ä»æ¥åå ãã¦ãã Deep Learning Accelerationåå¼·ä¼ - connpass ãé常ã«é¢ç½ãã£ã. ä¸åº¦èããã ãã§ã¯å ¨ç¶ç解ã§ããªãã£ãã®ã§ãå¾ããèªã¿è¿ãããã«å ¬éãããè³æãè«æãªã©ãã¡ã¢ãã¾ã¨ããã ç§èªèº«ã¯ä»äºã§ä¸åº¦tensorflowã§CNN触ã£ã¦ã¿ããããã§ãè¶ åå¿è ãªã®ã§ã ããããªæ¸ãæ¹ããã¦ããã¨ãããããããããã¾ããã¾(´ï¾Ðï¾;) [2017/09/04 11:00 追è¨] ééãã®ä¿®æ£/æ å ±è¿½è¨ ç®æ¬¡ ã¢ããã«ã¢ã¼ããã¯ãã£è¦³ç¹ããã®Deep Neural Networké«éå by Yusuke Uchida(@yu4u)ãã è³æ ããããé«éåã®æ¹æ³ Factorization conv(5 x 5) => conv(3 x 3) - conv(3 x 3) conv(3 x 3) -> conv(1 x 3) - conv (3 x
{{#tags}}- {{label}}
{{/tags}}