Kedis ku xidhan hagaha · Adag Heerka

Kediska Hadalka Is-Idaamaynta ee HuBERT

Hubi sida HuBERT u abuurto bartilmaameedyo ay kooxdu ka soo jeedo ee maqalka daboolan, waxay ku celcelisaa summadaha, oo ay qiimeeyaan matalaadda la bartay ee cabbirrada hadalka.

Wadooyinka haga ee la xidhiidhaHadalka Hubert Self Kormeeray
Su'aal 1 ee 8

Sidee HuBERT u dhalisaa yoolalka ay isku daydo inay saadaaliso inta lagu jiro tababarka?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Sidee HuBERT u dhalisaa yoolalka ay isku daydo inay saadaaliso inta lagu jiro tababarka?
    • Khadka-la'aanta k- macneheedu waa ururinta sifooyinka maqalka ee unugyada qarsoon
    • Qoraallada aadanaha
    • Calaamadaha qaylada random
    • Isbarbardhigga shaybaarada wax jeediya
  2. Qaabkee qoraalka caanka ah ayaa dhiirigeliyay nidaamka tababarka saadaasha ee HuBERT?
    • GPT-3
    • BERT
    • ResNet
    • Word2Vec
  3. Sidee HuBERT u wanaajisaa sumadaha bartilmaameedka ah ee tababarrada isdaba jooga ah?
    • Waxay shaqaaleysiisaa annotators badan
    • Waxay ku daraysaa qaylo cayriin dheeraad ah
    • Waxay dib u ururisaa iyadoo adeegsanaysa tusaalaha muuqaalkiisa oo la hagaajiyay
    • Weligood way qaboojisaa kooxihii ugu horreeyay
  4. Waa maxay sifooyinka sida caadiga ah lagu soo ururiyey HuBERT soocdiisii ugu horeysay?
    • Erayada qoraalka ugu dambeeya
    • pixels sawirka
    • Ku-xidhnaanta aqoonsiga afhayeenka
    • Astaamaha codka fudud ee MCCC
  5. Muxuu HuBERT u baran karaa qaab-dhismeed faa'iido leh xitaa marka kooxaysigeedu aanu qummanayn?
    • Waxa kaliya oo ay u baahan tahay in la saadaaliyo aqoonsiga kooxeed ee meelaha waji-xidhan, kaas oo sii ahaanaya hawl joogto ah
    • Waxay iska indha tirtaa gabi ahaanba kooxaha
    • Waxay adeegsataa koox-ururin oracle kaamil ah
    • Waxay xifdisanaysaa jaangooyooyinka imtixaanka
  6. Which pretraining challenge does HuBERT address by using offline cluster labels?
    • Speech has no temporal structure
    • There is no lexicon of sound units available during pretraining
    • Every speaker uses the same pitch
    • Audio contains no repeated patterns
  7. In the HuBERT paper, what relative WER reductions did the 1B-parameter model report on dev-other and test-other?
    • Exactly 50% on both sets
    • No measurable change on either set
    • Up to 19% on dev-other and 13% on test-other
    • Up to 13% on dev-other and 19% on test-other
  8. Which benchmark sets did the original HuBERT paper use for speech-recognition results?
    • ImageNet iyo COCO
    • GLUE and SuperGLUE
    • SQuAD and TriviaQA
    • LibriSpeech and Libri-light