Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cyo Gukata Icyitegererezo

Kugenzura gutema ibyubatswe kandi bitubatswe, imipaka idasanzwe-intungamubiri, kugarura itera, kwishyuza tike ya tombora, no gupima ibyoherejwe.

Inzira zijyanye no kuyoboraGukata Icyitegererezo
Ikibazo 1 Bya 8

Gukata icyitegererezo bikuraho iki kumurongo watojwe?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Gukata icyitegererezo bikuraho iki kumurongo watojwe?
    • Kuraho uburemere cyangwa imiterere itanga bike mubisohoka
    • Ongeraho ibipimo byinshi kugirango utezimbere
    • Guhugura icyitegererezo kubihe byinshi
    • Kongera ibyinjira byinjira
  2. Ni iki gitandukanya gutema byubatswe no gutema bitubatswe?
    • Gutema byubatswe bikora kumashusho gusa
    • Gutema byubatswe bikuraho ibice byose nka neuron cyangwa imitwe; itubatswe ikuraho uburemere bwa buri muntu
    • Gutema bitubatswe bisaba kwitoza guhera
    • Nta tandukaniro rifatika riri hagati yabo
  3. Ni ukubera iki gutema bitubatswe akenshi binanirwa kwihutisha icyitegererezo kubikoresho bisanzwe?
    • Irasiba ibice byinshi
    • Burigihe bigabanya ukuri kuri zeru
    • Uburemere buke matrices ikenera ibyuma byihariye cyangwa amasomero kugirango ukoreshe zeru
    • Yongera ibipimo bibarwa
  4. Ni ubuhe buryo busanzwe bwo gutema ibiti?
    • Kata byose icyarimwe, hanyuma wohereze
    • Banza ubanze, hanyuma ntuzongere kwitoza
    • Bisanzwe usibe ibice kugeza igihe ukuri kugabanuka
    • Gutoza, gukata byibuze-byingenzi, guhuza neza kugirango ukire, kandi usubiremo
  5. Hypothesis ya Tike ya Tombora isaba iki?
    • Imiyoboro yuzuye irimo sisitemu ntoya ishobora guhuza icyitegererezo cyuzuye mugihe yatojwe uhereye kubitangira neza
    • Gukata buri gihe bitezimbere
    • Gukata bisanzwe ni byiza
    • Moderi nini ntishobora guhagarikwa
  6. Nibihe bipimo byoroshye bikoreshwa muguhitamo uburemere bwo gukata?
    • Urutonde rwimyandikire yizina ryamazina
    • Ubunini bwuzuye bwibipimo
    • Ibara ry'ikarita yo gukora
    • Umubare w'ingero zamahugurwa
  7. When is pruning most likely to reduce inference latency on ordinary dense hardware?
    • When individual zeros are added but the dense kernel still computes them
    • When the model is renamed without changing its graph
    • When whole channels or other structures are removed to produce a smaller dense model
    • When a sparse checkpoint is stored in a larger file
  8. What should a team measure after pruning for a deployment target?
    • Only the number of zeros in a checkpoint
    • Only training accuracy before pruning
    • Whether the file extension changed
    • Task quality and latency or memory on the intended hardware