ãŸã€‚ ã¾ã ã ã„ã¶æªã‚“ã§ã„ãŸã‚Šã§ ã‚ã¾ã‚Šã‚ãƒ¬ã‚¤ã˜ã‚ƒãªã„ã‘ã©â€¦ã€‚é¡”ç”»åƒã‚’å¤šãåŽé›†ã§ãã¦ã„ã‚‹ã‚¢ã‚¤ãƒ‰ãƒ«90äººã®é¡”ç”»åƒã‹ã‚‰ãã‚Œãžã‚Œ120ä»¶ã‚’æŠ½å‡ºã—ãŸåˆè¨ˆ10800ä»¶ã‚’ã‚‚ã¨ã«å¦ç¿’ã•ã›ã¦ç”Ÿæˆã•ã›ãŸã‚‚ã®ã€‚ åˆ†é¡žã‚¿ã‚¹ã‚¯ã¨ã¯é€†æ–¹å‘ã®å¤‰æ›ã€è¤‡æ•°ã®ãƒ¢ãƒ‡ãƒ«å®šç¾©ãªã©ãŒã‚ã£ã¦ãªã‹ãªã‹ç†è§£ãŒé›£ã—ã„éƒ¨åˆ†ã‚‚ã‚ã£ãŸã‘ã©ã€ä½œã£ã¦ã¿ã‚‹ã¨ãã“ã¾ã§é›£ã—ãã¯ãªãã€å‡ºæ¥ä¸ŠãŒã£ã¦ã„ãéŽç¨‹ã‚’è¦‹ã‚‹ã®ãŒæ¥½ã—ã„ã— ã¨ã¦ã‚‚é¢ç™½ã„ã€‚ DCGANã¨ã¯ "Deep Convolutional Generative Adversarial Networks"ã€ç•¥ã—ã¦DCGANã€‚ã“ã¡ã‚‰ã®è«–æ–‡ã§æœ‰åã«ãªã£ãŸã€ã®ã‹ãªï¼Ÿ [1511.06434] Unsupervise
penalty 2016/05/16
tensorflow
ãƒªãƒ³ã‚¯
ã€ŒTensorFlow Tutorialã®æ•°å¦çš„èƒŒæ™¯ã€ ã‚¯ã‚¤ãƒƒã‚¯ãƒ„ã‚¢ãƒ¼ï¼ˆãƒ‘ãƒ¼ãƒˆ1ï¼‰
TensorFlowå‹‰å¼·ä¼šï¼ˆï¼“ï¼‰ ã§ç™ºè¡¨äºˆå®šã®è³‡æ–™ã§ã™ã€‚ http://connpass.com/event/27081/ å…¬é–‹å±¥æ´ 2016/03/25 ver1.0 å…¬é–‹ 2016/03/29 ver1.1 Jupyterã®ç´¹ä»‹ã‚’è¿½åŠ Read less
penalty 2016/04/13
TensorFlow
ãƒªãƒ³ã‚¯
1