[B! model] manabouã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

manabou id:manabou

modelã«é–¢ã™ã‚‹manabouã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (14)

${{author_name}}$

{{{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}}

ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ã®ãƒˆãƒãƒã‚¸ã«ã¤ã„ã¦ãƒ¡ãƒ¢ | æ¸¬åº¦ã‚¼ãƒã®æŠ¹èŒ¶ãƒãƒ§ã‚³
ã¯ã˜ã‚ã« ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ã«ã¤ã„ã¦ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ãªã©ã‚’è¦‹ã¦ã„ã‚‹ã¨ã€Œãƒˆãƒãƒã‚¸ã‚’ãã‚Œã„ã«æ•´ãˆã‚‹ã€ã¨ã‹ ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ã¨ã„ã£ãŸè¡¨ç¾ã«å‡ºãã‚ã™ã“ã¨ãŒã‚ã‚Šã¾ã™ã€‚ ã“ã®ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ã¨ã„ã†ã®ã¯ã€ã©ã†ã‚„ã‚‰çµŒé¨“çš„ã«ç²å¾—ã•ã‚Œã‚‹æ„Ÿè¦šã®ã‚ˆã†ã§ã™ã€‚ ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã§ã‚‚ã€ã©ã†ãªã£ãŸã‚‰ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ãªã®ã‹ ã¨ã„ã†ã“ã¨ã‚’ãã¡ã‚“ã¨èª¬æ˜Žã—ã¦ã„ã‚‹ã‚‚ã®ã¯å°‘ãªã„ã§ã™ã€‚ ã“ã®ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ãŒä¸€ä½“ã©ã®ã‚ˆã†ãªã‚‚ã®ãªã®ã‹ã«ã¤ã„ã¦èª¿ã¹ã¦ã¿ã¾ã—ãŸãŒã€ æ®‹å¿µãªãŒã‚‰ã¯ã£ãã‚Šã¨ã—ãŸç”ãˆã¯æ‰‹ã«å…¥ã‚Œã‚‹ã“ã¨ãŒã§ãã¾ã›ã‚“ã§ã—ãŸã€‚ ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ãŒä½•è€…ãªã®ã‹ã«ã¤ã„ã¦æ˜Žç¢ºãªç”ãˆã¯æ‰‹ã«ã¯ã„ã‚Šã¾ã›ã‚“ã§ã—ãŸãŒã€ ã€Œãã‚Œã„ãªãƒˆãƒãƒã‚¸ã€ã«ã¤ã„ã¦èª¿ã¹ã‚‹éŽç¨‹ã§æ‰‹ã«å…¥ã£ãŸçŸ¥è˜ã‚’ã¨ã‚Šã‚ãˆãšãƒ¡ãƒ¢ã—ã¦ãŠãã¾ã™ã€‚ ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ç”¨èªžã¨ã—ã¦ã®ãƒˆãƒãƒã‚¸ ãã‚‚ãã‚‚ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ç•Œéšˆã§ä½¿ã‚ã‚Œã‚‹ã€Œãƒˆãƒãƒã‚¸ã€ã¨ã„ã†ç”¨èªžãŒç‰¹æ®Šã§ã™ã€‚ ä¸€èˆ¬çš„ãªç”¨èªžã¨ã—ã¦ã®ã€Œãƒˆãƒãƒã‚¸ã€ã¨
manabou 2021/07/01
modeling

model

topology
ãƒªãƒ³ã‚¯
AIã§1æžšã®äººç‰©å†™çœŸã‹ã‚‰é«˜ç²¾ç´°3Dãƒ¢ãƒ‡ãƒ«ä½œæˆã€€Facebookãªã©é–‹ç™º
Innovative Techï¼š ã“ã®ã‚³ãƒ¼ãƒŠãƒ¼ã§ã¯ã€ãƒ†ã‚¯ãƒŽãƒã‚¸ãƒ¼ã®æœ€æ–°ç ”ç©¶ã‚’ç´¹ä»‹ã™ã‚‹Webãƒ¡ãƒ‡ã‚£ã‚¢ã€ŒSeamlessã€ã‚’ä¸»å®°ã™ã‚‹å±±ä¸‹è£•æ¯…æ°ãŒåŸ·ç†ã€‚æ–°è¦æ€§ã®é«˜ã„ç§‘å¦è«–æ–‡ã‚’å±±ä¸‹æ°ãŒãƒ”ãƒƒã‚¯ã‚¢ãƒƒãƒ—ã—ã€è§£èª¬ã™ã‚‹ã€‚ å—ã‚«ãƒªãƒ•ã‚©ãƒ«ãƒ‹ã‚¢å¤§å¦ã€Facebook Reality Labsã€Facebook AI Researchã«ã‚ˆã‚‹ç ”ç©¶ãƒãƒ¼ãƒ ãŒé–‹ç™ºã—ãŸã€ŒPIFuHDã€ã¯ã€1æžšã®äººç‰©å†™çœŸã‹ã‚‰è¡£æœã‚’ç€ãŸäººä½“3æ¬¡å…ƒãƒ¢ãƒ‡ãƒ«ã‚’é«˜è§£åƒã«å†æ§‹ç¯‰ã™ã‚‹æ·±å±¤å¦ç¿’ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã ã€‚æŒ‡ã‚„é¡”ã®ç‰¹å¾´ã€è¡£æœã®ã²ã ã¨ã„ã£ãŸç´°ã‹ã„æƒ…å ±ã‚„ã€è¦‹ãˆã¦ã„ãªã„ç®‡æ‰€ã¾ã§3Dã§å†ç¾ã™ã‚‹ã€‚ äººç‰©ã®3Dãƒ¢ãƒ‡ãƒ«ã‚’å¿ å®Ÿã«ä½œæˆã™ã‚‹å ´åˆã€å¾“æ¥ã¯ã€ŒLight Stageã€ã¨å‘¼ã°ã‚Œã‚‹ã€å…¨æ–¹ä½ã«å¤šæ•°ã®ã‚«ãƒ¡ãƒ©ã‚„ã‚»ãƒ³ã‚µãƒ¼ã‚’é…ç½®ã—ãŸçƒå½¢ãƒ‰ãƒ¼ãƒ åž‹ã‚ãƒ£ãƒ—ãƒãƒ£ãƒ¼ã‚·ã‚¹ãƒ†ãƒ ãŒç”¨ã„ã‚‰ã‚Œã¦ããŸã€‚ã—ã‹ã—ã€Light Stageã¯é«˜ä¾¡ãªãŸã‚ä¸€èˆ¬çš„ã«å°Žå…¥ã¯é›£ã—ã„ã€‚ãã“ã§1æžšã®ç”»åƒã‹ã‚‰æ·±å±¤å¦ç¿’ã‚’ç”¨ã„
manabou 2020/06/18
ai

model

image

3d

3dmodel

deeplearning

colab
ãƒªãƒ³ã‚¯
ã€Œãƒ‡ãƒ¼ã‚¿æ´»ç”¨ã®ãŸã‚ã®æ•°ç†ãƒ¢ãƒ‡ãƒªãƒ³ã‚°å…¥é–€ã€ã§ã€ã‚¢ã‚¤ãƒ‡ã‚¢ã‚’æ•°å¼ã«è½ã¨ã—è¾¼ã‚€ - yag's blog
ãƒ‡ãƒ¼ã‚¿æ´»ç”¨ã®ãŸã‚ã®æ•°ç†ãƒ¢ãƒ‡ãƒªãƒ³ã‚°å…¥é–€ ä½œè€…:æ°´ä¸Š ã²ã‚ã,ç†Šè°· é›„ä»‹,é«˜é‡Ž é›…å…¸,è—¤åŽŸ æ™´é›„ç™ºå£²æ—¥: 2020/04/15ãƒ¡ãƒ‡ã‚£ã‚¢: å˜è¡Œæœ¬ï¼ˆã‚½ãƒ•ãƒˆã‚«ãƒãƒ¼ï¼‰ ã€Œãƒ‡ãƒ¼ã‚¿æ´»ç”¨ã®ãŸã‚ã®æ•°ç†ãƒ¢ãƒ‡ãƒªãƒ³ã‚°å…¥é–€ã€ã‚’ã”æµè´ˆã„ãŸã ãã¾ã—ãŸã®ã§ã€ã“ã®ãƒ–ãƒã‚°ã§ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ç§ã¯è‘—è€…ã®ä¸€äººã‹ã‚‰ãŠå£°ãŒã‘ã„ãŸã ãå‡ºç‰ˆå‰ã«åŽŸç¨¿ã‚’ãƒ¬ãƒ“ãƒ¥ãƒ¼ã•ã›ã¦ã„ãŸã ãæ©Ÿä¼šãŒã‚ã‚Šã¾ã—ãŸã®ã§ã€å®Ÿã¯ã‚ã¨ãŒãã«ã“ã£ãã‚Šåå‰ãŒè¼‰ã£ã¦ã„ãŸã‚Šã—ã¾ã™ã€‚ æœ¬æ›¸ã«ã¤ã„ã¦ ã“ã®æœ¬ã¯ã€ãƒ“ã‚¸ãƒã‚¹ã«ãŠã‘ã‚‹æ§˜ã€…ãªèª²é¡Œã‚’æ•°ç†ãƒ¢ãƒ‡ãƒªãƒ³ã‚°ã‚’é€šã˜ã¦è§£æ±ºã™ã‚‹ãŸã‚ã®å…¥é–€æ›¸ã§ã™ã€‚ãƒ“ãƒƒã‚°ãƒ‡ãƒ¼ã‚¿ã‹ã‚‰äººå·¥çŸ¥èƒ½/AIã¸ã¨è¨€è‘‰ã‚’å¤‰ãˆã¤ã¤ã‚‚ã€ä¸–ã®ä¸ã®æµè¡Œã¨ã—ã¦æ³¨ç›®ã•ã‚Œç¶šã‘ã‚‹ä¼æ¥ã®ãƒ‡ãƒ¼ã‚¿æ´»ç”¨ã§ã™ãŒã€ã©ã†ã„ã£ãŸãƒ‡ãƒ¼ã‚¿ã«å¯¾ã—ã¦ä½•ã‚’ã™ã‚‹ã‹ã¨ã„ã†æ´»ç”¨ã®å®Ÿæ…‹ã¯å¤šç¨®å¤šæ§˜ã§ã™ã€‚ECã‚µã‚¤ãƒˆã«ãŠã‘ã‚‹å•†å“ã®ãƒ¬ã‚³ãƒ¡ãƒ³ãƒ‰ã«å§‹ã¾ã‚Šã€åºƒå‘Šé…ä¿¡ã®é…ä¿¡ã®æœ€é©åŒ–ã‚„æ¥å‹™åŠ¹çŽ‡åŒ–ã€ã¾ãŸã¯ç¤¾å†…ã«çœ ã£ã¦ã„ã‚‹æƒ…å ±ã‚’ç´ æ—©ãã‚¢ã‚¯ã‚»ã‚¹ã§
manabou 2020/04/20
data

model

book

math
ãƒªãƒ³ã‚¯
How to Deploy Machine Learning Models
Introduction The deployment of machine learning models is the process for making your models available in production environments, where they can provide predictions to other software systems. It is only once models are deployed to production that they start adding value, making deployment a crucial step. However, there is complexity in the deployment of machine learning models. This post aims to
manabou 2019/06/13
machinelearning

deployment

model
ãƒªãƒ³ã‚¯
BERTè§£èª¬ï¼šè‡ªç„¶è¨€èªžå‡¦ç†ã®ãŸã‚ã®æœ€å…ˆç«¯è¨€èªžãƒ¢ãƒ‡ãƒ« | AIå°‚é–€ãƒ‹ãƒ¥ãƒ¼ã‚¹ãƒ¡ãƒ‡ã‚£ã‚¢ AINOW
è‘—è€…ã®Rani Horevæ°ã¯ã‚¤ã‚¹ãƒ©ã‚¨ãƒ«å‡ºèº«ã§ã€ã‚¤ã‚¹ãƒ©ã‚¨ãƒ«å›½é˜²çœã®ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹ãƒ»ãƒãƒ¼ãƒ ãƒªãƒ¼ãƒ€ãƒ¼ã‚’å‹™ã‚ãŸã“ã¨ã‚‚ã‚ã‚‹AIç ”ç©¶è€…ã§ã™ã€‚åŒæ°ãŒè‹±èªžé•·æ–‡è¨˜äº‹ãƒ¡ãƒ‡ã‚£ã‚¢Mediumã«æŠ•ç¨¿ã—ãŸè¨˜äº‹ã§ã¯ã€GoogleãŒ2018å¹´10æœˆã«ç™ºè¡¨ã—ãŸè‡ªç„¶è¨€èªžå‡¦ç†ãƒ¢ãƒ‡ãƒ«ã§ã‚ã‚‹BERTã®ç‰¹å¾´ãŒè§£èª¬ã•ã‚Œã¦ã„ã¾ã™ã€‚ Google AI Languageãƒãƒ¼ãƒ ãŒ2018å¹´10æœˆã«ç™ºè¡¨ã•ã‚ŒãŸBERTã¯ã€è³ªç–‘å¿œç”ã‚„è‡ªç„¶è¨€èªžæŽ¨è«–ã¨ã„ã£ãŸæ§˜ã€…ãªè‡ªç„¶è¨€èªžå‡¦ç†ã®ã‚¿ã‚¹ã‚¯ã«ãŠã„ã¦å…ˆè¡Œã™ã‚‹è¨€èªžãƒ¢ãƒ‡ãƒ«ã‚’å‡Œé§•ã™ã‚‹æ€§èƒ½ã‚’å®Ÿç¾ã—ã¾ã—ãŸï¼ˆBERTã®ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ãƒ†ã‚¹ãƒˆã«é–¢ã™ã‚‹è©³ç´°ã¯æœ¬è¨˜äº‹æœ«å°¾ã®ï¼ˆâ€»è¨»7ï¼‰ã‚’å‚ç…§ï¼‰ã€‚è¿‘å¹´ã®è¨€èªžAIãƒ¢ãƒ‡ãƒ«ã®ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ã«ãªã£ã¦ã„ã‚‹ã‚¿ã‚¹ã‚¯ã¯ã€æœªçŸ¥ã®æ–‡å—åˆ—ã«å¯¾ã—ã¦é©åˆ‡ãªé–¢ä¿‚ã«ã‚ã‚‹å˜èªžã‚ã‚‹ã„ã¯æ–‡å—åˆ—ã‚’äºˆæ¸¬ã™ã‚‹ã¨ã„ã†äºˆæ¸¬å•é¡Œã‚’è§£ãã“ã¨ã¨åŒç¾©ã¨ãªã‚Šã¾ã™ã€‚ä¾‹ãˆã°è³ªç–‘å¿œç”ã‚¿ã‚¹ã‚¯ã«ãŠã„ã¦ã¯ã€å—ã‚«ãƒªãƒ•ã‚©ãƒ«ãƒ‹ã‚¢ã®ç•¥ç§°ã‚’å•ã†è³ªå•ã‚’æ„å‘³ã™ã‚‹æ–‡å—
manabou 2019/05/21
bert

nlp

model
ãƒªãƒ³ã‚¯
500å††ã§è‡ªä½œã§ãã‚‹æœ¬æ ¼ãƒžãƒ†ãƒªã‚¢ãƒ«ã‚¹ã‚ãƒ£ãƒŠï¼ - äººæŸ±ç³»CGãƒ¢ãƒ‡ãƒ©ãƒ¼ã®Tipsãƒ–ãƒã‚°
çš†æ§˜ã“ã‚“ã«ã¡ã¯ã€‚ ã”ç„¡æ²™æ±°ã—ã¦ãŠã‚Šã¾ã™ã€ã¾ã™ãã§ã™ã€‚ æ–°ã—ã„ä»•äº‹ã‚‚ãªã‹ãªã‹ã«ãƒãƒ¼ãƒ‰ã§ã€å®¶ã«å¸°ã‚Œãªã‹ã£ãŸã‚Šã€è·å ´ã«æ³Šã¾ã£ãŸã‚Šã€å¸°å®…ã§ãã¦ã‚‚ãƒã‚¿ãƒ³Qã§ãªã‹ãªã‹è¨˜äº‹ãŒæ›´æ–°ã§ããªã‹ã£ãŸã‚Šã¨â€¦é–¢ä¿‚å„æ‰€ã«ã”è¿·æƒ‘ã‚’ãŠã‹ã‘ã—ã¦ãŠã‚Šã¾ã™orz ã ã„ã¶å‰ã‹ã‚‰ä¼ç”»ã—ã¦ã„ãŸã€Žèª°ã«ã§ã‚‚å®‰ä¾¡ã«ä½œã‚Œã‚‹æœ¬æ ¼ãƒžãƒ†ãƒªã‚¢ãƒ«ã‚¹ã‚ãƒ£ãƒŠãƒ¼ã€ä¼ç”»ã§ã™ãŒã€æ˜¨å¹´æœ«ã®å‘ŠçŸ¥ã‹ã‚‰å¤§å¤‰é•·ã‚‰ããŠã¾ãŸã›ã—ã¦ã—ã¾ã„ã¾ã—ãŸã€‚ã‚ˆã†ã‚„ãå¿µé¡˜å¶ã£ã¦ãã®å…¨å®¹ã‚’ç´¹ä»‹ã‚’ã—ã¦å‚ã‚ŠãŸã„ã¨æ€ã„ã¾ã™ã€‚ from Substance Designer ã“ã‚Œã¯ã€100å‡ã®ç´ æã§ä½œæˆã—ãŸè£…ç½®ã¨ã€æ™®é€šã®ãƒŸãƒ©ãƒ¼ãƒ¬ã‚¹ï¼ˆSony Î±7rm2ï¼‰ã¨æ™®é€šã®ãƒžã‚¯ãƒãƒ¬ãƒ³ã‚ºï¼ˆSEL90M28Gï¼‰ã§èº«è¿‘ãªå¸ƒã‚’æ’®å½±ã—ã€æ™®é€šã®Lightroomã§å†™çœŸã‚’åˆ‡ã‚Šå‡ºã—ã€æ™®é€šã®SubstanceDesignerã§è³ªæ„Ÿå†ç¾ã‚’ã—ãŸã‚‚ã®ã§ã™ã€‚ è‡ªåˆ†ãŸã¡ã§ã‚‚é©šãã»ã©è‰¯ãã‚¹ã‚ãƒ£ãƒ³ã§ããŸã¨æ€ã„ã¾ã™ã€‚ ã‚‚ã¡ã‚ã‚“ã€å¿…è¦ãªé“å…·ã•ãˆã‚
manabou 2019/03/12
scanner

material

image

diy

hardware

camera

modeling

model

3d
ãƒªãƒ³ã‚¯
OpenAIã€ã¾ã“ã¨ã—ã‚„ã‹ãªãƒ•ã‚§ã‚¤ã‚¯ãƒ‹ãƒ¥ãƒ¼ã‚¹ã‚‚ç°¡å˜ç”Ÿæˆã®è¨€èªžãƒ¢ãƒ‡ãƒ«ã€ŒGPT-2ã€ã®é™å®šç‰ˆã‚’ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹åŒ–
OpenAIã€ã¾ã“ã¨ã—ã‚„ã‹ãªãƒ•ã‚§ã‚¤ã‚¯ãƒ‹ãƒ¥ãƒ¼ã‚¹ã‚‚ç°¡å˜ç”Ÿæˆã®è¨€èªžãƒ¢ãƒ‡ãƒ«ã€ŒGPT-2ã€ã®é™å®šç‰ˆã‚’ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹åŒ– ã‚¤ãƒ¼ãƒãƒ³ãƒ»ãƒžã‚¹ã‚¯æ°ãŒå…±åŒä¼šé•·ã‚’å‹™ã‚ã‚‹éžå–¶åˆ©ã®ç±³AIï¼ˆäººå·¥çŸ¥èƒ½ï¼‰ç ”ç©¶ä¼æ¥OpenAIã¯2æœˆ14æ—¥ï¼ˆç¾åœ°æ™‚é–“ï¼‰ã€è‡ªç„¶è¨€èªžã®æ–‡ç« ã‚’ç”Ÿæˆã™ã‚‹è¨€èªžãƒ¢ãƒ‡ãƒ«ã€ŒGPT-2ã€ã‚’ç™ºè¡¨ã—ãŸã€‚ ã“ã®æŠ€è¡“ãŒæ‚ªæ„ã‚ã‚‹ã‚¢ãƒ—ãƒªã«æµç”¨ã•ã‚Œã‚‹æ‡¸å¿µã‹ã‚‰ã€GPT-2ã‚’ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹åŒ–ã¯ã›ãšã€ç¸®å°ç‰ˆãƒ¢ãƒ‡ãƒ«ã¨è«–æ–‡ã®ã¿å…¬é–‹ã—ãŸã€‚ GPT-2ã¯ã€40Gãƒã‚¤ãƒˆåˆ†ã®ãƒãƒƒãƒˆä¸Šã®ãƒ†ã‚ã‚¹ãƒˆã®æ¬¡ã®å˜èªžã‚’äºˆæ¸¬ã™ã‚‹ãŸã‚ã ã‘ã«è¨“ç·´ã•ã‚ŒãŸTransf ormerãƒ™ãƒ¼ã‚¹ã®è¨€èªžãƒ¢ãƒ‡ãƒ«ã€‚800ä¸‡ã®Webãƒšãƒ¼ã‚¸ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã§è¨“ç·´ã—ã€15å„„ã®ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ã‚’æŒã¤ã€‚ å¤§ããªç‰¹å¾´ã¯ã€å›ºæœ‰ã®è¨“ç·´ç”¨ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã§ã¯ãªãã€Webãƒšãƒ¼ã‚¸ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã‚’ãã®ã¾ã¾ä½¿ã£ãŸã“ã¨ã ã€‚Wikipediaã‚„ãƒ‹ãƒ¥ãƒ¼ã‚¹ã€æ›¸ç±ãªã©ç‰¹å®šã®ãƒ‡ãƒ¼ã‚¿ã§è¨“ç·´ã•ã‚ŒãŸä»–ã®è¨€èªžãƒ¢ãƒ‡ãƒ«ï¼ˆç±³Googleã®BERTãª
manabou 2019/02/15
openai

gpt2

fakenews

nlp

model

language
ãƒªãƒ³ã‚¯
2018å¹´ã®è¨€èªžãƒ¢ãƒ‡ãƒ«æ¦‚è¦
Tung Nguyen2019-02-072013å¹´ã«Deep Belief Networkã«ã¤ã„ã¦åˆã‚ã¦èªã‚“ã ã¨ãã€ç§ã¯ãƒ‡ã‚£ãƒ¼ãƒ—ãƒ©ãƒ¼ãƒ‹ãƒ³ã‚°(æ·±å±¤å¦ç¿’)ã«ã¯ã¾ã‚Šã¾ã—ãŸã€‚ç¾åœ¨ã€Clovaã®è‡ªç„¶è¨€èªžå‡¦ç†ã«ãŠã‘ã‚‹æ·±å±¤å¦ç¿’ã‚’ç ”ç©¶é–‹ç™ºã—ã¦ãŠã‚Šã¾ã™ã€‚ ã“ã‚“ã«ã¡ã¯ã€Clovaãƒãƒ¼ãƒ ã®Tungã§ã™ã€‚Â Clovaã¯ã€Clova Friendsã‚„Clova Waveãªã©ã¨ã„ã£ãŸã‚¹ãƒžãƒ¼ãƒˆãƒ‡ãƒã‚¤ã‚¹ã«æè¼‰ã•ã‚Œã¦ã„ã‚‹ç§ãŸã¡ã®AIãƒ—ãƒ©ãƒƒãƒˆãƒ•ã‚©ãƒ¼ãƒ ã§ã™ã€‚Â è£½å“ã®è©³ç´°ã«ã¤ã„ã¦ã¯ã“ã¡ã‚‰ã‚’ã”è¦§ãã ã•ã„ã€‚ 2018å¹´ã®è‡ªç„¶è¨€èªžå‡¦ç†ï¼ˆNLPï¼‰åˆ†é‡Žã«ãŠã„ã¦ç¶šã€…ã¨ç™ºè¡¨ã•ã‚ŒãŸå¼·åŠ›ãªè¨€èªžãƒ¢ãƒ‡ãƒ« -Â ELMoã€ULMFitÂ ã€OpenAI Transf ormerã€BERTã«ã¤ã„ã¦æŒ¯ã‚Šè¿”ã‚ŠãŸã„ã¨æ€ã„ã¾ã™ã€‚Â å„ãƒ¢ãƒ‡ãƒ«ãã‚Œãžã‚Œã®å†…éƒ¨è§£èª¬ã¯ã‚¤ãƒ³ã‚¿ãƒ¼ãƒãƒƒãƒˆä¸Šã«ã‚ã£ãŸã‚Šã—ã¾ã™ãŒã€ä»Šå›žã¯å°‘ã—ç•°ãªã£ãŸè¦‹æ–¹ã‚’æŒã£ã¦æ¥ã‚‹ã“ã¨ã‚’è©¦ã¿ã¾ã™ã€‚å…¨ãƒ¢ãƒ‡ãƒ«ã®ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£
manabou 2019/02/08
nlp

language

model

machinelearning
ãƒªãƒ³ã‚¯
[R] äºˆæ¸¬ãƒ¢ãƒ‡ãƒ«ã‚’ä½œã‚‹ã«ã¯ formula ã‚’æ´»ç”¨ã›ã‚ˆ - ill-identified diary
æ¦‚è¦ formula ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã¯å¤‰æ•°å¤‰æ›ã‚„äº¤äº’ä½œç”¨é …ãªã©, å¤šå½©ãªè¡¨ç¾ãŒã§ãã‚‹. xgboost ã‚„ glmnet ã§ã¯ model.matrix() ã‚’ä½µç”¨ã™ã‚‹ã“ã¨ã§ formula ã‚’åˆ©ç”¨ã§ãã‚‹. çµ±è¨ˆãƒ¢ãƒ‡ãƒªãƒ³ã‚°/æ©Ÿæ¢°å¦ç¿’ã§äºˆæ¸¬ãƒ¢ãƒ‡ãƒ«ã‚’æ§‹ç¯‰ã™ã‚‹ã¨ã, äºˆæ¸¬æ€§èƒ½ã®å‘ä¸Šã®ãŸã‚, ã—ã°ã—ã°å¤‰æ•°ã‚’å…¥ã‚Œæ›¿ãˆãŸã‚Š, å¤‰æ›ã—ãŸã‚Šã¨ã„ã£ãŸæŽ¨æ•²ãŒå¿…è¦ã¨ãªã‚‹. R ã§ã¯ã“ã†ã„ã†ã¨ãã« formula ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã‚’ä½¿ã†ã¨, ã„ã¡ã„ã¡ãƒ‡ãƒ¼ã‚¿ãƒ•ãƒ¬ãƒ¼ãƒ ã«å¤‰æ›å¾Œã®æ•°å€¤ã‚’è¿½åŠ ã—ãŸã‚Šæ›¸ãæ›ãˆãŸã‚Šã™ã‚‹å¿…è¦ãŒãªããªã‚‹. glmnet ã‚„ xgboost ãªã©, formula ãŒç›´æŽ¥ä½¿ãˆãªã„ã‚‚ã®ã§ã‚‚ model.matrix() ç‰ã‚’ä½µç”¨ã™ã‚Œã°å¯èƒ½ã§ã‚ã‚‹*1. formula ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã‚’è§£èª¬ã—ãŸè¨˜äº‹ã‚’æŽ¢ã™ã¨, ã‹ãªã‚Šå‰ã‹ã‚‰å˜åœ¨ã™ã‚‹. ä¾‹ãˆã°ä»¥ä¸‹ã®è¨˜äº‹. m884.hateblo.jp ãªãŠ, ä¸Šè¨˜ã¯ã‚¿ã‚¤ãƒˆãƒ«ãŒã€Œf
manabou 2017/05/02
r

model

formula
ãƒªãƒ³ã‚¯
AWS ã¯å½¢å¼æ‰‹æ³•ã®å¤¢ã‚’è¦‹ã‚‹ã‹ï¼Ÿ - ãƒ¢ãƒ‡ãƒ«æ¤œæŸ»å™¨ Alloy ã«ã‚ˆã‚‹ã‚¤ãƒ³ãƒ•ãƒ©è¨è¨ˆ
NGK2015B ã§ä½¿ç”¨ã—ãŸã‚¹ãƒ©ã‚¤ãƒ‰ã§ã™ã€‚ãƒ¢ãƒ‡ãƒ«æ¤œæŸ»å™¨ Alloy ã‚’ç”¨ã„ã¦ã€AWS ã®ã‚»ã‚ãƒ¥ãƒªãƒ†ã‚£ã‚°ãƒ«ãƒ¼ãƒ—è¨å®šã‚’è‡ªå‹•ã§æ¤œæŸ»ã—ã¾ã™ã€‚
manabou 2015/12/10
aws

alloy

model

infrastructure
ãƒªãƒ³ã‚¯
Rails ã§ fat model ã‚’é¿ã‘ã‚‹ãŸã‚ã®ã€ã‚ã¾ã‚ŠçŸ¥ã‚‰ã‚Œã¦ã„ãªã„æ–¹æ³•ã«ã¤ã„ã¦ - ãŠã‚‚ã—ã‚webã‚µãƒ¼ãƒ“ã‚¹é–‹ç™ºæ—¥è¨˜
ã“ã®ã‚¨ãƒ³ãƒˆãƒªã§æ›¸ã„ãŸå†…å®¹ã¯ã€ã»ã¼ Growing Rails Applications in Practice ã®å†…å®¹ãŒå…ƒã«ãªã£ã¦ã„ã¾ã™ã€‚è‹±èªžã§ã™ãŒã€ã“ã“ã§æŒ™ã’ãŸå†…å®¹ä»¥å¤–ã«ã‚‚ã‚³ãƒ¼ãƒ‰ã‚’ç¶ºéº—ã«ä¿ã¤ãƒ†ã‚¯ãƒ‹ãƒƒã‚¯ãŒæ›¸ã‹ã‚Œã¦ãŠã‚Šã€ã‹ã¤ãƒšãƒ¼ã‚¸æ•°ã‚‚å°‘ãªãèªã¿ã‚„ã™ã„ã§ã™ã€‚ã‚³ãƒ¼ãƒ‰ã‚’ç¶ºéº—ã«ä¿ã¤ã®ãŒå¥½ããªæ–¹ã¯ä¸€èªã—ã¦ã¿ã‚‹ã“ã¨ã‚’ãŠã™ã™ã‚ã—ã¾ã™ã€‚ ã¯ã˜ã‚ã« Rails ã§ fat model ã‚’é¿ã‘ã‚‹ãŸã‚ã®æ–¹æ³•ã¯ã€7 Patterns to Refactor Fat ActiveRecordÂ Models ã‚’å§‹ã‚ã¨ã—ã¦ã€å¤šãã®ã‚„ã‚Šæ–¹ãŒå˜åœ¨ã—ã¾ã™*1ã€‚ validation ã‚„ callback ã¯ ActiveRecord(ä»¥ä¸‹AR) ã‚’ç¶™æ‰¿ã›ãšã¨ã‚‚åˆ©ç”¨ã™ã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚7 Patterns to Refactor Fat ActiveRecordÂ Models ã® ã€Œ3. Extract Form Objects
manabou 2015/08/25
rails

model

callback

activerecord

ruby
ãƒªãƒ³ã‚¯
MySQLãƒ»PostgreSQLãƒ¦ãƒ¼ã‚¶ãƒ¼ã‚°ãƒ«ãƒ¼ãƒ—ï¼ˆMyNAãƒ»JPUGï¼‰åˆåŒDBå‹‰å¼·ä¼šã§ç™ºè¡¨ã—ãŸè³‡æ–™ã‚’å…¬é–‹ã—ã¾ã—ãŸã€‚ã€Œãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«ã«ã¤ã„ã¦çŸ¥ã£ã¦ãŠãã¹ã7ã¤ã®ã“ã¨ ã€œNoSQLã«æ‰‹ã‚’å‡ºã™å‰ã«ã€œã€
è¡¨é¡Œã®é€šã‚Šã€MyNAã¨JPUGã®åˆåŒDBå‹‰å¼·ä¼šã§ç™ºè¡¨ã‚’ã—ãŸã®ã§è³‡æ–™ã‚’å…¬é–‹ã—ãŸã€‚ å†…å®¹ã®è©³ç´°ã¯ã‚¹ãƒ©ã‚¤ãƒ‰ãã®ã‚‚ã®ã‚’è¦‹ã¦ã„ãŸã ãã¨ã—ã¦ã€è¨€ã„ãŸã„ã“ã¨ã®ä¸»æ—¨ã¯ã“ã†ã§ã‚ã‚‹ã€‚ä¸–ã®ä¸ã«å®Œç’§ãªãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«ã¯ãªã„ã®ã§ã€NoSQLã¯å½“ç„¶ã®å¦‚ãå¿…è¦ã«ãªã‚‹ã€‚ã ãŒã€ä½•ã§ã‚‚ã‹ã‚“ã§ã‚‚NoSQLã‚’ä½¿ãˆã°è‰¯ã„ã¨ã„ã†ã‚‚ã®ã§ã¯ãªã„ã€‚ã‚€ã—ã‚ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ãŒå¿…è¦ã¨ã—ã¦ã„ã‚‹ãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«ãŒä½•ã‹ã¨ã„ã†ã“ã¨ã‚’ã‚ˆãç†è§£ã—ã€æœ¬å½“ã«å¿…è¦ãªå ´åˆã«ã“ãã€NoSQLã‚’ä½¿ã†ã¹ããªã®ã§ã‚ã‚‹ã€‚ã¤ã¾ã‚Šã€Œã”åˆ©ç”¨ã¯è¨ˆç”»çš„ã«ï¼ã€ã¨ã„ã†ã“ã¨ã ã€‚ å¤§åˆ‡ãªã®ã¯ã€æ§˜ã€…ãªãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«ã‚’ç†è§£ã—ã€ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã«ã¨ã£ã¦ãƒ™ã‚¹ãƒˆãªè£½å“ã‚’é¸æŠžã™ã‚‹ã¨ã„ã†ã“ã¨ã ã€‚ãƒ™ã‚¹ãƒˆãªã®ãŒRDBã‹ã‚‚çŸ¥ã‚Œãªã„ã—ã€ãã†ã§ãªã„ã‹ã‚‚çŸ¥ã‚Œãªã„ã€‚æœ€é©ãªãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«ã‚’é¸æŠžã—ãŸå ´åˆã«ã€å‡ºæ¥ä¸ŠãŒã£ãŸã‚‚ã®ã®æ€§èƒ½ã‚‚æœ€é«˜ã«ãªã‚‹ã—ã€é–‹ç™ºåŠ¹çŽ‡ã‚‚æœ€ã‚‚è‰¯ããªã‚‹ã€‚ãƒ‡ãƒ¼ã‚¿ãƒ™ãƒ¼ã‚¹ã®ä¸»æµã¯RDBã ãŒã€ãã‚Œã¯ãƒªãƒ¬ãƒ¼ã‚·ãƒ§ãƒŠãƒ«ãƒ¢ãƒ‡ãƒ«ãŒã‚«ãƒãƒ¼ã§
manabou 2015/07/08
mysql

slide

database

model

datamodel
ãƒªãƒ³ã‚¯
ã‚ãªãŸãŒçŸ¥ã‚‰ãªã„ ãƒªãƒ¬ãƒ¼ã‚·ãƒ§ãƒŠãƒ«ãƒ¢ãƒ‡ãƒ«
Luaæ–‡åŒ–ã®ä¼æ‰¿!? WFSã«ãŠã‘ã‚‹ã‚¤ãƒ™ãƒ³ãƒˆã‚¹ã‚¯ãƒªãƒ—ãƒˆæ´»ç”¨è¡“ã€œã™ã¹ã¦ã¯ã‚ˆã‚Šè‰¯ã„ã‚³ãƒ³ãƒ†ãƒ³ãƒ„åˆ¶ä½œã®ãŸã‚ã«ã€œ
manabou 2014/11/14
database

slide

sql

model

mysql
ãƒªãƒ³ã‚¯
ã‚¤ãƒŸãƒ¥ãƒ¼ã‚¿ãƒ–ãƒ«ãƒ‡ãƒ¼ã‚¿ãƒ¢ãƒ‡ãƒ«(å…¥é–€ç·¨)
ãƒãƒ³ãƒ‰ãƒŽãƒ¼ãƒˆ Tå—å½¢ERãƒ¢ãƒ‡ãƒ«ã€€ã‚»ãƒŸãƒŠãƒ¼è³‡æ–™ã€€(Author; S.Toriyabe SYSTEMS DESIGN Co.,Ltd. Japan)
manabou 2014/10/21
immutable

data

database

CRUD

model

datamodel

programming
ãƒªãƒ³ã‚¯
1