Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Qaababka Fiiro Gaarka ah

Is barbar dhig isku xidhka dareenka maxaliga ah, kala furfuran, random, iyo caalamiga ah iyo sida kernels-ka la hagaajiyay uga boodaan galitaanka aan la isticmaalin ee habaynta galinta dheer.

Wadooyinka haga ee la xidhiidhaQaababka Fiiro Gaarka Leh
Su'aal 1 ee 8

Waa maxay sababta feejignaanta buuxda ugu qaalisan tahay taxanaha dheer?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay sababta feejignaanta buuxda ugu qaalisan tahay taxanaha dheer?
    • Qiimaheedu waxa uu ku koraa labajibbaaran ee dhererka isku xigxiga
    • Ma isticmaali karto GPU-yada
    • Keliya waxay ku shaqaysaa ereyo gaagaaban
    • Wax xusuus ah uma baahna haba yaraatee
  2. Waa maxay dareenka simbiriirixidda-daaqadda (maxalli)?
    • Calaamad kastaa waxay u dhigantaa xog-ururin random
    • Calaamad kastaa waxay u dhigantaa oo keliya deriskeeda u dhow
    • Calaamaduhu weligood ma xaadiraan wax
    • Calaamad kastaa waxay ka qaybgashaa dhammaan calaamadaha laba jeer
  3. Waa maxay ujeedada 'calaamadaha caalamiga' ee dareenka aan badnayn?
    • Si loo yareeyo qaabka
    • Si loo kaydiyo sawirada
    • Si aad u noqoto xarun wax walba ka soo qaybgasha oo wax walba ay ka qayb galaan
    • Si loo beddelo ereyada
  4. Qaabkee caddeeyey in isku darka random, daaqada, iyo dareenka caalamiga ahi ay ilaalinayso muujinta feejignaan buuxda?
    • ResNet
    • Word2Vec
    • AlexNet
    • BigBird
  5. Qiyaas ahaan sidee ayay tirada isku xidhka u cabbiraysaa dareenka yar ee sida wanaagsan loo qaabeeyey?
    • Qiyaas ahaan toosan oo leh dherer isku xiga
    • Afar ahaan dhererka isku xigxiga leh
    • Kuma xidhna dhererka
    • Si weyn oo leh dherer isku xigxig leh
  6. What do strided or dilated sparse-attention patterns allow?
    • Xisaabiyaa iyaga oo markaas ka dhigaya eber
    • Skip ahead to reach distant context with fewer attention links
    • Waxay u diraan server kale
    • Waxay u kaydisaa si joogto ah
  7. How does Longformer combine local and global attention?
    • Every token attends to every other token
    • Only random edges are allowed
    • Most tokens use a sliding local window, while selected global tokens connect across the sequence
    • Each token attends only to the final token
  8. What is the goal of learned, content-based sparse attention?
    • A full dense connection to every token
    • A rule for removing all future tokens
    • A learned vocabulary for each attention head
    • Avoid spending attention compute and memory on content unrelated to the current query