ãã£ã±GPTãä»çµã¿ããåå¼·ããããã¨ããæ¬ãããã¤ãè¦ã¤ããã®ã§ã¾ã¨ãã¦ããã¾ãã
ã¾ãçè«çãªæ¦è¦ã
æ©æ¢°å¦ç¿ãããã¥ã¼ã©ã«ãããã¯ã¼ã¯ãCNNã§ã®ç»åå¦çããã©ã³ã¹ãã©ã¼ãã¼ã§ã®èªç¶è¨èªå¦çãé³å£°èªèã»åæãããã¦ããããçµã¿åããããã«ãã¢ã¼ãã«ã¨ç« ãé²ãã®ã§ãçè«çãªæ¦è¦³ãå¾ãã®ã«ããã¨æãã¾ãã
æåã¯æ°å¼ãå¤ãã®ã ãã©ãΣãforæã ã¨æãã¤ã¤ãå®ç¾©ã説ææä¸ã«åã¾ã£ã¦ãPerlãããã¡ãæªãè¨å·ä¸»ä½è¨èªã ã¨æãã°èªãããã©ããã©ãããã®ã§é£ã°ãã¦ãåé¡ãªãã¨æãã¾ãã
ã§ããããããªãä½ãã
ãã¼ã¯ãã¤ã¶ã¼ããå
¨é¨ã¤ãã£ã¦ããã¾ããTensorFlowã§BERTãã¤ãã£ã¦GPT2ãã¤ããã¨ããæµãã
ã¾ã ãã£ã¦ãªããã©ãããåçµãã¦ããã°ãã ãããç解ã§ããæ°ãããã
ããããã¯è£è¶³çãªãã®ãã
PyTorchã¨ãTensorFlowã¨ããããããããªã¼ã¨æã£ããããã®ãããã§ããããç·´ç¿ããã¨ããããã
GAN(æµå¯¾ççæãããã¯ã¼ã¯)ã§ç»åçæãªãããPyTorchã¨TensorFlowã§ä½ã£ã¦ããã¾ãã
ãã¨ããã©ã³ã¹ãã©ã¼ãã¼ããã£ã¨ç¥ãããã¨ããå ´åã«ã¯ãã®æ¬ããã§ãã
txt2imgã¯æ¦è¦ã§ã³ã¼ãã¯ãªãã§ããããã©ã³ã¹ãã©ã¼ãã¼ã«ããç»åèªèããããã試ããã³ã¼ããè¼ã£ã¦ã¾ããæè¦çã«ã¤ãããªãç»åå¦çã®ã»ããããããã
ããã¨ãããããã¢ãã«ã触ãã¨transformersã¨ããã©ã¤ãã©ãªã使ãã®ã§ã馴æãã§ããã¨ããããããã¥ã¡ã³ããå²ã¨ãã£ãããã¦ããã®ã§ãããèªãã ãChatGPTã«èãããã§ããããã ãã©ãæ¬ãã»ãã人ããã
ãã¨å²ã¨ãsã®ã¤ããªãtransformerã¯æè¡ã®ååã§sãã¤ãtransformersã¯ã©ã¤ãã©ãªã®ååã¿ããã«ãªã£ã¦ããã®ã§ãæ
£ãã¦ããã¨ããããã
é£ä¼ã«èªããã