10 Manyan Buɗaɗɗen Tushen Kayan Aikin Hannu na Artificial don Linux

A cikin wannan sakon, za mu rufe kaɗan daga cikin manyan kayan aikin fasaha na wucin gadi (AI) don yanayin yanayin Linux. A halin yanzu, AI yana ɗaya daga cikin fannonin da suka ci gaba a fannin kimiyya da fasaha, tare da babban mai da hankali kan gina software da kayan masarufi don warware matsalolin rayuwa ta yau da kullun a fannoni kamar kiwon lafiya, ilimi, tsaro, masana'antu, banki da sauransu.

A ƙasa akwai jerin dandamali da yawa waɗanda aka tsara kuma aka haɓaka don tallafawa AI, waɗanda zaku iya amfani da su akan Linux da yuwuwar sauran tsarin aiki da yawa. Ka tuna wannan jeri ba a shirya shi cikin kowane takamaiman tsari na sha'awa ba.

1. Zurfin Koyo Don Java (Deeplearning4j)

Deeplearning4j darajar kasuwanci ce, tushen buɗe ido, toshewa da wasa, rarraba zurfin koyo don harsunan shirye-shiryen Java da Scala. An tsara shi musamman don aikace-aikacen da ke da alaƙa da kasuwanci, kuma an haɗa shi tare da Hadoop da Spark a saman CPUs da GPUs masu rarraba.

An saki DL4J a ƙarƙashin lasisin Apache 2.0 kuma yana ba da tallafin GPU don ƙima akan AWS kuma an daidaita shi don ƙirar ƙananan sabis.

Ziyarci Shafin Gida: http://deeplearning4j.org/

2. Caffe - Tsarin Koyo mai zurfi

Caffe tsari ne na zamani da bayyana zurfin ilmantarwa dangane da sauri. An sake shi a ƙarƙashin lasisin BSD 2-Clause, kuma ya rigaya yana tallafawa ayyukan al'umma da yawa a fannoni kamar bincike, samfuran farawa, aikace-aikacen masana'antu a fannoni kamar hangen nesa, magana da multimedia.

Ziyarci Shafin Gida: http://caffe.berkeleyvision.org/

3. H20 - Tsarin Koyon Injin Rarraba

H20 buɗaɗɗen tushe ne, mai sauri, mai daidaitawa da tsarin ilmantarwa na inji, tare da nau'ikan algorithms sanye take kan tsarin. Yana goyan bayan aikace-aikacen da ya fi wayo kamar zurfin koyo, haɓaka gradient, dazuzzuka bazuwar, ƙirar layi na gaba ɗaya (watau koma baya na logistic, Elastic Net) da ƙari mai yawa.

Kayan aiki ne na kasuwanci wanda ke da alaƙa da bayanan ɗan adam don yanke shawara daga bayanai, yana bawa masu amfani damar zana fahimta daga bayanan su ta amfani da ƙirar ƙira mai sauri da inganci.

Ziyarci Shafin Gida: http://www.h2o.ai/

4. MLlib - Library Learning Machine

MLlib buɗaɗɗen tushe ne, mai sauƙin amfani da ɗakin karatu na koyo na inji wanda aka haɓaka azaman ɓangaren Apache Spark. Yana da sauƙin turawa kuma yana iya aiki akan gungu na Hadoop da bayanai.

MLlib kuma yana jigilar kayayyaki tare da tarin algorithms don rarrabuwa, koma baya, shawarwari, tari, nazarin rayuwa da ƙari mai yawa. Mahimmanci, ana iya amfani da shi a cikin Python, Java, Scala da yarukan shirye-shiryen R.

Ziyarci Shafin Gida: https://spark.apache.org/mllib/

5. Apache Mahout

Mahout wani tsari ne na budadden tushe wanda aka ƙera don gina aikace-aikacen koyo na inji, yana da fitattun siffofi guda uku da aka jera a ƙasa:

  1. Yana samar da wurin aiki mai sauƙi kuma mai iya daidaitawa
  2. Yana ba da nau'ikan algorithms da aka riga aka shirya don Scala + Apache Spark, H20 da Apache Flink
  3. Ya haɗa da Samaras, wurin aikin gwajin lissafin vector tare da daidaitawar R-kamar

Ziyarci Shafin Gida: http://mahout.apache.org/

6. Buɗe Laburaren Neural Networks (BuɗenNN)

OpenNN kuma ɗakin karatu ne na buɗe tushen tushen da aka rubuta a cikin C++ don zurfafa ilmantarwa, ana amfani da shi don ƙaddamar da hanyoyin sadarwa. Koyaya, yana da kyau kawai ga ƙwararrun masu shirye-shiryen C++ da mutanen da ke da ƙwarewar koyon injin. Yana da alaƙa da gine-gine mai zurfi da babban aiki.

Ziyarci Shafin Gida: http://www.opennn.net/

7. Oriya 2

Oryx 2 ci gaba ne na aikin Oryx na farko, an haɓaka shi akan Apache Spark da Apache Kafka a matsayin sake fasalin gine-ginen lambda, kodayake sadaukar da kai don cimma nasarar koyo na injin na ainihin lokaci.

Dandali ne don haɓaka aikace-aikace da jigilar kaya tare da wasu aikace-aikace haka nan don tace haɗin gwiwa, rarrabuwa, koma baya da dalilai na tari.

Ziyarci Shafin Gida: http://oryx.io/

8. OpenCyc

OpenCyc ita ce tashar buɗe ido zuwa ga mafi girma kuma mafi girman tushen tushen ilimin gabaɗaya da injin tunani na gama gari na duniya. Ya haɗa da ɗimbin sharuddan Cyc da aka tsara a cikin ƙayyadaddun ilimin ilimin halittar jiki don aikace-aikace a fannoni kamar:

  1. Tsarin ƙirar yanki mai wadata
  2. Tsarin ƙwararrun ƙwararrun yanki
  3. Fahimtar rubutu
  4. Haɗin bayanan Semantic da wasannin AI da ƙari da yawa.

Ziyarci Shafin Gida: http://www.cyc.com/platform/opencyc/

9. Apache SystemML

SystemML dandamali ne mai buɗe ido na wucin gadi don koyan inji manufa don manyan bayanai. Babban fasalinsa shine - yana gudana akan R da Python-kamar syntax, mai da hankali kan manyan bayanai kuma an tsara su musamman don babban matakin lissafi. An bayyana yadda yake aiki da kyau a shafin farko, gami da nunin bidiyo don bayyananniyar kwatanci.

Akwai hanyoyi da yawa don amfani da shi ciki har da Apache Spark, Apache Hadoop, Jupyter da Apache Zeppelin. Wasu sanannun shari'o'in amfani da shi sun haɗa da motoci, zirga-zirgar jirgin sama da banki na zamantakewa.

Ziyarci Shafin Gida: http://systemml.apache.org/

10. NuPIC

NuPIC tsarin buɗaɗɗen tushe don koyan injina wanda ya dogara akan Heirarchical Temporary Memory (HTM), ka'idar neocortex. An aiwatar da shirin HTM da aka haɗa a cikin NuPIC don nazarin bayanan yawo na ainihi, inda yake koyon tsarin tushen lokaci da ke cikin bayanai, yana tsinkayar ƙima mai zuwa tare da bayyana duk wani rashin daidaituwa.

Fitattun abubuwanta sun haɗa da:

  1. Ci gaba da koyan kan layi
  2. Tsarin lokaci da sararin samaniya
  3. Bayanan yawo na ainihi
  4. Hasashen da yin samfuri
  5. Gano mai ƙarfi mai ƙarfi
  6. Ƙwaƙwalwar lokaci mai matsayi

Ziyarci Shafin Gida: http://numenta.org/

Tare da haɓakawa da ci gaba da bincike a cikin AI, za mu daure mu shaida ƙarin kayan aikin da ke tasowa don taimakawa wannan fannin fasaha ya yi nasara musamman don magance ƙalubalen kimiyya na yau da kullun tare da dalilai na ilimi.

Kuna sha'awar AI, menene ra'ayin ku? Ba mu ra'ayoyinku, shawarwarinku ko duk wani ra'ayi mai inganci game da batun ta sashin sharhin da ke ƙasa kuma za mu yi farin cikin ƙarin sani daga naku.