ã¯ããã« Rubyã§ãªã«ã大è¦æ¨¡è¨èªã¢ãã« (Large Language Models, LLMs) ã«é¢ãããã¨ããã£ã¦ã¿ããã¨æã£ã¦ãSentencePieceã®binding gemããªããã¨ã«æ°ã¥ãã¦ä½ã£ãã github.com SentencePieceã¯ãã¥ã¼ã©ã«ãããã¯ã¼ã¯ãªè¨èªå¦çåãã«ä½ãããTokenizer/Detokenizerã§ãæãä¸ããããã¨åèªã«åå²ãã¦IDãå²ãå½ã¦ããéã«IDã®ãªã¹ãããæãåæ§æãããã¨ãã§ããããã®éå¤æãå¯è½ãªãã¨ã§End to Endã®...ãã¨ã¯ä½è ã®Kudo Takuããã®Qiitaè¨äºã«ã qiita.com ã¤ã³ã¹ãã¼ã« sentencepiece.rbã¯ãSentencePieceèªä½ãå梱ãã¦ããªãã®ã§ãåããã±ã¼ã¸ããã¼ã¸ã£ã¼ã§ãããã¤ã³ã¹ãã¼ã«ãããUbuntuã§ããã°aptã§ãmacOSã§ããã°homebre
ã¯ããã« llama.cppã¯Metaã®å¤§è¦æ¨¡è¨èªã¢ãã« (Large Language Models, LLMs) ã§ããLLaMAããéååãããã¨ã§CPUä¸ã§åä½ããããã®ã§ãããC/C++ã§å®è£ ããã¦ããã®ã§ãRuby bindingsãä½ã£ã¦ãRubyããå¼ã³åºããããã«ããã github.com 使ãæ¹ ã¤ã³ã¹ãã¼ã«ã¯é常ã®native extensionsãªgemã¨åæ§ã§ããã $ gem install llama_cpp ã¢ãã«ãç¨æãã¦éååããå¿ è¦ãããããllama.cppãcloneãã¦ãã¦ãUsageã®éãã«ããã°ã§ããã GitHub - ggerganov/llama.cpp: Port of Facebook's LLaMA model in C/C++ ãã¨ã¯ãæ®éã«requireããã°ä½¿ããã require 'llama_cpp' # ãã©ã¡ã¼ã¿ã
Ruby Business Users Conference 2017
èæ¯ ç¾å¨ãTensorFlowãChainerä»å¤æ°ã®DeepLearningç¨ã©ã¤ãã©ãªãå ¬éããã¦ãã¾ãã æ¬æ ¼çãªã¢ããªã±ã¼ã·ã§ã³ã§ä½¿ãã«ã¯å®è¡ã¹ãã¼ããã¯ãªãªãã£ãæ¡å¼µæ§ãããã¥ã¡ã³ããã³ãã¥ããã£ã®å å®çå¤ãã®é¢ã§ããããã®ä¸ããé¸æãã¦ä½¿ç¨ããã®ãéæ¿ãªç¶æ³ã§ãããã¡ãããç§ãã¡ã¤ã³ã§ã¯ãããã使ããã¦ããã£ã¦ãã¾ãããããã®ã©ã¤ãã©ãªãä¾ãã°tensorFlowã§ã¯computatoin graphãæ§ç¯ãoperationã追å ãã¦ãããå®è¡ã¨ããã¤ã¡ã¼ã¸ã§ï¼è¡åãæ°å¼ã§åãæ±ãã¤ã¡ã¼ã¸ï¼ãæ ¹åºã«ããå¤å ¸çãªãã¥ã¼ãã³ã®çµåã¨ããèãæ¹ãé ããã¦ããæ°ããã¾ãããããããã®ãã¨ã¯å¿ãã¦çªãé²ãã§ãã¾ã£ã¦ãè¯ãæ°ã¯ãã¾ãããèªåã®ç解ãæ·±ããæå³ã§ããã¥ã¼ãã³æåã§ã¹ã¯ã©ãããããã¥ã¼ã©ã«ãããã¯ã¼ã¯ãæ¸ãã¦ã¿ã¾ããã 使ç¨è¨èªã¯æ©æ¢°å¦ç¿åéã§ã¯Pythonã«æ®å¿µãªããé ã
ãªãªã¼ã¹ãé害æ å ±ãªã©ã®ãµã¼ãã¹ã®ãç¥ãã
ææ°ã®äººæ°ã¨ã³ããªã¼ã®é ä¿¡
å¦çãå®è¡ä¸ã§ã
j次ã®ããã¯ãã¼ã¯
kåã®ããã¯ãã¼ã¯
lãã¨ã§èªã
eã³ã¡ã³ãä¸è¦§ãéã
oãã¼ã¸ãéã
{{#tags}}- {{label}}
{{/tags}}