Posted 杩涘嚮鐨凜oder

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了相关的知识,希望对你有一定的参考价值。

缂栬€呮寜锛欻uggingface绉戠爺璐熻矗浜篢homas Wolf绠€瑕佷粙缁嶄簡鏅€傝瘝宓屽叆銆佸彞宓屽叆鐨勬渶鏂版柟娉曘€?/span>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=鍥剧墖鏉ユ簮锛欽ohn Christian Fjellestad

璇嶅祵鍏ワ紙word embeddings锛?/span>鍜?span>鍙ュ祵鍏ワ紙sentence embeddings锛?/span>宸茬粡鎴愪负浠讳綍鍩轰簬娣卞害瀛︿範鐨勮嚜鐒惰瑷€澶勭悊绯荤粺涓嶅彲鎴栫己鐨勯儴鍒嗐€?/p>

璇嶅祵鍏ュ拰鍙ュ祵鍏ュ皢鍗曡瘝鍜屽彞瀛?馃摐 缂栫爜涓哄浐瀹氶暱搴︾殑瀵嗛泦鍚戦噺馃搻 锛屾垙鍓ф€у湴鏀瑰杽鏂囨湰鏁版嵁鐨勫鐞嗐€?/p>

涓€涓儹闂ㄧ殑瓒嬪娍鏄?span>鏅€傚祵鍏ワ紙Universal Embeddings锛?/span>锛氬湪澶у瀷璇枡搴撲笂棰勮缁冪殑宓屽叆锛屽彲浠ユ彃鍏ヨ澶氫笅娓哥殑浠诲姟妯″瀷锛堟儏鎰熷垎鏋愩€佸垎绫汇€佺炕璇戔€︹€︼級锛岄€氳繃骞跺叆鍦ㄨ緝澶ц妯$殑鏁版嵁闆嗕笂瀛︿範鍒扮殑璇?鍙ヨ〃绀鸿嚜鍔ㄦ敼鍠勬ā鍨嬭〃鐜般€?/p>

杩欐槸涓€绉?span>杩佺Щ瀛︿範锛坱ransfer learning锛?/span>鐨勫舰寮忋€傛渶杩戯紝杩佺Щ瀛︿範鍦ㄤ竴浜涢噸瑕佺殑浠诲姟锛堟瘮濡傛枃鏈垎绫伙級涓婃垙鍓ф€у湴鎻愬崌浜哊LP妯″瀷鐨勮〃鐜般€傝鍙傞槄Jeremy Howard鍜孲ebastian Ruder鎻愬嚭鐨刄LMFiT妯″瀷銆?/p>

灏界闀挎湡浠ユ潵鍙ュ瓙鐨勬棤鐩戠潱琛ㄧず瀛︿範鏄富娴侊紝鏈€杩戝嚑涓湀锛?017骞存湯/2018骞村垵锛夛紝鎴戜滑鐪嬪埌浜嗚澶氶潪甯告湁瓒g殑宸ヤ綔锛屾樉绀轰簡鍚戠洃鐫e涔犲拰澶氫换鍔″涔犺浆鍚戠殑瓒嬪娍銆?/p>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=鏈€杩戠殑鏅€傝瘝/鍙ュ祵鍏ョ爺绌?/p>

鏈枃瀵瑰綋鍓嶆渶鍏堣繘鐨勬櫘閫傝瘝/鍙ュ祵鍏ュ仛浜嗕竴涓畝鐭殑鍒濇鎬荤粨銆傛垜浠皢鎻忚堪涓婅〃涓姞绮楃殑妯″瀷锛?/p>

  • 寮哄姏/杩呴€熺殑鍩虹嚎锛欶astText銆佽瘝琚嬶紙Bag-of-Words锛?/p>

  • 褰撳墠鏈€鍏堣繘妯″瀷锛欵LMo銆丼kip-Thoughts銆丵uick-Thoughts銆?InferSent銆丮ILA/MSR鐨凣eneral Purpose Sentence Representations銆丟oogle鐨刄niversal Sentence Encoder

濡傛灉浣犲笇鏈涗簡瑙d竴浜涜瘝宓屽叆鍦?017骞翠箣鍓嶇殑杩涘睍锛屽彲浠ュ弬鑰僑ebastian Ruder鍐欑殑Word embeddings in 2017: Trends and future directions銆俁uder杩樺啓杩囦竴绡囦粙缁嶈瘝宓屽叆鎶€鏈殑鏂囩珷On word embeddings銆?/p>

璁╂垜浠粠璇嶅祵鍏ュ紑濮嬨€?/p>

璇嶅祵鍏ユ渶杩戠殑杩涘睍

杩戜簲骞存潵鎻愬嚭浜嗗ぇ閲忚瘝宓屽叆鏂规硶銆傚叾涓渶甯哥敤鐨勬ā鍨嬫槸word2vec鍜孏loVe锛岃繖涓や釜妯″瀷閮芥槸鍩轰簬鍒嗗竷鍋囪锛坉istributional hypothesis锛?/span>鐨勬棤鐩戠潱鏂规硶銆傦紙鏍规嵁鍒嗗竷鍋囪锛屽嚭鐜板湪鐩稿悓涓婁笅鏂囦腑鐨勫崟璇嶅€惧悜浜庡叿鏈夌浉浼肩殑鍚箟锛夈€?br>

灏界鏈変竴浜涘伐浣滈€氳繃骞跺叆璇箟鎴栬娉曠煡璇嗙瓑澧炲己杩欎簺鏃犵洃鐫f柟娉曪紝绾棤鐩戠潱鏂规硶鍦?017-2018骞存湡闂村彇寰椾簡鏈夎叮鐨勮繘灞曪紝鍏朵腑鏈€閲嶅ぇ鐨勬槸FastText锛坵ord2vec鐨勬墿灞曪級鍜?span>ELMo锛堝綋鍓嶆渶鍏堣繘鐨勪笂涓嬫枃璇嶅悜閲忥級銆?/p>

FastText鐢盩omas Mikolov鍥㈤槦寮€鍙戯紝杩欎竴鍥㈤槦姝f槸鍦?013骞存彁鍑簑ord2vec鐨勫洟闃熴€?/p>

FastText鐨勪富瑕佹敼杩涙槸鍖呭惈浜嗗瓧绗︾殑n鍏冭娉曪紝浠庤€屽彲浠ヤ负璁粌鏁版嵁涓病鏈夊嚭鐜?/span>鐨勫崟璇嶈绠楄瘝琛ㄧず銆?/p>

FastText鍚戦噺璁粌鏋佷负杩呴€燂紝鍚屾椂鎻愪緵浜嗗熀浜庣淮鍩虹櫨绉戝拰Crawl璁粌鐨?57绉嶈瑷€鐨勯璁粌璇嶅悜閲忊€斺€旇繖鏄緢妫掔殑鍩虹嚎銆?/p>

ELMo锛堟繁搴︿笂涓嬫枃璇嶈〃绀猴級鏈€杩戝皢璇嶅祵鍏ョ殑鏈€浣宠〃鐜版彁鍗囦簡涓嶅皯銆傚畠鏄敱AI2寮€鍙戠殑锛屽苟鍦∟AACL 2018涓婅璇勪负鏈€浣宠鏂囥€?/p>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=Elmo锛堣姖楹昏瑙掕壊锛?/p>

鍦‥LMo涓紝宓屽叆鍩轰簬涓€涓?span>鍙屽眰鐨勫弻鍚戣瑷€妯″瀷锛坆iLM锛?/span>鐨勫唴閮ㄧ姸鎬?/span>璁$畻锛孍LMo涔熸槸鍥犳寰楀悕鐨勶細Embeddings from Language Models锛堟潵鑷瑷€妯″瀷鐨勫祵鍏ワ級銆?/p>

ELMo鐨勭壒鎬э細

  • ELMo鐨勮緭鍏ユ槸瀛楃鑰屼笉鏄崟璇嶃€傝繖浣垮緱瀹冨彲浠ュ埄鐢ㄥ瓙瀛楋紙sub-word锛夊崟鍏冧负璇嶆眹琛ㄤ互澶栫殑鍗曡瘝璁$畻鏈夋剰涔夌殑琛ㄧず锛堝拰FastText绫讳技锛夈€?/p>

  • ELMo鏄痓iLM鐨勫灞傛縺娲荤殑杩炴帴锛坈oncatenation锛?/span>銆傝瑷€妯″瀷鐨勪笉鍚屽眰缂栫爜浜嗗崟璇嶇殑涓嶅悓淇℃伅銆傝繛鎺ユ墍鏈夊眰浣垮緱ELMo鍙互缁勫悎澶氱璇嶈〃绀猴紝浠ユ彁鍗囦笅娓镐换鍔$殑琛ㄧ幇銆?/p>

濂戒簡锛岀幇鍦紝璁╂垜浠湅鐪嬫櫘閫傚彞宓屽叆銆?/p>

鏅€傚彞宓屽叆鐨勫叴璧?/span>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=鐩墠鏈夊緢澶氱浉浜掔珵浜夌殑瀛︿範鍙ュ祵鍏ョ殑鏂规銆傚敖绠$畝鍗曠殑鍩虹嚎锛堜緥濡傚钩鍧囪瘝宓屽叆锛夋寔缁彁渚涘己鍔涚殑缁撴灉锛屽湪2017骞翠笅鍗婂勾鍜?018骞翠笂鍗婂勾鍑虹幇浜嗕竴浜涘垱鏂扮殑鏃犵洃鐫e拰鐩戠潱鏂规硶锛屼互鍙婂浠诲姟瀛︿範鏂规銆?/p>

璇嶈鏂规硶

杩欎竴棰嗗煙鐨勪竴鑸叡璇嗘槸锛岀洿鎺?span>骞冲潎涓€涓彞瀛愮殑璇嶅悜閲?/span>杩欎竴绠€鍗曟柟娉曪紙鎵€璋撹瘝琚嬫柟娉曪級锛屼负璁稿涓嬫父浠诲姟鎻愪緵浜嗗己鍔涚殑鍩虹嚎銆?/p>

Arora绛夊幓骞村湪ICLR鍙戣〃鐨勮鏂嘇 Simple but Tough-to-Beat Baseline for Sentence Embeddings鎻愪緵浜嗕竴涓緢濂界殑绠楁硶锛氶€夋嫨涓€绉嶆祦琛岀殑璇嶅祵鍏ワ紝缂栫爜鍙ュ瓙涓鸿瘝鍚戦噺鐨勭嚎鎬у姞鏉冪粍鍚堬紝鐒跺悗杩涜鐩稿悓鎴愬垎绉婚櫎锛堟牴鎹瑕佷富鎴愬垎绉婚櫎鍚戦噺鎶曞奖锛夈€傝繖涓€閫氱敤鏂规硶鍏锋湁娣卞埢鑰屽己澶х殑鐞嗚鍔ㄦ満锛屽熀浜庡湪璇瘒鍚戦噺涓婇殢鏈鸿璧颁互鐢熸垚鏂囨湰鐨勭敓鎴愬紡妯″瀷銆?/p>

杈惧鏂藉鐗瑰伐涓氬ぇ瀛︽渶杩戞彁鍑轰簡Concatenated p-mean Embeddings锛堣繛鎺-鍧囧€煎祵鍏ワ級锛屼竴涓瘮Arora绛夌殑绠楁硶鏇村己鍔茬殑璇嶈鍩虹嚎瀹炵幇锛堟劅璋aser鍚戞垜鎻愬強杩欎竴宸ヤ綔锛夈€?/p>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=璇嶈鏂规硶涓㈠け浜嗚瘝搴忥紝浣嗘槸淇濈暀浜嗘暟閲忔儕浜虹殑璇箟鍜岃娉曞唴瀹?/p>

鏃犵洃鐫f柟娉?/strong>

Jamie Kiros绛夊湪2015骞存彁鍑轰簡Skip-thoughts鍚戦噺锛屼竴绉嶅熀浜庢棤鐩戠潱璁粌鐨勬柟娉曘€?/p>

鏃犵洃鐫f柟妗堝皢瀛︿範鍙ュ瓙宓屽叆浣滀负瀛︿範棰勬祴鍙ュ瓙鎴栧瓙鍙ョ殑涓嬩竴鍙ョ殑鍓骇鍝併€傝繖涓€鏂规硶鍙互锛堢悊璁轰笂锛夊埄鐢ㄤ换浣曞寘鍚繛璐彞瀛愩€佸瓙鍙ョ殑鏂囨湰鏁版嵁闆嗐€?/p>

Skip-thoughts鍚戦噺鏄棤鐩戠潱瀛︿範鍙ュ祵鍏ョ殑鍏稿瀷渚嬪瓙銆傚彲浠ユ妸瀹冪湅鎴愯瘝宓屽叆鐨剆kip-gram妯″瀷鐨勫彞瀛愮増鏈細缁欏畾鍙ュ瓙锛岄娴嬪叾鍛ㄥ洿鐨勫彞瀛?/span>銆傝妯″瀷鍖呮嫭涓€涓熀浜嶳NN鐨勭紪鐮佸櫒-瑙g爜鍣ㄦ灦鏋勶紝鏍规嵁褰撳墠鍙ュ瓙閲嶅缓鍛ㄥ洿鐨勫彞瀛愩€?/p>

Skip-thoughts璁烘枃鐨?span>璇嶆眹鎵╁睍鏂规寰堟湁瓒o細閫氳繃瀛︿範RNN璇嶅祵鍏ョ┖闂村拰word2vec涔嬬被鐨勮瘝宓屽叆鐨勭嚎鎬у彉鎹紝鏉ュ鐞嗗湪璁粌涓湭瑙佺殑鍗曡瘝銆?/p>

鏈€杩戯紝Quick-thoughts鍚戦噺鍙戝睍浜哠kip-thoughts鍚戦噺銆傝璁烘枃鍦ㄤ粖骞寸殑ICLR涓婃姤鍛婅繃銆傚湪杩欎竴宸ヤ綔涓紝缁欏畾鍓嶄竴鍙ワ紝棰勬祴涓嬩竴鍙ョ殑浠诲姟琚噸鏁翠负涓€涓垎绫讳换鍔★細鐢ㄤ竴涓粠鍊欓€夊彞瀛愪腑閫夋嫨涓嬩竴鍙ョ殑鍒嗙被鍣ㄥ彇浠d簡瑙g爜鍣?/span>銆傝繖鍙互鐪嬫垚鏄鐢熸垚闂鐨勫垽鍒€艰繎銆?/p>

璇ユā鍨嬬殑涓€澶т紭鍔挎槸璁粌閫熷害锛堝拰Skip-thoughts妯″瀷鏈夋暟閲忕骇鐨勫樊寮傦級锛屽洜姝わ紝鍦ㄥぇ瑙勬ā鏁版嵁闆嗕笂锛屽畠鏄竴涓緢鏈夌珵浜夊姏鐨勬柟妗堛€?/p>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=Quick-thoughts鐨勫垎绫诲櫒浠庝竴缁勫彞宓屽叆涓€夋嫨涓嬩竴鍙?/p>

鐩戠潱鏂规硶

闀挎湡浠ユ潵锛屼汉浠涓猴紝鐩告瘮鏃犵洃鐫f柟娉曪紝鐩戠潱瀛︿範鍙ュ祵鍏ョ粰鍑虹殑宓屽叆璐ㄩ噺姣旇緝浣庛€傜劧鑰岋紝閮ㄥ垎鏄洜涓篒nferSent鐨勬彁鍑猴紝鏈€杩戣繖涓€鍋囧畾琚帹缈讳簡銆?/p>

鍜屼箣鍓嶆弿杩扮殑鏃犵洃鐫f柟娉曚笉鍚岋紝鐩戠潱瀛︿範闇€瑕佹爣娉ㄨ繃鐨勬暟鎹泦锛岃鏁版嵁闆嗗寘鎷负鏌愪釜浠诲姟娣诲姞鐨勬敞閲婏紝姣斿鑷劧璇█鎺ㄦ柇锛堜緥濡傦紝钑存兜鍙ュ锛夋垨鏈哄櫒缈昏瘧锛堢炕璇戝彞瀵癸級銆傝繖灏辨彁鍑轰簡闂锛氳閫夋嫨鍝竴绉嶄换鍔★紵澶氬ぇ鐨勬暟鎹泦鑳芥彁渚涗紭璐ㄥ祵鍏ワ紵鎴戜滑灏嗗湪涓嬩竴灏忚妭锛堝浠诲姟瀛︿範锛夎璁鸿繖浜涢棶棰樸€傚湪姝や箣鍓嶏紝璁╂垜浠湅涓?017骞村彂琛ㄧ殑InferSent杩欎竴绐佺牬銆?/p>

InferSent鏄竴涓緢鏈夎叮鐨勬柟娉曪紝瀹冪殑鏋舵瀯寰堢畝鍗曘€傚畠浣跨敤鍙ュ瓙缂栫爜鍣ㄥ湪Sentence Natural Language Inference dataset锛堜竴涓寘鍚?7涓囧彞瀛愬鐨勬暟鎹泦锛屾瘡涓彞瀛愬鏍囨敞涓轰腑鎬с€佸啿绐併€佽暣娑典笁涓被鍒腑鐨勪竴涓級涓婅缁冧竴涓垎绫诲櫒銆傚彞瀛愬涓殑鍙ュ瓙鍧囦娇鐢ㄧ浉鍚岀殑缂栫爜鍣ㄧ紪鐮侊紝鍒嗙被鍣ㄥ湪鐢变袱涓彞宓屽叆鏋勬垚鐨勮〃绀哄涓婅缁冦€傚彞瀛愮紪鐮佸櫒涓哄弻鍚慙STM鍔犱笂鏈€澶ф睜鍖栥€?/p>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?></p> 
<p class=澶氫换鍔″涔?/strong>

涔嬪墠鎴戜滑鎻愬埌杩囷紝鐩戠潱瀛︿範闇€瑕侀€夋嫨涓烘煇涓€浠诲姟鏍囨敞鐨勬暟鎹泦锛?/p>

鍝鐩戠潱璁粌浠诲姟鑳藉瀛︿範鍦ㄤ笅娓镐换鍔′腑閫氱敤鎬ф洿濂界殑鍙ュ祵鍏ワ紵

澶氫换鍔″涔犲彲浠ョ湅鎴愭槸Skip-thoughts銆両nferSent浠ュ強鐩稿叧鏃犵洃鐫?鐩戠潱瀛︿範鏂规鐨勬帹骞匡紝澶氫换鍔″涔犻€氳繃灏濊瘯鍦ㄤ竴绉嶈缁冩柟妗堜腑缁勫悎澶氱璁粌鐩爣鍥炵瓟涓婇潰鐨勯棶棰樸€?/p>

2018骞翠笂鍗婂勾鍙戣〃浜嗕竴浜涘浠诲姟瀛︿範鏂归潰鐨勫伐浣溿€傝鎴戜滑绠€鍗曚粙缁嶄笅MILA/MSR鐨?span>General Purpose Sentence Representation锛堜竴鑸洰鐨勫彞瀛愯〃绀猴級鍜孏oogle鐨?span>Universal Sentence Encoder锛堟櫘閫傚彞缂栫爜鍣級銆?/p>

MILA/MSR鐨勫伐浣滃湪ICLR 2018涓婂仛浜嗘姤鍛婏紙arXiv:1804.00079锛夛紝鍦ㄨ繖涓€宸ヤ綔涓紝Subramanian绛夎瀵熷埌锛屼负浜嗚兘澶熸帹骞垮埌鑼冨洿骞挎硾鐨勫绉嶄换鍔★紝鏈夊繀瑕佺紪鐮佸悓涓€鍙ュ瓙鐨勫涓柟闈€?/p>

鍥犳锛孲ubramanian绛夊埄鐢ㄤ竴涓?span>涓€瀵瑰鐨勫浠诲姟瀛︿範妗嗘灦锛岄€氳繃鍒囨崲浠诲姟瀛︿範鏅€傚彞宓屽叆銆備粬浠€夋嫨鐨?涓换鍔★紙棰勬祴鍚庡彞/鍓嶅彞銆佺缁忔満鍣ㄧ炕璇戙€佺煭璇粨鏋勮В鏋愩€佽嚜鐒惰瑷€鎺ㄦ柇绛夛級鍏变韩閫氳繃鍙屽悜GRU寰楀埌鐨勫彞瀛愬祵鍏ャ€傝瘯楠屾樉绀猴紝閫氳繃澶氳绁炵粡缈昏瘧浠诲姟鑳芥洿濂藉湴瀛︿範鍙ユ硶鎬ц川锛岄€氳繃瑙f瀽浠诲姟鍒欒兘鏇村ソ鍦板涔犻暱搴﹀拰璇嶅簭锛岃€岃缁冪缁忚瑷€鎺ㄦ柇鍙互鏇村ソ鍦扮紪鐮佽娉曚俊鎭€?/p>

Google鍦?018骞翠笂鍗婂勾鍙戣〃鐨?span>鏅€傚彞缂栫爜鍣?/span>閲囩敤浜嗗悓鏍风殑鏂规硶銆備粬浠殑缂栫爜鍣ㄤ娇鐢ㄤ簡涓€涓湪澶氱鏁版嵁婧愬拰澶氱浠诲姟涓婅缁冪殑杞崲鍣ㄧ綉缁滐紝浠ヤ究鍔ㄦ€佸湴瀹圭撼骞挎硾鐨勮嚜鐒惰瑷€鐞嗚В浠诲姟銆俆ensorFlow鎻愪緵浜嗕竴涓妯″瀷鐨勯璁粌鐗堟湰銆?/p>

缁撹

鏈€杩戝嚑涓湀鏉ワ紝鏅€傝瘝宓屽叆鍜屽彞宓屽叆杩欎竴棰嗗煙杩樻湁涓嶅皯鏈夎叮鐨勮繘灞曪紝鍖呮嫭璇勪及杩欎簺宓屽叆鐨勮〃鐜扮殑鏂规硶锛屼互鍙婂叧浜庡叾鍐呭湪鐨勫亸缃殑鐮旂┒锛堝鏅€傚祵鍏ヨ€岃█锛岃繖鏄竴涓ぇ闂锛夈€傛湰鏂囨病鑳借璁鸿繖浜涙渶鏂扮殑涓婚锛屼笉杩囨垜浠湪鏂囨湯鎻愪緵浜嗕竴浜涢摼鎺ャ€?br>

鎴戝笇鏈涗綘鍠滄杩欑瘒绠€鐭殑鍥為【銆傚鏋滀綘鍠滄鐨勮瘽锛岃鐐硅禐銆?/p>

鐩稿叧璧勬簮

  • Perone绛夋渶杩戝彂琛ㄤ簡涓€绡囨瘮杈僂LMo銆両nferSent銆丟oogle鏅€傚彞缂栫爜鍣ㄣ€乸-mean銆丼kip-thought绛夊祵鍏ユ柟娉曠殑璁烘枃锛歛rXiv:1806.06259

  • Hironsan鐨刧ithub浠撳簱鏀堕泦浜嗗緢澶氳瘝宓屽叆鐩稿叧鐨勮祫婧愶細Hironsan/awesome-embedding-models

  • 鍙ュ祵鍏ヨ鏂囷細Skip-Thoughts锛坅rXiv:1506.06726锛夈€丵uick-Thoughts锛坥penreview/rJvJXZb0W锛夈€丏iscSent锛坅rXiv:1705.00557锛夈€両nferSent锛坅rXiv:1705.02364锛夈€丮ILA/MSR鐨凣eneral Purpose Sentence Representations锛坅rXiv:1804.00079锛? Google鐨刄niversal Sentence Encoder锛坅rXiv:1803.11175锛夈€丟oogle鐨処nput-Ouput Sentence learning on dialog锛坅rXiv:1804.07754锛夈€?/p>

  • 濡傛灉浣犲鍙ュ祵鍏ョ殑璇勪及鏂瑰紡鎰熷叴瓒o紝鍒敊杩嘑alebook鏈€杩戠殑宸ヤ綔SentEval锛屼互鍙奛YU銆乁W銆丏eepMind鐨勭爺绌朵汉鍛樻渶杩戝彂琛ㄧ殑GLUE璇勬祴妗嗘灦銆?/p>


鏉ユ簮锛歁edium
缂栬瘧锛歸eakish


鎺ㄨ崘闃呰

1

2


3


4


5



宕斿簡鎵?/strong>

闈欒鍗氬鍗氫富锛?/span>銆奝ython3缃戠粶鐖櫕寮€鍙戝疄鎴樸€嬩綔鑰?/span>

闅愬舰瀛?/span>

鑷劧璇█澶勭悊鍩虹煶 Embedding 鏈€鏂拌繘灞曟眹鎬?> 
           </section> 
          </section> 
         </section> 
        </section> 
        <section class=
鐐?

杩欓噷鈥滈槄璇诲師鏂団€濓紝鏌ョ湅鏇村

以上是关于的主要内容,如果未能解决你的问题,请参考以下文章

Python 操作Redis

python爬虫入门----- 阿里巴巴供应商爬虫

Python词典设置默认值小技巧

《python学习手册(第4版)》pdf

Django settings.py 的media路径设置

Python中的赋值,浅拷贝和深拷贝的区别