5æã®åãã«ãTensorFlowéçºè 蚌ææžãååŸããããšã«ããŸããããã®ããã«ãã¹ãã«ãåäžããããã¬ãŒãã³ã°ããã°ã©ã ãéçºãã2æ¥åïŒ6æ3æ¥ïŒã«èªå®è©Šéšã®ã¿ã¹ã¯ãå®äºããŸãããç§ã¯è©Šéšã«åæ ŒããããšãããããŸããã
ç§ããããã©ã®ããã«ãããããããŠããªããåãããã«ã©ã®ããã«ããããæããŠãã ãã
åŸ ã€ã TensorFlowãšã¯äœã§ããïŒ
TensorFlowã¯ãªãŒãã³ãœãŒã¹ã®æ°å€ã·ã¹ãã ã§ãããããŒã¿ãååŠçããŠã¢ãã«åãïŒéåžžã¯ãã£ãŒãã©ãŒãã³ã°ã«ãã£ãŠããŒã¿ããã¿ãŒã³åãïŒããœãªã¥ãŒã·ã§ã³ãå šäžçã«å±éã§ããŸãã
Googleã¯TensorFlowã䜿çšããŠããã¹ãŠã®æ©æ¢°åŠç¿ãµãŒãã¹ããµããŒãããŠããŸããããããããããèªãã§ããããã€ã¹ã¯ã以åã«äœããã®åœ¢ã§TensorFlowã䜿çšããããšããããŸãã
éåžžãéåžžã«ç解ããããPythonïŒããã¯è©Šéšã«å¿ èŠãªãã®ã§ãïŒãŸãã¯JavaScriptïŒtensorflow.jsïŒã§TensorFlowã䜿çšããŠã³ãŒããèšè¿°ããCã§èšè¿°ãããããã€ãã®åºæ¬çãªé¢æ°ãå®è¡ããŸãããããã®é¢æ°ã¯ãåè¿°ã®ã³ãã³ããå®è¡ããŸãïŒå€ãã®æ°å€èšç®ãå®è¡ããŸãïŒ ã
ããã§TensorFlowãšã¯äœããããããŸããããTensorFlowéçºè èªå®ãšã¯äœã§ããïŒãããŠããªãããªãã¯ããã«èå³ãããã®ã§ããããïŒ
TensorFlowéçºè èªå®ãšã¯äœã§ããïŒ
ãæ³åã®ãšãããTensorFlowéçºè èªå®ã¯ãTensorFlowã䜿çšããèœåãå®èšŒããæ¹æ³ã§ãã
ç¹ã«ãååž°åæãã³ã³ãã¥ãŒã¿ãŒããžã§ã³ïŒç»åã®ãã¿ãŒã³ãèŠã€ããïŒãèªç¶èšèªåŠçïŒããã¹ãã®ãã¿ãŒã³ãèŠã€ããïŒãæç³»åäºæž¬ïŒå°æ¥ã®äºæž¬ïŒãªã©ãããŸããŸãªã¿ã¹ã¯ã®ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ãæ§ç¯ãããšãã«TensorFlowïŒPythonããŒãžã§ã³ïŒã䜿çšããèœåéå»ã®ã€ãã³ãã®æ°ã«åºã¥ãåŸåïŒã
TensorFlowéçºè 蚌ææžãå¿ èŠãªã®ã¯ãªãã§ããïŒ
äžçªã®çç±ã¯æ¥œããã£ãã§ããä»äºã§å°ããã£ã¬ã³ãžããŠãè²·ã£ãæ°ããæ¬ãèªãå£å®ãèŠã€ããããªããŸããïŒè©³ããã¯åŸã§èª¬æããŸãïŒã
ããããããã«2ã€ã®çç±ããããŸãã
- , , .
- .
å°æ¥ã®éçšäž»ã«ã€ããŠèšãã°ãHacker Newsã®Who's HiringããŒãžïŒãœãããŠã§ã¢éçºè ã®æ¯æã®æ±äººãäžèŠ§è¡šç€ºããããŒãžïŒã®ããŒã¿ã«åºã¥ããšãTensorFlowã¯ä»ã®ãã£ãŒãã©ãŒãã³ã°ãã¬ãŒã ã¯ãŒã¯ãããé²ãã§ããããã§ãã
Hacker Newsã®Who's HiringããŒãžã®ããŸããŸãªåºçç©ã§èšåãããŠããé »åºŠã«åºã¥ããããŸããŸãªãã£ãŒãã©ãŒãã³ã°ãã¬ãŒã ã¯ãŒã¯ã®æ¯èŒã泚ïŒTensorFlow 2.x以éãKerasã¯åºæ¬çã«TensorFlowã®äžéšã§ãã泚2ïŒçŸåšã®äžççãªç¶æ³ã«ããããœãããŠã§ã¢éçºè ã®å šäœçãªæ¡çšçã¯äœäžããŠããŸãã
æ絊蚌ææžã¯å°±è·ã®ä¿èšŒã§ã¯ãªãããšãæ確ã«ããããšæããŸãããã ããã¹ãã«ãå€æŽããããªã³ã©ã€ã³åŠç¿ã®äžçã§ã¯ãããã¯ããªããäœãã§ãããã瀺ããã1ã€ã®æ¹æ³ã§ãã
ããã¯ãããªããåãçµãã§ããå人çãªãããžã§ã¯ãã®æ¢åã®ãªã¹ããžã®çŽ æŽãããè¿œå ã ãšæããŸã-ã³ãŒã¹ã¯åºæ¬çãªç¥èã圢æãããããžã§ã¯ãã¯å ·äœçãªç¥èã圢æããŸãã
ã§ã¯ãããã¯ã©ã®ããã«è¡ãããã®ã§ããããïŒ
è©Šéšã®æºåæ¹æ³
èå³ããããšæã£ããšããèªå®ããã°ã©ã ã®ãŠã§ããµã€ãã«ã¢ã¯ã»ã¹ããTensorFlow Developerèªå®ã¬ã€ããèªã¿ãŸããã
ããã2ã€ã®ãªãœãŒã¹ãããã«ãªãã¥ã©ã ãäœæããŸããã
ã«ãªãã¥ã©ã ã¯ãè©Šéšã«åæ Œããããã«å¿ èŠãªã¹ãã«ãéçºããããã«ç§ãåŠãã ããšãåæ ããŠããŸã
è©Šéšã®æºåãå§ããåã«ãTensorFlowã䜿çšããŠããã€ãã®ãããžã§ã¯ããæ§ç¯ããå®è·µçãªçµéšããã£ãããšã«æ³šæããŠãã ããã
çµéšè±å¯ãªTensorFlowãŸãã¯ãã£ãŒãã©ãŒãã³ã°ã®å®è·µè ã¯ã次ã®ãã¬ãŒãã³ã°ããã°ã©ã ãç§ãšåãããŒã¹ïŒåèš3é±éïŒã§å®äºããå¯èœæ§ããããŸãïŒããããããéãïŒã
åå¿è ã¯å¿ èŠãªã ãæéãè²»ããããšãã§ããŸãã䟡å€ã®ããã¹ãã«ãç¿åŸããã«ã¯æéãããããŸãã
åãªãœãŒã¹ã®ã¿ã€ãã³ã°ãã³ã¹ãïŒç±³ãã«ïŒãããã³ãŠãŒãã£ãªãã£ã¬ãã«ïŒè©Šéšã«åæ ŒïŒããªã¹ãããŸãããã¿ã€ãã³ã°ã¯ç§ã®çµéšã«åºã¥ããŠããŸãã
èªåã§ã«ãªãã¥ã©ã ãäœæãããå Žåã¯ã以äžã®ãã§ãã¯ãªã¹ãã®ãããªãã®ããå§ãããŸãã
泚ïŒã¢ãã£ãªãšã€ããªã³ã¯ã¯ææãªãœãŒã¹ã«äœ¿çšãããŸãããããã§ãªãœãŒã¹ã®äŸ¡æ Œãå€ããããšã¯ãããŸããããè³æã®1ã€ã«ã¢ã¯ã»ã¹ã§ããå Žåã¯ããã®äžéšãåãåããŸãããã®ãéã䜿ã£ãŠè³æãäœæããŸãã
1. TensorFlowéçºè èªå®ãã³ãããã¯
æéïŒ 1æéã
è²»çšïŒç¡æã
ãŠãŒãã£ãªãã£ã¬ãã«ïŒå¿ é ã
ãã®ãªãœãŒã¹ãæåã®ã¹ãããã«ãªãã¯ãã§ããè©Šéšã§æ±ããããããã¯ã«ã€ããŠèª¬æããŸãããããèªãã§ãããããäžåºŠèªãã§ãã ããã
TensorFlowãšæ©æ¢°åŠç¿ã«æ £ããŠããªãå Žåããããããããèªãã§ãããããçš®é¡ã®äºæã«æ¯ããŸããå¿é ããªãã§ãã ããã以äžã®ãªãœãŒã¹ã¯ããããã«æ £ããã®ã«åœ¹ç«ã¡ãŸãã
2. TensorFlow Coursera
æéïŒ 3é±éïŒäžçŽãŠãŒã¶ãŒïŒãã3ãæïŒåå¿è ïŒã
è²»çšïŒ 7æ¥éã®ç¡æè©Šçšã®åŸãæé¡59ãã«ã§ã財æ¿çæ¯æŽãèŠæ±ã§ããŸãã Courseraã«ã¢ã¯ã»ã¹ã§ããªãå Žåã¯ãYouTubeã§åçã®ç¡æããŒãžã§ã³ãã芧ãã ããã
ãŠãŒãã£ãªãã£ã¬ãã«ïŒ 10/10ã
ããã¯ãè©Šéšã«æãé¢é£ã®ãããªãœãŒã¹ã§ãïŒãããŠäžè¬çã«TensorFlowã䜿ãå§ããŸãïŒã泚ææ·±ããªã¹ããŒã¯TensorFlowèªå®ã¬ã€ãã«æ°ã¥ãããã®å°éåéã®èŒªéã¯ã»ãŒåãã§ãã
圌ã¯Lawrence MoroniãšAndrew Ngã2ã€ã®TensorFlowã¿ã€ã¿ã³ãšæ©æ¢°åŠç¿ã«ãã£ãŠæããããŠãããè©Šéšã®æºåã«1ã€ã®ãªãœãŒã¹ã®ã¿ãéžæããå¿ èŠãããå Žåãããã¯ãã®ã³ãŒã¹ã«ãªããŸãã
ç§ã¯çããããªåœ¢åŒã«æ¬åœã«æè¬ããã§ããã ãæ©ãå®çšçãªäŸã«çŠç¹ãåãããŸãããåã»ã¯ã·ã§ã³ã®çµããã«ããå€æ°ã®ã³ãŒããã¡ã€ã«ã¯ãçåŸãåŠç¿ããã®ã«éåžžã«åœ¹ç«ã¡ãŸãã
ããã°ã©ãã³ã°æŒç¿ã®ãã³ãïŒã³ãŒãã®ç©ºçœãåããã ãã§ãªãããã¹ãŠãèªåã§èšè¿°ããŠãã ããã
3. Scikit-LearnãKerasãTensorFlow 2nd Editionã䜿çšããå®è·µçãªæ©æ¢°åŠç¿ã
æéïŒ 3é±éïŒã«ããŒããŒã«ããŒãæŒç¿ãªãïŒã3ãæïŒã«ããŒããŒã«ããŒãæŒç¿ãè¡ãïŒã
è²»çšïŒäŸ¡æ Œã¯Amazonã«ãã£ãŠç°ãªããŸãããçŽã®ããŒãžã§ã³ã55ãã«ã§è³Œå ¥ããŸããããã¹ãŠã®ã³ãŒãã¯GitHubã§ç¡æã§è¡šç€ºã§ããŸãã
æçšæ§ïŒ 7/10ïŒäžéšã®ç« ãè©Šéšåºæã§ã¯ãªãããïŒã
ãã®æ¬ã¯700ããŒãžãè¶ ããã³ã³ãã¥ãŒã¿åŠç¿ã®ã»ãŒãã¹ãŠã®åŽé¢ãã«ããŒããŠãããããè©Šéšä»¥å€ã®ãããã¯ãå«ãŸããŠããŸããããããè©Šéšã«åæ Œããã ãã§ãªããæ©æ¢°åŠç¿ã®å°æ¥ã®åŠç¿ã®ããã®ç¢ºåºããåºç€ãç¯ãããšã«èå³ããã人ã¯å¿ èªã§ãã
æ©æ¢°åŠç¿ãåããŠäœ¿çšããå Žåã¯ããã®æ¬ãèªãã®ã¯ïŒæ ã®å§ãã«ïŒé£ããã§ããããç¹°ãè¿ããŸãããå¿é ããå¿ èŠã¯ãããŸãããæ¥ãã§è¡ãå Žæããªããæçšãªã¹ãã«ãç¿åŸããã«ã¯æéãããããŸãã
ãã®ããã«èšããŸããããæ¬ã®å質ã«ã€ããŠã®ã¢ã€ãã¢ãç¥ãããå Žåã¯ãæ©æ¢°åŠç¿ãšã³ãžãã¢ãšããŠåãããã«é転ããŠããæã®æåã®çãèªã¿ãŸããããããŠãç§ãæ¬ã§èªãã ããšã¯ãå°±æ¥äžã«æãé »ç¹ã«åœ¹ç«ã£ããšèšããã§ãããã
第2çââãåãã§ãããææ°ã®ããŒã«ãšãã¯ããã¯ãã€ãŸãè©Šéšã®ããŒã¹ãšãªã£ãŠããTensorFlow 2.xãã«ããŒããããã«æŽæ°ãããŠããç¹ãç°ãªããŸãã
è©Šéšã«é¢é£ããç« ã®ã¿ãå¿ èŠãªå Žåã¯ã以äžããèªã¿ãã ããã
- 第10ç« ïŒKerasã«ãã人工ãã¥ãŒã©ã«ãããã¯ãŒã¯ã®æŠèŠ
- 第11ç« ïŒãã£ãŒããã¥ãŒã©ã«ãããã¯ãŒã¯ã®ãã¬ãŒãã³ã°
- 第12ç« ïŒã«ã¹ã¿ã ã¢ãã«ãšTensorFlowãã¬ãŒãã³ã°
- 第13ç« ïŒTensorFlowã«ããããŒã¿ã®èªã¿èŸŒã¿ãšååŠç
- 第14ç« ïŒç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ã䜿çšãããã£ãŒãã³ã³ãã¥ãŒã¿ãŒããžã§ã³
- 第15ç« ïŒãªã«ã¬ã³ãããã³ç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ã䜿çšããã·ãŒã±ã³ã¹åŠç
- 第16ç« ïŒãªã«ã¬ã³ããã¥ãŒã©ã«ãããã¯ãŒã¯ã䜿çšããèªç¶èšèªããã¹ãåŠçãšæ³šæ
ããããçé¢ç®ãªåŠçã®å Žåã¯ãæ¬å šäœãèªãã§æŒç¿ãè¡ãããšããå§ãããŸãïŒãã¹ãŠã§ã¯ãªããããããŸããããããªãã®èå³ã«æãåã£ããã®ã§ãïŒã
4. MITããã®ãã£ãŒãã©ãŒãã³ã°ã®çŽ¹ä»
æéïŒ 3æéïŒç§ã¯3ã€ã®è¬çŸ©ããèŠãŠããŸããã§ããïŒãã1æ¥ïŒåè¬çŸ©ã«1æéããããŠåŸ©ç¿ã«1æéïŒã
è²»çšïŒç¡æã
ãŠãŒãã£ãªãã£ã¬ãã«ïŒ 8/10ã
äžçã¯ã©ã¹ã®å€§åŠããã®äžçã¯ã©ã¹ã®æ·±å±€åŠç¿ã³ãŒã¹ãç§ã¯ãããç¡æã§ããããšãèšåããã®ãå¿ããŠããŸãããïŒ
æåã®3ã€ã®è¬çŸ©ããã£ãŒãã©ãŒãã³ã°ïŒæŠèŠïŒãç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ïŒéåžžã¯ã³ã³ãã¥ãŒã¿ãŒããžã§ã³ã«äœ¿çšïŒããªã«ã¬ã³ããã¥ãŒã©ã«ãããã¯ãŒã¯ïŒéåžžã¯ããã¹ãåŠçã«äœ¿çšïŒã®ã»ã¯ã·ã§ã³ã¯ãè©Šéšã«ãšã£ãŠæãéèŠã§ãã
ããããç¹°ãè¿ãã«ãªããŸãããå€åãªãªã¹ããŒãã³ãŒã¹å šäœãå®äºããããšã¯æçã§ãã
GitHubã§æäŸãããŠããã©ããšã³ãŒããç¹ã«TensorFlowã®æŠèŠã確èªããŠãã ããããç¹°ãè¿ããŸãããç§ã¯èªå·±ã³ãŒãã£ã³ã°ã®éèŠæ§ãååã«èª¬æããããšã¯ã§ããŸããã
5. PyCharmã䜿ãå§ãã
æéïŒ 3æéïŒã³ã³ãã¥ãŒã¿ã®é床ã«ãã£ãŠç°ãªããŸãïŒã
è²»çšïŒç¡æã
ãŠãŒãã£ãªãã£ã¬ãã«ïŒ 10/10ïŒPyCharmã䜿çšããããã«å¿ èŠïŒã
è©Šéšã¯PyCharmïŒPythonéçºããŒã«ïŒã§è¡ãããŸããç§ã¯è©Šéšåã«PyCharmã䜿çšããããšããªãã£ãã®ã§ãè©Šéšãéå§ããåã«å°ãªããšãå°ãã¯æ £ããŠããããšããå§ãããŸãã
PyCharmãç解ããããã«ãYouTubeã§äžé£ã®çŽ¹ä»ãããªãèŠèŽããŸããããéåžžã«ç°¡åã§ããããããããã®ãã¿ã³ã®æ©èœã§ããã
ããããäž»ãªãã¹ãã¯ãTensorFlow 2.xãåé¡ãªãåäœããããšãšããã£ãŒããã¥ãŒã©ã«ãããã¯ãŒã¯ã劥åœãªæéå ã«æäœã§ããããšïŒç§ã®MacBook Proã«ã¯Nvidia GPUããªãïŒããã§ãã¯ããããšã§ããã
ãããã®åŽé¢ããã¹ãããããã«ã次ã®2ã€ã®TensorFlowãã¥ãŒããªã¢ã«ãããŒã«ã«ãã·ã³ã«ã³ããŒããŸããã
ããããåŸã§èŠãããã«ãè©Šéšãåãå§ããçŽåŸã«åé¡ãçºçããŸããã
ããã«
Coursera / YouTubeã®deeplearning.aiããã®ãã㪠-è©Šéšã«ã¯ããã°ã©ãã³ã°ã¿ã¹ã¯ã®å®è¡ãå«ãŸããŸãïŒPythonã³ãŒããèšè¿°ããå¿ èŠããããŸãïŒããèšè¿°ããã³ãŒãã®è£åŽã§äœãèµ·ããããç¥ãããå ŽåïŒç·åœ¢ä»£æ°ãæ°åŠçåæïŒãç§ã¯èŠãŸãå¯èœãªéããããã®ãããªãããšãã°ãããããã±ãŒãžã®åŸé éäžæ³ãããããªãå Žåã¯ããdeeplearning.aiãããããåŸé éäžæ³ãã®
TensorFlowããã¥ã¡ã³ããæ¢ããŸããTensorFlowãã©ã¯ãã£ã·ã§ããŒã«ãªãå Žåã¯ãããã¥ã¡ã³ããèªãããšãã§ããå¿ èŠããããŸããäœããåãããªãå Žåã¯ãã³ãŒããæžããŠèªåã§ã³ã¡ã³ãããŠãã ããã
YouTubeã§ã®TensorFlowã«ããããã°ã©ãã³ã°ïŒãã¬ã€ãªã¹ãïŒ -YouTubeåç»ã§Courseraã䜿çšããTensorFlowå°éåéã®ã»ãšãã©ã¯ãåãè¬åž«ã«ãã£ãŠæããããŠããŸãã
è©Šéšã®æºåæ¹æ³
äžèšã®ãªãœãŒã¹ãæŠåšã«ãNotionã§èšç»ãç«ãŠãŸããã
Notionã®TensorFlowéçºè èªå®ããã°ã©ã ãäœãããå¿ èŠããããã远跡ããããã«ãç§ã¯ããã°ãã®æ¹æ³è«ãšããŸããŸãªãªãœãŒã¹ãšã¡ã¢ã䜿çšããŸããããªã³ã¯ãã¯ãªãã¯ãããšãå³äžé ã«ãã[è€è£œ]ãã¿ã³ãã¯ãªãã¯ããŠç¬èªã®ã³ããŒãäœæã§ããŸãã
5æã®æ¯æãèµ·ããŠãæžããŠãæ©ããŠããPractical Machine Learningããšããæ¬ã1æéèªãã åŸãTensorFlowãå®éã«2ã3æé䜿ã£ãŠäœæ¥ããŸããïŒæåã«è¬çŸ©ã«åå ããŠãããGoogle Colabã§ãã¹ãŠã®ã³ãŒãã£ã³ã°æŒç¿ãè¡ããŸããïŒãåã¢ãžã¥ãŒã«ã®æåŸã§ã察å¿ããMITã®ãã£ãŒãã©ãŒãã³ã°ã®æŠèŠã®è¬çŸ©ãèŠèŽããŸããã
ããšãã°ãTensorFlowã®å®çšçãªå°éåéã®ã³ã³ãã¥ãŒã¿ãŒããžã§ã³ã®ã»ã¯ã·ã§ã³ãçµãããšããã«ãMITã®ç³ã¿èŸŒã¿ãã¥ãŒã©ã«ãããã¯ãŒã¯ïŒã³ã³ãã¥ãŒã¿ãŒããžã§ã³ã¢ã«ãŽãªãºã ã®äžçš®ïŒã«é¢ããè¬çŸ©ãèŠãŸããã
ãã®ããªãã«ã¢ãããŒãã¯ç¹ã«å¹æçã§ããããšã蚌æãããŠããŸãã
ãã®æ¬ã§ç 究ãããæŠå¿µã¯ãCourseraå°éåéã®ã³ãŒãäŸã«ãã£ãŠè£åŒ·ãããæçµçã«ã¯MITã®ãããªè³æã«ãã£ãŠèŠçŽãããŸããã
ã¿ã€ãã³ã°ãç¥ãããã«ã5æ11æ¥ã«è©Šéšã®æºåãéå§ãã6æ3æ¥ã«åæ ŒããŸããã
ç§ã®èŠ³å¯ïŒNotionã§ïŒãšææžãã®ããã¯ããŒã¯ãããç§ã¯å¹³åããŠ1æéããã20ããŒãžãå匷ãã2ã3æéã®å匷ïŒæ°æŽãããªãïŒã§çŽ1é±éã®ã³ãŒã¹å 容ãåŠç¿ããŸããã
æåŸã«ãè©Šéšã®æ°æ¥åã«ãPyCharmãããŠã³ããŒãããŠã調æ»ããã³ãŒããµã³ãã«ã®ããã€ããããŒã«ã«ãã·ã³ã§åäœããããšã確èªããŸããã
詳现-è©Šéšäžã«è¡ãããããš
ããã§ãããªãã¯èšç·ŽãçµããŸãããïŒããã§ïŒ
ããŠã2ã€ã®éèŠãªèŠçŽ ããå§ããŸãããã
è©Šéšè²»çšïŒ 100ãã«ïŒåéšã«å€±æããåŸãååéšãããŸã§2é±éåŸ ã€å¿ èŠããããŸããåéšã«å€±æãããã³ã«ãåŸ ã¡æéãé·ããªããŸãïŒã
æéïŒ 5æéãè©Šéšã®åãã«ééãããªããã°ã3æéã§ç°¡åã«åæ ŒãããšæããŸãããã ããå¶éæéãé·ããªããšãã³ã³ãã¥ãŒã¿ãŒã§ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ããã¬ãŒãã³ã°ããã®ã«ååãªæéãäžããããŸãïŒè©Šéšãå§ãŸãåã«ãã¹ãŠãæ©èœããããšã確èªããŠãã ããïŒã
è©Šéšã¯ã©ã®ããã«æé ãããŸããïŒ
å ¬å¹³ã§ã¯ãªãã®ã§ãããã§ã¯ããŸãæããã«ããŸãããTensorFlow Developer Handbookãèªãã ãã§ãè©Šéšã®äž»èŠãªã»ã¯ã·ã§ã³ãæ確ã«ãªããŸãã
ããã¥ã¢ã«ã«èšèŒãããŠããåãã¯ãããžãŒãïŒäžèšã®ãªãœãŒã¹ã䜿çšããŠïŒå®è·µãããšãæºåãæŽããŸãã
è©Šéšã®ãã¥ã¢ã³ã¹
ã¢ãã«ãã¬ãŒãã³ã° -ã³ã³ãã¥ãŒã¿ãŒã§ãã£ãŒãã©ãŒãã³ã°ã¢ãã«ãããã«æããããšãã§ããªãå ŽåïŒè©äŸ¡åºæºã®äžéšã¯ãã¬ãŒãã³ã°æžã¿ã¢ãã«ã®ãã¬ãŒã³ããŒã·ã§ã³ã§ãïŒãç¡æã®ã¯ã©ãŠãGPUã䜿çšããŠGoogle Colabã§ãã¬ãŒãã³ã°ããè©Šéšã«é©ããã«ã¿ãã°ã«é 眮ããŠããŠã³ããŒãã§ããŸãã PyCharmçµç±ã§éä¿¡ããŸãã
My Broken Python Interpreter-è©Šéšæºåè³æã§ã¯ãè©Šéšã«åæ Œããã«ã¯Python 3.7ãå¿ èŠã§ããããšã匷調ããŠããŸããç§ãå§ãããšããç§ã¯Python 3.7.3ãæã£ãŠããŸããããããŠãäœããã®çç±ã§ãTensorFlowãåæ¥ã«PyCharmã䜿çšããŠããŒã«ã«ãã·ã³ã§å®è¡ãããŠãããšããŠããè©Šéšãéå§ãããšïŒTensorFlowç°å¢ãèªåçã«äœæãããŸãïŒããã¹ãŠãå£ããŸããã
ã€ãŸããTensorFlowã³ãŒããå°ãªããšã1è¡å®è¡ãããã³ã«ããšã©ãŒãçºçããŸããã
RuntimeError: dictionary changed size during iteration
çŸåšã®ãšããããããè©Šéšãã€ã³ã¹ããŒã«ããTensorFlowã®ããŒãžã§ã³ïŒ2.0.0ïŒãªã®ãããããšãç¹å®ã®ããŒãžã§ã³ã®PythonïŒ3.7.3ïŒãªã®ãããããŸããã
ããããããã€ãã®åªããšã®å¥¥ã«æ¿åã®æ€çŽ¢åŸã¯GitHubäžã®å€ãåé¡ã¹ã¬ãããç§ãæ¥ãå šäœã§ãç§ã䜿ã£ãŠããã®Pythonç§ã®ããŒãžã§ã³ã®ãœãŒã¹ã³ãŒããå€æŽããªããã°ãªããªããã®ã§ãå¥åŠãªä¿®æ£ïŒç¹ã«lincache.pyã®48è¡ç®ïŒ ã
# Previous line 48 of lincache.py
for mod in sys.modules.values():
# Updated line 48 of linecache.py
for mod in list(sys.modules.values()): # added list()
泚ïŒããã¯è©Šéšæéäžã«ã®ã¿äœ¿çšããããããç°¡åãªä¿®æ£ã§ãããé·æçãªã¡ãªããããããã©ãããçµæã«åœ±é¿ããããã©ããã¯ããããŸããã
ç§ã®å¿ æ»ã®æ€çŽ¢äžã«ãPyCharmã§äœ¿çšããŠããTensorFlowã®ããŒãžã§ã³ãæŽæ°/åã€ã³ã¹ããŒã«ããããšã代æ¿æ段ã§ãããšèªã¿ãŸããïŒäŸïŒ2.0.0-> 2.2.xïŒãè©ŠããŠã¿ãŸãããããŸããããŸããã§ããããPyCharmã®åå¿è ãšããŠããŠãŒã¶ãŒãšããŠäœããééã£ãŠããããšãèªããŸãã
ä¿®æ£åŸãåé¡ãªãè©Šéšãçµããããšãã§ããŸããã
è©ŠéšçµäºåŸã®æµã
è©Šéšã«åæ Œãããšãã¡ãŒã«éç¥ãéä¿¡ãããŸãã ãããã§ãšããããªãã¯åæ ŒããŸããããŸãã¯ãæ®å¿µãªããä»åã¯åæ ŒããŸããã§ããã以å€ã®ã¬ãã¥ãŒã¯ãããŸããã
åŠå®çãªçµæããªããã°ãè©Šéšäžã«ããªãæ確ãªæ瀺ãå±ããŸã-åæ Œãããã©ããïŒã¢ãã«ã玹ä»ãããã³ã«ã圌女ã¯ããŒã¯ãååŸããŸãïŒã
åæ Œããããããã§ãšãããããŸãïŒ
ã¡ãŒã«ã®ãã©ãŒã ã«èšå ¥ããŠãTensorFlow Certified Developer Networkã«ç¢ºå®ã«è¿œå ãããããã«ããŠãã ããã
è©Šéšã«åæ Œãã確èªã¡ãŒã«ã®ãã©ãŒã ã«èšå ¥ãããšãæ°é±éã§Google Developers Certification Networkã«ã¢ã¯ã»ã¹ã§ããããã«ãªããŸãã
泚ïŒå·çã®æç¹ã§ã¯ãç§ã¯ããã«ããŸããã§ããã1ã2é±éããããŸãã
ç»é²ãšã¯ãçµéšè±å¯ãªTensorFlowéçºè ãæ¢ããŠãã誰ãããèªå®ã®çš®é¡ãçµéšãå°åã«åºã¥ããŠããªããèŠã€ããããšãã§ããããšãæå³ããŸãã
æåŸã«ãæ°é±é以å ã«ãæ£åŒãªèšŒææžãšTensorFlowéçºè ãããžãã¡ãŒã«ã§å±ããŸãïŒãŸã ç§ã®ãã®ãåãåã£ãŠããŸããïŒãäœæ¥ãããããžã§ã¯ãã«ããããè¿œå ããããšãã§ããŸãã
ã質å
ã³ãŒã¹ãåè¬ããæ¬ãèªãã§èªåã§ç·Žç¿ããããšã¯ã§ããŸããïŒæ¬åœã«èšŒææžãå¿ èŠã§ããïŒ
ãã¡ããã§ããŸããçµå±ã®ãšãããèªå®ã§ââã¯ãªãã¹ãã«ãç®æãå¿ èŠããããŸãã蚌ææžãæã€ããšã¯è¯ãããšã§ãããå¿ é ã§ã¯ãããŸããã
蚌ææžã¯å¿ èŠãªããšèšãå Žå-ãªããããåãåã£ãã®ã§ããïŒ
ç§ã¯ææŠããŠãããã«åãçµãããã«åãããã§ããæ¥ä»ãæå®ãããšïŒããšãã°ãã6æ3æ¥ã«è©Šéšã«åæ ŒããŸãããïŒãå匷ãããããããŸããã§ããã
ç¡æã®ãªãœãŒã¹ã§ãããè¡ãããšã¯ã§ããŸããïŒ
ã§ããŸããã TensorFlowã®ããã¥ã¡ã³ãã調ã¹ãããšã§ãå¿ èŠãªã¹ãã«ããã¹ãŠç¿åŸã§ããŸããå®éãäœããç·Žç¿ããå¿ èŠããããšãã¯ãããã¥ã¡ã³ãïŒã³ãŒãã®åè¡ïŒããäŸãã³ããŒããåè¡ã®ç解ãç·Žç¿ããŠãããèªåãèŠããã®ãç¹°ãè¿ããŠã¿ãŸãã
ãªãPyTorchã§ã¯ãªãã®ã§ããïŒ
PyTorchã倧奜ãã§ãããããã圌ãã¯èªå®ãæäŸããŠããŸããããããããªããç§ã¯ããããïŒæ¥œãã¿ã®ããã«ïŒåæ Œããã§ããããããã«ãäž¡æ¹ã®ãã¬ãŒã ã¯ãŒã¯ïŒPyTorchãšTensorFlowïŒã®çµéšè±å¯ãªãŠãŒã¶ãŒã¯ãæè¿ã®æŽæ°ã«ãã2ã€ã®ãã¬ãŒã ã¯ãŒã¯ãéåžžã«é¡äŒŒããŠããããšã«æ°ä»ãã§ããããããã«ãTensorFlowã¯äŒæ¥ã®äžçã§ãåªããŠããŸãïŒäžèšã®ã°ã©ããåç §ïŒã
æ©æ¢°åŠç¿ã«ã€ããŠäœãç¥ããŸãããã©ãããå§ããã°ããã§ããïŒ
èšäºãèªããåå¿è ã®ããã®æ©æ¢°åŠç¿ãžã®5ã€ã®ã¹ããããã
è©Šéšã«åæ ŒããGoogle Developers Certification Networkã«ç»é²ããŸããã次ã«äœãããã°ããã§ããïŒ
äœæããæãæ¥ãŸããïŒåŠãã ã¹ãã«ã䜿çšããŠãäžçã§èŠãããã®ãäœæããŸãããããŠãããªãã®äœåãå ±æããããšãå¿ããªãã§ãã ããã誰ããããèŠããã¯æ±ºããŠããããŸããã
äœãèšåããŸããã§ãããïŒã³ã¡ã³ããæ®ããããã¡ãŒã«ã§è³ªåãããããŠãã ããããããŠç§ãçããŸãã
PSããããªãèŠããæ¹ã¯ããã®èšäºã®ãããªçãäœããŸããã
SkillFactoryã®ææãªã³ã©ã€ã³ã³ãŒã¹ãåè¬ããŠãã¹ãã«ãšçµŠäžã®æ³šç®ã®è·æ¥ããŒãããååŸããæ¹æ³ã®è©³çŽ°ãã芧ãã ããã
- æ©æ¢°åŠç¿ã³ãŒã¹ïŒ12é±éïŒ
- ããŒã¿ãµã€ãšã³ã¹ã®å°éè·ããŒããããã¬ãŒãã³ã°ããïŒ12ãæïŒ
- (9 )
- «Python -» (9 )