28. 28 Appendix: APIä¸è¦§(1/3) Model ⢠constructor: ã¢ãã«ã«å¿ è¦ãªæ§æè¦ç´ (é ã層)ãªã©ã®å®ç¾© ⢠forward(inference): constructorã§å®ç¾©ããæ§æè¦ç´ ãå©ç¨ããå ¥åãåº åã«ãã(ä¼æ¬)ããã»ã¹ãå®ç¾©ããã ⢠å¦ç¿ä¸ã¨ããã§ãªãå ´åã§æ§æãå¤ããå ´å(Dropoutãªã©)ããããå¼æ° ã«åããâ»ããã§lossãåºããªããã¨(åºãã¦ãããããoutputãã¡ãã㨠è¿ã) ModelAPI ⢠constructor: æä½éModelã®ãã¹ãåå¾ããèªã¿è¾¼ã ⢠predict: é åãªã©ã®ä¸è¬çãªå¤æ°ãããModelãå©ç¨ããäºæ¸¬å¤ãè¿ã 29. 29 Appendix: APIä¸è¦§(2/3) Trainer ⢠constructor: modelã¨å¦ç¿ã«å¿ è¦ãªãã©ã¡ã¼ã¿ã¼ãåãåãã DataProcess
{{#tags}}- {{label}}
{{/tags}}