serihiroã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ - ã¯ã¦ãªãƒ

serihiro id:serihiro

serihiroã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (11,321)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

Streamlit ã‚¢ãƒ—ãƒªã‚’ç°¡å˜ã«ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—ã‚¢ãƒ—ãƒªåŒ–ã™ã‚‹ã‚³ãƒžãƒ³ãƒ‰ã‚’ä½œã£ãŸ
Streamlit ã¯ç°¡å˜ã«ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹ç³»ã®ã‚¦ã‚§ãƒ–ã‚¢ãƒ—ãƒªã‚’æ§‹ç¯‰ã§ãã‚‹ Python ãƒ©ã‚¤ãƒ–ãƒ©ãƒªã§ã™ã€‚ã‚¦ã‚§ãƒ–ã‚¢ãƒ—ãƒªãªã®ã§ã€ã‚µãƒ¼ãƒãƒ¼ã§èµ·å‹•ã—ã¦ãƒ–ãƒ©ã‚¦ã‚¶ã§ã‚¢ã‚¯ã‚»ã‚¹ã™ã‚‹ã“ã¨ã«ãªã‚Šã¾ã™ãŒã€ãƒãƒ¼ã‚«ãƒ«ã§å‹•ä½œã™ã‚‹ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—ã‚¢ãƒ—ãƒªã«ã—ã¦é…å¸ƒã—ãŸã„çŠ¶æ³ãŒã‚ã‚Šã¾ã—ãŸã€‚ Streamlitã‚¢ãƒ—ãƒªã‚’ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—ã‚¢ãƒ—ãƒªåŒ–ã™ã‚‹æ–¹æ³•ã¯ã„ãã¤ã‹ã‚ã‚‹[1][2]ã‚ˆã†ã§ã—ãŸãŒã€åˆ¶é™ãŒã‚ã£ãŸã‚Šæ‰‹é–“ãŒã‹ã‹ã£ãŸã‚Šã™ã‚‹ã‚ˆã†ã§ã—ãŸã€‚ ã€Œã„ã‚„ã€ãŠæ‰‹è»½ã«ã‚¢ãƒ—ãƒªã«ã—ãŸã„ã‚“ã˜ã‚ƒã£ã€ã¨ã„ã†ã“ã¨ã§ã€ç°¡å˜ã«ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—ã‚¢ãƒ—ãƒªåŒ–ã§ãã‚‹ Streamlit Desktop App ã‚’ä½œã‚Šã¾ã—ãŸã€‚ã“ã®ã‚³ãƒžãƒ³ãƒ‰ã‚’ä½¿ç”¨ã™ã‚‹ã¨ã€ã‚³ãƒžãƒ³ãƒ‰ä¸€ç™ºã§ Streamlitã‚¢ãƒ—ãƒªã‚’ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—ã‚¢ãƒ—ãƒªã«å¤‰æ›ã§ãã¾ã™[3]ã€‚ã‚¦ã‚£ãƒ³ãƒ‰ã‚¦ã‚’é–‰ã˜ãŸã¨ãã« Stremlit ã®ã‚µãƒ¼ãƒãƒ¼ãƒ—ãƒã‚»ã‚¹ãŒè‡ªå‹•ã§ Kill ã•ã‚Œã‚‹ã‚ˆã†ã«ã—ãŸã‚Šã¨ã€ç´°ã‹ã„å·¥å¤«ã‚‚å…¥ã£ã¦ã„ã¾ã™ã€‚ ãªãŠã€ãƒ‡ã‚¹ã‚¯ãƒˆãƒƒãƒ—
serihiro 2024/12/15
ãƒªãƒ³ã‚¯
LLMã«ã¯"Super Weights"ãŒã‚ã‚‹ã¨ã„ã†è©±ã¨ã€ç¾å®Ÿã®ãƒ’ãƒˆã®è„³ã¨ã®é–¢é€£æ€§ã‚’è€ƒãˆã¦ã¿ã‚‹ - æ¸‹è°·é§…å‰ã§åƒããƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã®ãƒ–ãƒã‚°
X (Twitter)ã‚’çœºã‚ã¦ã„ãŸã‚‰ã€é¢ç™½ãã†ãªè«–æ–‡ãŒæµã‚Œã¦ãã¾ã—ãŸã€‚ãã‚ŒãŒã“ã¡ã‚‰ã§ã™ã€‚ å®Ÿéš›ã«æµã‚Œã¦ããŸã®ã¯ã“ã¡ã‚‰ã®ç´¹ä»‹è¨˜äº‹ãªã‚“ã§ã™ãŒã€ãã®è¦ç´„ã‚’èªã‚“ã é™ã‚Šã§ã‚‚ãªã‹ãªã‹ã«èˆˆå‘³æ·±ã„ç¾è±¡ã§ã‚ã‚‹ã‚ˆã†ã«æ€ã‚ã‚Œã¾ã™ã€‚ ã¨ã„ã†ã“ã¨ã§ã€ä½•ç•ªç…Žã˜ã‹ã‚‚ã¯ã‚„åˆ†ã‹ã‚Šã¾ã›ã‚“ãŒã“ã®ãƒ–ãƒã‚°ã§ã‚‚å‚™å¿˜éŒ²çš„ã«å–ã‚Šä¸Šã’ã¦ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚ãŒã€ãŸã ãã‚Œã ã‘ã§ã¯é¢ç™½ããªã„ã®ã§ã€ã“ã®è«–æ–‡ã‚’èªã‚“ã§åƒ•ãŒå€‹äººçš„ã«è€ƒãˆãŸã€Œç¾å®Ÿã®ãƒ’ãƒˆã®è„³ã¨ã®é–¢é€£æ€§ã€ã«ã¤ã„ã¦ã‚‚è«–ã˜ã¦ã¿ã‚‹ã“ã¨ã«ã—ã¾ã™ã€‚ è«–æ–‡ã®æ¦‚è¦ LLMã®"Super Weights"ã®å½¹å‰² "Super Weights"ã¯ãƒ‡ãƒ¼ã‚¿å…¥åŠ›ã‚’ã›ãšã¨ã‚‚ç‰¹å®šã§ãã‚‹ "Super Weights"ãŒLLMã®é‡ååŒ–ã«æžœãŸã™æ„ç¾© ç¾å®Ÿã®ãƒ’ãƒˆã®è„³ã¨ã®æ¯”è¼ƒ ãƒžã‚¯ãƒã«è¦‹ã‚Œã°ã€Œä¼¼ã¦ã„ã‚‹ã€ çŸ¥è¦šãƒ»èªçŸ¥æ©Ÿèƒ½ã«çµžã£ã¦è¦‹ã‚Œã°ã€Œç•°ãªã‚‹ã€ ã‚³ãƒ¡ãƒ³ãƒˆãªã© è«–æ–‡ã®æ¦‚è¦ åŸºæœ¬çš„ã«ã¯å†’é ã«ãƒªãƒ³ã‚¯ã—ãŸã¾ã¨ã‚è¨˜äº‹ã§ã‚‚ç´¹ä»‹ã•ã‚Œã¦ã„ã‚‹é€šã‚Šã§
serihiro 2024/12/15
ãƒªãƒ³ã‚¯
è©•ä¾¡é§†å‹•é–‹ç™ºï¼ˆEval-driven developmentï¼‰ï¼šLLMã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³é–‹ç™ºã«ãŠã‘ã‚‹èª²é¡Œã¨ã‚¢ãƒ—ãƒãƒ¼ãƒ - LayerX ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ãƒ–ãƒã‚°
ã“ã®è¨˜äº‹ã¯ã€LayerX Tech Advent Calendar 2024 ã® 12æ—¥ç›®ã®è¨˜äº‹ã§ã™ã€‚ tech.layerx.co.jp ã“ã‚“ã«ã¡ã¯ã€LayerXã®AIãƒ»LLMäº‹æ¥éƒ¨ãƒ—ãƒãƒ€ã‚¯ãƒˆãƒžãƒãƒ¼ã‚¸ãƒ£ãƒ¼ã®é‡Žç•‘(@isseinohata)ã§ã™ã€‚ AIãƒ»LLMäº‹æ¥éƒ¨ã§ã¯ç”ŸæˆAIãƒ—ãƒ©ãƒƒãƒˆãƒ•ã‚©ãƒ¼ãƒ ã€ŒAi Workforceã€ã‚’é–‹ç™ºã—ã¦ã„ã¾ã™ã€‚ getaiworkforce.com LLMã‚’ç”¨ã„ãŸã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³é–‹ç™ºã«ã¯ç‹¬è‡ªã®ç‰¹å¾´ã‚„èª²é¡ŒãŒå˜åœ¨ã—ã¦ãŠã‚Šã€Ai Workforceã®é–‹ç™ºãƒãƒ¼ãƒ ã‚‚ã€æ—¥ã€…æ§˜ã€…ãªãƒãƒ£ãƒ¬ãƒ³ã‚¸ã«å‘ãåˆã£ã¦ã„ã¾ã™ã€‚ä»Šå›žã¯ã€ãã®ä¸ã§ã‚‚ç‰¹ã«LLMã®ã€Œå‡ºåŠ›ã®ä¸ç¢ºå®šã•ã€ã«èµ·å› ã™ã‚‹é–‹ç™ºãƒ—ãƒã‚»ã‚¹ã®èª²é¡Œã‚’è§£æ±ºã™ã‚‹ãŸã‚ã®æ–¹æ³•ã¨ã—ã¦ã€è©•ä¾¡é§†å‹•é–‹ç™ºã¨ã„ã†ã‚¢ãƒ—ãƒãƒ¼ãƒã‚’ã”ç´¹ä»‹ã—ã¾ã™ã€‚ è©•ä¾¡é§†å‹•é–‹ç™ºã‚’ç´¹ä»‹ã™ã‚‹å‰ã«ã€LLMã‚’ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã«çµ„ã¿è¾¼ã‚€ä¸Šã§ã®ç‰¹å¾´ã‚„èª²é¡Œã«ã¤ã„ã¦ã€ç°¡å˜ã«ã¾ã¨ã‚ã¦ã¿ã¾ã™ã€‚ LLMã‚’
serihiro 2024/12/15
ãƒªãƒ³ã‚¯
Geminiã¨Speech-to-Textã§å®Ÿç¾ã™ã‚‹é«˜ç²¾åº¦ãªæ–‡å—èµ·ã“ã— - enechain Tech Blog
ã“ã®è¨˜äº‹ã¯enechain Advent Calendar 2024ã®14æ—¥ç›®ã®è¨˜äº‹ã§ã™ã€‚ ã¯ã˜ã‚ã« ã“ã‚“ã«ã¡ã¯ã€‚enechainã§çµ±è¨ˆãƒ»æ©Ÿæ¢°å¦ç¿’ãƒ¢ãƒ‡ãƒ«ã®æ§‹ç¯‰ã‚„LLMï¼ˆå¤§è¦æ¨¡è¨€èªžãƒ¢ãƒ‡ãƒ«ï¼‰ã®æ´»ç”¨æŽ¨é€²ã‚’æ‹…å½“ã—ã¦ã„ã‚‹@udon_tempuraã§ã™ã€‚ ç§é”ã®ãƒãƒ¼ãƒ ã§ã¯ã€ä»¥å‰ç´¹ä»‹ã—ãŸä¼šè°å‹•ç”»è¦ç´„ã®ãƒŽã‚¦ãƒã‚¦ã‚’æ´»ç”¨ã—ã€ ç¤¾å†…ç”¨ã®å‹•ç”»ãƒ»éŸ³å£°æ–‡å—èµ·ã“ã—ãƒ„ãƒ¼ãƒ«ã‚’æ§‹ç¯‰ãƒ»é‹ç”¨ã—ã¦ã„ã¾ã™ã€‚ æœ¬è¨˜äº‹ã§ã¯ã€ãã®ç²¾åº¦å‘ä¸Šã®ä»•çµ„ã¿ã«ã¤ã„ã¦ç´¹ä»‹ã—ã¾ã™ã€‚ å®Ÿè£…ã‚‚æ¯”è¼ƒçš„å°‘ãªãå®Ÿç¾ã§ããŸãŸã‚ã€ã‚»ã‚ãƒ¥ãƒªãƒ†ã‚£ãªã©ã®é–¢ä¿‚ã«ã‚ˆã‚Šå†…è£½ã§æ–‡å—èµ·ã“ã—ã‚’æ§‹ç¯‰ã—ã¦ã„ã‚‹æ–¹ã€…ã®å‚è€ƒã«ãªã‚Œã°å¹¸ã„ã§ã™ã€‚ ã¯ã˜ã‚ã« èƒŒæ™¯ã¨èª²é¡Œ ã‚·ã‚¹ãƒ†ãƒ è¦ä»¶ ã‚·ã‚¹ãƒ†ãƒ å®Ÿè£… ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£ å…¥åŠ›ãƒ‡ãƒ¼ã‚¿å½¢å¼ ãƒ‰ãƒ¡ã‚¤ãƒ³å›ºæœ‰ç”¨èªžã®ç®¡ç† å‡¦ç†ãƒ•ãƒãƒ¼ Cloud Speech-to-Textã«ã‚ˆã‚‹æ–‡å—èµ·ã“ã— Gemini 1.5 Flashã«ã‚ˆã‚‹èªè˜å‡¦ç† Gemini 1.5 Proã«ã‚ˆ
serihiro 2024/12/15
ãƒªãƒ³ã‚¯
è‡ªå‹•åŒ–ã™ã‚‹LLMã‚·ã‚¹ãƒ†ãƒ ã®å“è³ªç®¡ç†: LLM-as-a-judge ã®ä½œã‚Šæ–¹ - Gaudiy Tech Blog
ã“ã‚“ã«ã¡ã¯ã€‚ãƒ•ã‚¡ãƒ³ã¨å…±ã«æ™‚ä»£ã‚’é€²ã‚ã‚‹ã€Web3ã‚¹ã‚¿ãƒ¼ãƒˆã‚¢ãƒƒãƒ— Gaudiy ã® seya (@sekikazu01)ã¨ç”³ã—ã¾ã™ã€‚ LLM-as-a-Judgeã€åå‰ã®é€šã‚Šä½•ã‹ã—ã‚‰ã®è©•ä¾¡ã«LLMã‚’ä½¿ã†æ‰‹æ³•ã§ã™ã€‚ LLMã¯éžæ§‹é€ åŒ–ãƒ‡ãƒ¼ã‚¿ã§ã‚‚ã†ã¾ã„ã“ã¨è§£é‡ˆã—ã¦ãã‚Œã‚‹å„ªã‚Œã‚‚ã®ãªãŸã‚ã€é€šå¸¸ã®ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã§ã¯åˆ¤æ–ã™ã‚‹ã“ã¨ãŒé›£ã—ã„ã“ã¨ã§ã‚‚ã€ãã‚Œã‚‰ã—ãåˆ¤æ–ã‚„ã‚¹ã‚³ã‚¢ã‚’ã¤ã‘ã¦ãã‚Œã¾ã™ã€‚ æœ¬è¨˜äº‹ã§ã¯ãã‚“ãªLLM-as-a-Judgeã‚’ã€LLMã‚’ä½¿ã£ãŸã‚·ã‚¹ãƒ†ãƒ ã®æŒ¯ã‚‹èˆžã„ã®æ”¹å–„ã®ç¢ºèª and ãƒ‡ã‚°ãƒ¬æ¤œçŸ¥ç›®çš„ã§ä½œã£ã¦ã„ãéš›ã®ã€ç§ãªã‚Šã®æµã‚Œã‚„Tipsãªã©ã‚’ç´¹ä»‹ã—ã¦ã„ãã¾ã™ã€‚ ãªãœLLMã§Judgeã—ãŸã„ã®ã‹ï¼Ÿ ã¯ã˜ã‚ã«ã€ŒãªãœLLMã§ã®è©•ä¾¡ãŒã—ãŸã„ã®ã‹ã€ã€ãã®ç›®çš„ã‚’æ•´ç†ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ ã¾ãšå¤§ä¸Šæ®µã®ç›®çš„ã¨ã—ã¦ã€LLMã‚·ã‚¹ãƒ†ãƒ ãŒæ”¹å–„ã—ã¦ã„ã‚‹ã‹ã€ã¾ãŸã¯åŠ£åŒ–(ãƒ‡ã‚°ãƒ¬)ã—ã¦ãªã„ã‹ã‚’æ¤œçŸ¥ã—ãŸã„ã¨ã„ã†ã®ãŒã‚ã‚Šã¾ã™ã€‚ LLMã‚’ä½¿ã£
serihiro 2024/12/13
LLM
ãƒªãƒ³ã‚¯
é¢å€’ãªã“ã¨ã¯LLMã«ã‚„ã‚‰ã›ã‚ˆã† Guest:ã‹ã‚‰ã‚ã’ "Beginning LLM Level.9"
serihiro 2024/12/13
LLM
ãƒªãƒ³ã‚¯
The Little Book of Deep Learning
serihiro 2024/12/13
tutorial

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
https://x.com/SnowySnowly/status/1859081379264319643
serihiro 2024/12/13
ãŸã¾ã«è¦‹ãŸããªã‚‹
ãƒªãƒ³ã‚¯
Inside NotebookLM with Raiza Martin and Steven Johnson
serihiro 2024/12/13
ãƒªãƒ³ã‚¯
ã€AWS re:Invent 2024ã€‘RAG é–¢é€£ã®5å€‹ã®ã‚»ãƒƒã‚·ãƒ§ãƒ³ã®å¦ã³ã‚’æ¿ƒç¸®ã—ã¦ãŠå±Šã‘ã—ã¾ã™ - ã‚«ãƒŸãƒŠã‚· ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ãƒ–ãƒã‚°
ã“ã‚“ã«ã¡ã¯ã€AWS re:Invent ã‹ã‚‰å¸°å›½ã—ãŸ a2 (@Atsuhiro_tim) ã§ã™ã€‚ã™ãå®¶ãŒã‚µãƒ©ãƒ€ã‚’ã¤ã‘ã¦ã‚‚ $5 ã§ã€ãã®ç¾Žå‘³ã—ã•ã¨å®‰ã•ã«æ¶™ã‚’æµã—ã¦ã„ã¾ã™ã€‚ ã•ã¦ã€AWS re:Invent 2024 ã®ã‚»ãƒƒã‚·ãƒ§ãƒ³ã‚«ã‚¿ãƒã‚°ã‚’è¦‹ã‚‹ã¨ã€ä»Šå¹´ã‚‚ GenAI ãŒçŒ›å¨ã‚’æŒ¯ã‚‹ã£ã¦ã„ãŸã“ã¨ãŒã‚ã‹ã‚Šã¾ã™ã€‚ æ˜¨å¹´ã¯ Gen AI x ã€‡ã€‡ ãŒå¤šã‹ã£ãŸã®ã§ã™ãŒã€ä»Šå¹´ã¯ä¸€æ©é€²ã‚“ã§ã€ RAG x ã€‡ã€‡ ã‚„ Agent x ã€‡ã€‡ ãŒå‡ºã¦ãã¾ã—ãŸã€‚RAG ã«ã¤ã„ã¦ã¯æ©Ÿèƒ½ãƒªãƒªãƒ¼ã‚¹ã®ãƒ‹ãƒ¥ãƒ¼ã‚¹ã¯èªè˜ã—ã¦ã„ãŸã‚‚ã®ã®ã€å®Ÿéš›ã«è§¦ã‚‹ã“ã¨ãŒãªã‹ã£ãŸã®ã§ã€ä»Šå›žã® re:Invent ã§ Gen AI ã‚„ RAG é–¢é€£ã®ã‚»ãƒƒã‚·ãƒ§ãƒ³ã«å‚åŠ ã—ã¦ãã¾ã—ãŸã€‚è¤‡æ•°ã®ã‚»ãƒƒã‚·ãƒ§ãƒ³ã®å¦ã³ã‚’ã¾ã¨ã‚ã¦ã‚·ã‚§ã‚¢ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ å‚åŠ ã—ãŸã‚»ãƒƒã‚·ãƒ§ãƒ³ï¼š Explore generative AI use cases with LangChain
serihiro 2024/12/13
ãƒªãƒ³ã‚¯
ä»•äº‹ãŒå¤šã™ãŽã¦çµ‚ã‚ã‚‰ãªã„äººãŒçŸ¥ã‚‰ãªã„â€ãƒžãƒ‹ãƒ£ãƒ¼ãƒŠã®æ³•å‰‡â€ï½œæ®‹æ¥æ¿€æ¸›ã€é€†è»¢ã®ä»•äº‹è¡“ - STUDY HACKERï¼ˆã‚¹ã‚¿ãƒ‡ã‚£ãƒ¼ãƒãƒƒã‚«ãƒ¼ï¼‰ï½œç¤¾ä¼šäººã®å‹‰å¼·æ³•ï¼†è‹±èªžå¦ç¿’
ã€Œã¨ã«ã‹ãä»•äº‹ãŒå¤šã™ãŽã¦å›°ã£ã¦ã„ã‚‹ï¼ã€ ã‚¿ã‚¹ã‚¯ãŒæ¬¡ã‹ã‚‰æ¬¡ã¸ã¨é™ã£ã¦ãã¦ã€ä»•äº‹ãŒçµ‚ã‚ã‚‰ãªã„çŠ¶æ³ã§æ‚©ã‚“ã§ã„ã¾ã›ã‚“ã‹ï¼Ÿã€€åŠ¹æžœçš„ãªå¯¾å‡¦æ³•ãŒè¦‹ã¤ã‹ã‚‰ãšã€ãƒãƒ£ãƒƒãƒˆã«è¿”ä¿¡ã—ãªãŒã‚‰è³‡æ–™ä½œæˆã€ãƒ¡ãƒ¼ãƒ«ãƒã‚§ãƒƒã‚¯ã—ãªãŒã‚‰é›»è©±å¯¾å¿œ...ã€‚ã“ã‚“ãªåƒãæ–¹ã‚’ç¶šã‘ã¦ã„ã¦ã¯ã€é™ç•ŒãŒæ¥ã¦ã—ã¾ã„ã¾ã™ã€‚ ã€Œä»•äº‹ã®åŠ¹çŽ‡åŒ–ã®ãŸã‚ã«ã¯ã€ãƒžãƒ«ãƒã‚¿ã‚¹ã‚¯ã§ã“ãªã™ã—ã‹ãªã„ã€ ãã†è€ƒãˆã¦ã—ã¾ã†äººã‚‚å¤šã„ã¯ãšã§ã™ã€‚ç¢ºã‹ã«ä¸€è¦‹ã€è¤‡æ•°ã®ä»•äº‹ã‚’åŒæ™‚ã«ã“ãªã™ã“ã¨ãŒã€ä»•äº‹ãŒå¤šã™ãŽã‚‹çŠ¶æ³ã¸ã®å¯¾å‡¦æ³•ã«æ€ãˆã¾ã™ã€‚ ã—ã‹ã—ã€ã€Œãƒžãƒ«ãƒã‚¿ã‚¹ã‚¯ã§ä»•äº‹ã‚’ã“ãªãã†ã€ã¨ã„ã†ãã®ç™ºæƒ³è‡ªä½“ãŒã€å®Ÿã¯ä»•äº‹ãŒçµ‚ã‚ã‚‰ãªã„åŽŸå› ã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ã€‚æœ¬è¨˜äº‹ã§ã¯ã€åŠ¹çŽ‡çš„ãªä»•äº‹ã®é€²ã‚æ–¹ã«æ‚©ã‚€æ–¹ã«ã€æ„å¤–ãªè§£æ±ºæ–¹æ³•ã‚’ã”ç´¹ä»‹ã—ã¾ã™ã€‚ãã‚Œã¯ã€ã€Œæ˜Žæ—¥ã§ãã‚‹ã“ã¨ã¯ä»Šæ—¥ã‚„ã‚‰ãªã„ã€ã¨ã„ã†é€†èª¬çš„ãªå¯¾å‡¦æ³•ã§ã™ã€‚ ãƒžãƒ«ãƒã‚¿ã‚¹ã‚¯ãŒä»•äº‹ã‚’çµ‚ã‚ã‚‰ãªãã™ã‚‹ç†ç”± ã‚·ãƒ³ã‚°ãƒ«ã‚¿ã‚¹ã‚¯ã§ä»•äº‹ãŒå¤šã™ãŽã‚‹çŠ¶æ³ã«å¯¾å‡¦ã™ã‚‹ ä»•äº‹åŠ¹çŽ‡åŒ–ã®æ–°ã—ã„å¯¾å‡¦æ³•ã€Œãƒž
serihiro 2024/12/13
ãƒªãƒ³ã‚¯
ç”ŸæˆAIã§è‡ªå‹•ãƒ†ã‚¹ãƒˆã‚’æ¥½ã«ä½œã‚ŠãŸã„ï¼ - Tabelog Tech Blog
ã“ã®è¨˜äº‹ã¯ é£Ÿã¹ãƒã‚°ã‚¢ãƒ‰ãƒ™ãƒ³ãƒˆã‚«ãƒ¬ãƒ³ãƒ€ãƒ¼2024 ã®12æ—¥ç›®ã®è¨˜äº‹ã§ã™ðŸŽ…ðŸŽ„ ç›®æ¬¡ ç›®æ¬¡ ã¯ã˜ã‚ã« è‡ªå‹•ãƒ†ã‚¹ãƒˆä½œæˆã®èª²é¡Œ ãƒ†ã‚¹ãƒˆã‚±ãƒ¼ã‚¹ã‚’è€ƒãˆã‚‹ã“ã¨ã®é›£ã—ã• ãƒ†ã‚¹ãƒˆã‚³ãƒ¼ãƒ‰ã«è½ã¨ã—è¾¼ã‚€ä½œæ¥ã®è² æ‹… ãƒ†ã‚¹ãƒˆå¯¾è±¡ã®ã‚³ãƒ¼ãƒ‰ä¾‹ RSpecã§ã®ãƒ†ã‚¹ãƒˆã‚³ãƒ¼ãƒ‰ä¾‹ è‡ªå‹•ãƒ†ã‚¹ãƒˆä½œæˆã®èª²é¡ŒãŒã‚‚ãŸã‚‰ã™å½±éŸ¿ ç”ŸæˆAIã¨è‡ªå‹•ãƒ†ã‚¹ãƒˆ è‡ªå‹•ãƒ†ã‚¹ãƒˆä½œæˆã®åŠ¹çŽ‡åŒ–ã‚’ç›®æŒ‡ã—ã¦ å°Žå…¥ã®æ¡ä»¶ Difyã‚’æ´»ç”¨ã—ãŸãƒãƒ£ãƒƒãƒˆãƒœãƒƒãƒˆ ãƒãƒ£ãƒƒãƒˆãƒœãƒƒãƒˆã®åˆ©ç”¨æ–¹æ³• ãƒ†ã‚¹ãƒˆç”Ÿæˆã®éšœå®³ å®Ÿè£…ã‚³ãƒ¼ãƒ‰ã‚’ãã®ã¾ã¾é€ã£ãŸå ´åˆã®å•é¡Œç‚¹ è‰¯ã„è‡ªå‹•ãƒ†ã‚¹ãƒˆç”ŸæˆãŒå¯èƒ½ãªã‚±ãƒ¼ã‚¹ ãƒ—ãƒãƒ³ãƒ—ãƒˆã®å·¥å¤« ãƒ—ãƒãƒ³ãƒ—ãƒˆã‚„å¯¾è©±ã®å·¥å¤«ã§ã¯è§£æ±ºã§ããªã„ã“ã¨ è¨è¨ˆã®é‡è¦æ€§ ã¾ã¨ã‚ ãƒã‚¸ãƒƒã‚¯ã‚’åˆ‡ã‚Šå‡ºã—ãŸå¯¾è±¡ã§ã®æœ‰åŠ¹æ€§ è‡ªå‹•ãƒ†ã‚¹ãƒˆã¨è¨è¨ˆã®ç›¸ä¹—åŠ¹æžœ æœ€å¾Œã« ã¯ã˜ã‚ã« ã“ã‚“ã«ã¡ã¯ã€‚é£Ÿã¹ãƒã‚°é–‹ç™ºæœ¬éƒ¨ ã‚¦ã‚§ãƒ–é–‹ç™º2éƒ¨ ç¬¬1ãƒ—ãƒãƒ€ã‚¯ãƒˆãƒãƒ¼ãƒ ã§ä¸»ã«ãƒãƒƒã‚¯ã‚¨ãƒ³ãƒ‰å‘¨ã‚Šã®é–‹ç™ºã‚’æ‹…å½“ã—ã¦ã„ã¾ã™ã€ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã®é«˜ç”°ã§ã™
serihiro 2024/12/13
ãƒªãƒ³ã‚¯
è‡ªå¾‹çš„ã«ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚’ã™ã‚‹AIã‚¨ãƒ¼ã‚¸ã‚§ãƒ³ãƒˆã€ŒJulesã€ã€GoogleãŒç™ºè¡¨ã€‚äººé–“ãŒã‚¿ã‚¹ã‚¯ã‚’ä¸Žãˆã‚‹ã¨å®Ÿè£…è¨ˆç”»ã‚’ä½œæˆã€ã‚³ãƒ¼ãƒ‰ã®ç”Ÿæˆã‚„å¤‰æ›´ã€ãƒã‚°ãƒ•ã‚£ã‚¯ã‚¹ãªã©ã‚’å®Ÿè¡Œ
Googleã¯ã€äººé–“ãŒã‚¿ã‚¹ã‚¯ã‚’ä¸Žãˆã‚‹ã¨è‡ªå¾‹çš„ã«å®Ÿè£…è¨ˆç”»ã‚’ç«‹ã¦ã¦ã‚³ãƒ¼ãƒ‰ã®ç”Ÿæˆã‚„å¤‰æ›´ã€ãƒã‚°ãƒ•ã‚£ã‚¯ã‚¹ãªã©ã‚’å®Ÿè¡Œã—ã¦ãã‚Œã‚‹AIã‚¨ãƒ¼ã‚¸ã‚§ãƒ³ãƒˆã€ŒJulesã€ã‚’ç™ºè¡¨ã—ã¾ã—ãŸã€‚ åŒç¤¾ãŒç™ºè¡¨ã—ãŸæœ€æ–°ã®ç”ŸæˆAIãƒ¢ãƒ‡ãƒ«ã§ã‚ã‚‹Gemini 2.0ãŒç”¨ã„ã‚‰ã‚Œã¦ã„ã¾ã™ã€‚ ã‚¿ã‚¹ã‚¯ã‚„ã‚¤ã‚·ãƒ¥ãƒ¼ã‚’ä¸Žãˆã‚‹ã¨ã€ãã‚Œã‚’èµ·ç‚¹ã«è‡ªå¾‹çš„ãªãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚’è¡Œã†ç”ŸæˆAIã‚’ç”¨ã„ãŸã‚µãƒ¼ãƒ“ã‚¹ã¯ã€å…ˆæ—¥æ£å¼ã‚µãƒ¼ãƒ“ã‚¹åŒ–ã•ã‚ŒãŸDevinã‚„ã€ç¾åœ¨ãƒ†ã‚¯ãƒ‹ã‚«ãƒ«ãƒ—ãƒ¬ãƒ“ãƒ¥ãƒ¼ä¸ã®GitHub Copilot Workspaceãªã©ã€ã™ã§ã«å…ˆè¡Œã—ã¦ã„ã‚‹ã‚µãƒ¼ãƒ“ã‚¹ãŒå˜åœ¨ã—ã¾ã™ã€‚ å‚è€ƒï¼šGitHubã€ã€ŒCopilot Workspaceã€ãƒ†ã‚¯ãƒ‹ã‚«ãƒ«ãƒ—ãƒ¬ãƒ“ãƒ¥ãƒ¼ã‚’é–‹å§‹ã€‚ã»ã¨ã‚“ã©å…¨ã¦ã®é–‹ç™ºå·¥ç¨‹ã‚’AIã§è‡ªå‹•åŒ– Googleã‚‚Gemini 2.0ã«ã‚ˆã£ã¦ã€åŒæ§˜ã«è‡ªå¾‹çš„ãªãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ãŒå¯èƒ½ãªãƒ¬ãƒ™ãƒ«ã®ã‚µãƒ¼ãƒ“ã‚¹ã‚’é–‹ç™ºå¯èƒ½ã«ãªã£ãŸã¨ã„ã†ã“ã¨ã§ã—ã‚‡ã†ã€‚ JulesãŒå‹•ä½œã™ã‚‹æ§˜å Google
serihiro 2024/12/13
ãƒªãƒ³ã‚¯
ãƒžã‚¹ã‚¯æ°ã®ç´”è³‡ç”£ã€äººé¡žåˆã®60å…†å††è¶…ãˆ
ã€12æœˆ12æ—¥ AFPã€‘ä¸–ç•Œã®å¯Œè±ªä¸Šä½500äººã®ç´”è³‡ç”£ã‚’ãƒ©ãƒ³ã‚¯ä»˜ã‘ã™ã‚‹ã€Œãƒ–ãƒ«ãƒ¼ãƒ ãƒãƒ¼ã‚°ãƒ»ãƒ“ãƒªã‚ªãƒã‚¢æŒ‡æ•°ã€ã«ã‚ˆã‚‹ã¨ã€ä¸–ç•Œä¸€ã®å¯Œè±ªã§ã‚ã‚‹ç±³å®Ÿæ¥å®¶ã‚¤ãƒ¼ãƒãƒ³ãƒ»ãƒžã‚¹ã‚¯æ°ã®ç´”è³‡ç”£ãŒ11æ—¥æ™‚ç‚¹ã§4000å„„ãƒ‰ãƒ«ï¼ˆç´„61å…†å††ï¼‰ã‚’è¶…ãˆãŸã€‚ç´”è³‡ç”£ãŒ4000å„„ãƒ‰ãƒ«å°ã«åˆ°é”ã—ãŸå€‹äººã¯ãƒžã‚¹ã‚¯æ°ãŒä¸–ç•Œã§åˆã‚ã¦ã€‚ ãƒ–ãƒ«ãƒ¼ãƒ ãƒãƒ¼ã‚°é€šä¿¡ã«ã‚ˆã‚‹ã¨ã€ãƒžã‚¹ã‚¯æ°çŽ‡ã„ã‚‹éžä¸Šå ´ã®å®‡å®™ä¼æ¥ã‚¹ãƒšãƒ¼ã‚¹Xã®ä¼æ¥ä¾¡å€¤ã‚’ç´„3500å„„ãƒ‰ãƒ«ï¼ˆç´„53å…†å††ï¼‰ã¨è©•ä¾¡ã—ã€æœ€å¤§12å„„5000ä¸‡ãƒ‰ãƒ«ç›¸å½“ã®åŒç¤¾æ ªã‚’æŠ•è³‡å®¶ãŒè³¼å…¥ã™ã‚‹ã“ã¨ã§ä¼šç¤¾å´ã¨åˆæ„ã—ãŸã®ã‚’å—ã‘ã€ãƒžã‚¹ã‚¯æ°ã®ç´”è³‡ç”£ã¯ç´„500å„„ãƒ‰ãƒ«è·³ãä¸ŠãŒã‚Šã€4400å„„ãƒ‰ãƒ«ï¼ˆç´„67å…†å††ï¼‰ã«é”ã—ãŸã€‚ ãƒžã‚¹ã‚¯æ°ã®è³‡ç”£ã¯ä¸»ã«ã€ã‚¹ãƒšãƒ¼ã‚¹Xã¨é›»æ°—è‡ªå‹•è»Šï¼ˆEVï¼‰å¤§æ‰‹ãƒ†ã‚¹ãƒ©ã®æ ªå¼ã®è©•ä¾¡é¡ã«åŸºã¥ã„ã¦ã„ã‚‹ã€‚ä¸¡ç¤¾ã®æ ªä¾¡ã¯å…ˆæœˆã€ãƒ‰ãƒŠãƒ«ãƒ‰ãƒ»ãƒˆãƒ©ãƒ³ãƒ—å‰å¤§çµ±é ˜ãŒå¤§çµ±é ˜é¸ã«å‹åˆ©ã—ãŸã®ã‚’å—ã‘ã€å¤§å¹…ä¸Šæ˜‡ã€‚å¤§çµ±é ˜é¸ä»¥æ¥ã®ãƒ†ã‚¹ãƒ©æ ªã®å€¤ä¸ŠãŒã‚Šå¹…ã¯ç´„65ï¼…ã«é”ã—
serihiro 2024/12/12
ãƒªãƒ³ã‚¯
SQLiteã‚’Rustã§æ›¸ãç›´ã—ãŸã€ŒLimboã€ãŒæµ·å¤–ã§è©±é¡Œã« â€” å®Œå…¨ãªéžåŒæœŸI/Oã®ã‚µãƒãƒ¼ãƒˆã€WASMå¯¾å¿œã€ãƒ¡ãƒ¢ãƒªå®‰å…¨æ€§ã®ç¢ºä¿ãªã©ã€Rustã®ãƒ¡ãƒªãƒƒãƒˆã‚’å…¨é¢çš„ã«äº«å—
serihiro 2024/12/12
ãƒªãƒ³ã‚¯
å±¥æ´ãƒ†ãƒ¼ãƒ–ãƒ«ã‹ã‚‰æœ€æ–°ã®1ä»¶ã‚’å–ã£ã¦ãã‚‹æ–¹æ³• - ããƒ¼ã ã„ãªã‚‹ã‚‰ããŒãå¸³
ä¾‹ãˆã°æ¬¡ã®ã‚ˆã†ãªãƒ†ãƒ¼ãƒ–ãƒ«ãŒã‚ã£ãŸã¨ã™ã‚‹ã€‚ -- PostgreSQL CREATE TABLE history ( id SERIAL PRIMARY KEY, user_id INTEGER NOT NULL, data TEXT, created_at TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ); -- MySQL CREATE TABLE history ( id INT AUTO_INCREMENT PRIMARY KEY, user_id INT NOT NULL, data TEXT, created_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP ); INSERT INTO history (user_id, data, created_at) VALUES (1, 'First
serihiro 2024/12/11
ãƒªãƒ³ã‚¯
Sora
Transf orm text and images into immersive videos. Animate stories, visualize ideas, and bring your concepts to life.
serihiro 2024/12/11
ãƒªãƒ³ã‚¯
NVIDIAè£½ã‚°ãƒ©ãƒœã‚„Appleãƒãƒƒãƒ—ã®AIæŽ¨è«–å‡¦ç†èƒ½åŠ›ã‚’ã¾ã¨ã‚ãŸä¸€è¦§è¡¨ã€ã©ã®ã‚°ãƒ©ãƒœã‚„Macã‚’è²·ãˆã°ã„ã„ã®ã‹ã®å‚è€ƒã«ãªã‚‹
ä¸€æ˜”å‰ã¾ã§ã¯ã‚°ãƒ©ãƒ•ã‚£ãƒƒã‚¯ãƒœãƒ¼ãƒ‰ã®ãƒ¡ã‚¤ãƒ³ç”¨é€”ã¯ã‚²ãƒ¼ãƒ ãªã©ã®3Dã‚°ãƒ©ãƒ•ã‚£ãƒƒã‚¯å‡¦ç†ã§ã—ãŸãŒã€è¿‘å¹´ã§ã¯ã€Œãƒãƒ¼ã‚«ãƒ«ã§AIã§å‹•ä½œã•ã›ã‚‹ã€ã¨ã„ã†ã“ã¨ã‚’ç›®çš„ã«ã‚°ãƒ©ãƒ•ã‚£ãƒƒã‚¯ãƒœãƒ¼ãƒ‰ã‚’é¸æŠžã™ã‚‹äº‹ä¾‹ã‚‚å¢—ãˆã¦ã„ã¾ã™ã€‚å¤§é‡ã®NVIDIAè£½ã‚°ãƒ©ãƒ•ã‚£ãƒƒã‚¯ãƒœãƒ¼ãƒ‰ã‚„Appleè£½ãƒãƒƒãƒ—ã§å¤§è¦æ¨¡è¨€èªžãƒ¢ãƒ‡ãƒ«ã€ŒLLaMA 3ã€ã®æŽ¨è«–å‡¦ç†ã‚’å®Ÿè¡Œã—ãŸéš›ã®å‡¦ç†æ€§èƒ½ã‚’ã¾ã¨ã‚ãŸã‚¦ã‚§ãƒ–ãƒšãƒ¼ã‚¸ã€ŒGPU-Benchmarks-on-LLM-Inferenceã€ã‚’è¦‹ã¤ã‘ãŸã®ã§ã€å†…å®¹ã‚’ã¾ã¨ã‚ã¦ã¿ã¾ã—ãŸã€‚ GitHub - XiongjieDai/GPU-Benchmarks-on-LLM-Inference: Multiple NVIDIA GPUs or Apple Silicon for Large Language Model Inference? https://github.com/XiongjieDai/GPU-Benchmarks-on-
serihiro 2024/12/09
ãƒªãƒ³ã‚¯
ã€è¥¿å·å’Œä¹…ã®ä¸å®šæœŸã‚³ãƒ©ãƒ ã€‘ M4 Maxæè¼‰ã€ŒMacBook Pro 14ã€ã®ãƒ¡ãƒ¢ãƒª128GBã‚’LLMã§ã»ã¼ä½¿ã„åˆ‡ã£ã¦ã¿ãŸï¼
serihiro 2024/12/09
ãƒªãƒ³ã‚¯
æ©ãã‚‚ãƒ»ãƒ©ãƒ³ã‚‚ãƒ»ã‚ã—ãƒ©ãƒœ
è…°ç—›ã€ã²ã–ç—›ã€è‚©ã“ã‚Šã®æ…¢æ€§åŒ– Oè„šã€éª¨ç›¤ã®ã‚†ãŒã¿ã€å§¿å‹¢ã®å´©ã‚Œã€æ§˜ã€…ãªä½“ã®ä¸èª¿ï½¥ï½¥ï½¥ æ¹¿å¸ƒã‚„é›»æ°—æ²»ç™‚ã§ã‚ˆããªã£ã¦ã‚‚ ãã®æ™‚ã ã‘ã€‚ ã¨ã„ã†äº‹ã‚ã‚Šã¾ã›ã‚“ã‹ï¼Ÿ ç—›ã¿ã«ã¯ã€€å¿…ãšã€ŒåŽŸå› ã€ãŒã‚ã‚‹ã‚‚ã®ã§ã™ã€‚ é–“é•ã£ãŸèº«ä½“ã®ä½¿ã„æ–¹ã‚’ç¹°ã‚Šè¿”ã™ã“ã¨ã§ã€€ãã®éƒ¨ä½ãŒå£Šã‚Œ ã€Œçµæžœã€ç—›ã¿ãŒç”Ÿã˜ã¾ã™ã€‚ åŽŸå› ã‚’æ”¹å–„ã›ãšã«ã€€ç—›ã¿ãŒè§£æ¶ˆã•ã‚Œãªã„ã®ã¯ã€€å½“ãŸã‚Šå‰ã®è©±ãªã®ã§ã™ã€‚ æœ¬å½“ã«è¦‹ç›´ã•ãªã‘ã‚Œã°ãªã‚‰ãªã„ã®ã¯ ã€Œèº«ä½“ã®ãƒãƒ©ãƒ³ã‚¹ã€ã¨ã€Œæ©ãæ–¹ã€ã§ã™ã€‚ ã‚ã—ãƒ©ãƒœã§ã¯ è¶³ã¨æ©ãã®å°‚é–€å®¶ãŒ ç‹è‚‰ã‚„éª¨æ ¼ã®çŠ¶æ…‹ã€å‹•ãã‚’ãƒã‚§ãƒƒã‚¯ã€‚ å®Ÿéš›ã®æ©è¡Œã‚’ã¿ã¦ æ£ã—ã„èº«ä½“ã®ä½¿ã„æ–¹ã‚’å°Žãæ–½è¡“ã¨æŒ‡å°Žã‚’ã„ãŸã—ã¾ã™ã€‚ â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²â–½â–²
serihiro 2024/12/04
ãƒ©ãƒ³ãƒ‹ãƒ³ã‚°

æ•´ä½“
ãƒªãƒ³ã‚¯
1 2 3 4 5 6 7 8 9 10 æ¬¡ã®ãƒšãƒ¼ã‚¸