Ku-xidhka Maqalka iyo Kediska Barashada Matalaada

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay gundhig maqal ah?
   • Dhaqdhaqaaq dhererkiisu go'an yahay oo tirooyin ah oo matalaya dhawaaqa muuqaalkiisa macnaha leh
   • Nooc MP3 ah oo la cufan ee faylka maqalka ah
   • Qoraal qoraal ah oo erayada lagu hadlo
   • Heerka mugga lagu qiyaasay decibels
2. Maxay tahay sababta barashada is-maamulku ay aad muhiim ugu tahay dhegeysiga maqalka?
   • Waxay u baahan tahay maqal si loogu duubo istuudiyaha
   • Waxay ka barataa matalo faa'iido leh oo maqal ah oo aan la calaamadin, iyadoo yaraynaysa baahida calaamadaha aadanaha
   • Waxa kaliya oo ay ka shaqeysaa muusigga, ma aha hadalka
   • Waxay dammaanad qaadaysaa saxnaanta qoraalka
3. Sidee ayuu Wav2Vec 2.0 wax u bartaa xilliga tababarka ka hor?
   • Adigoo maqalka u turjumaya Faransiis
   • Adigoo si sax ah u xifdisanaya feyl kasta oo tababar ah
   • Adigoo qarinaya taako maqal ah iyo saadaalinta cutubka saxda ah ee la qiyaasi karo ee ka dhex jira kuwa wax jeediya
   • Adiga oo cabbiraya cabbirka faylka clip kasta
4. Muxuu moodelka sida CLAP ku toosiyaa meel la wadaago dhexda?
   • Laba muuqaal oo kala duwan
   • Pixels iyo histograms midabka
   • Heerka batteriga iyo isticmaalka CPU
   • Qeexitaannada qoraalka maqalka ah iyo kuwa ku habboon
5. Kahor inta aan codka la siin shabakada neerfaha, waa maxay isbeddelka guud ee inta badan lagu dabaqo?
   • U beddelashada spectrogram ama matalaadda shaandhada la bartay
   • U daabacaadda sida muusig xaashida ah
   • Tirtirka dhammaan inta jeer ee ka sarreeya maqalka aadanaha si joogto ah
   • U beddelashada sawirka JPEG ee daboolka albamka
6. Waa kuwee shaqada dhabta ah ee si toos ah ugu tiirsan isbarbardhigga ku-xidhnaanta maqalka?
   • Xisaabinta qiimaha heesta iibkeeda
   • Dheejinta hadalka, iyadoo la ogaanayo cidda hadlay goorta
   • Doorashada farshaxanka daboolka albamka
   • Dejinta dhalaalka shaashadda
7. During contrastive audio-text training, what happens to a matching audio clip and its text description?
   • They are converted into the same raw waveform
   • The text embedding is discarded
   • Their embeddings are pulled closer while mismatched pairs are pushed apart
   • The audio is assigned a speaker name as a class label
8. Why can one self-supervised audio embedding be reused across several downstream tasks?
   • It stores the original waveform without compression
   • It contains a manually written label for every possible sound
   • It works only for one speaker and one recording
   • It captures general acoustic structure that can support tasks with little extra labeled data