Ikibazo gihuza ibibazo · Hagati Urwego

Uruvange rwimbitse Ikibazo

Gerageza MoD yize token router, itunganijwe hejuru-k ubushobozi, inzira isigaye kubimenyetso byasimbuwe, imiterere ihagaze, hamwe nuburyo bwo kwirinda inzira.

Inzira zijyanye no kuyoboraUruvange rwimbitse
Ikibazo 1 Bya 8

Niki Kuvanga Ubujyakuzimu bitandukana kubimenyetso?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Niki Kuvanga Ubujyakuzimu bitandukana kubimenyetso?
    • Ni bangahe kubara (ibice) buri kimenyetso cyakira
    • Amagambo yakoreshejwe
    • Amahugurwa yamakuru
    • Ururimi rusohoka
  2. Nigute MoD ituma ingengo yimari ya compte iteganijwe?
    • Mugihe cyo guta igice cyuburemere bwikitegererezo
    • Mugutunganya neza-k igice cyibimenyetso kuri buri gice
    • Ukoresheje icyitegererezo gitandukanye kuri buri nteruro
    • Muguhagarika ibitekerezo rwose
  3. Bigenda bite kubimenyetso router idahitamo kumurongo runaka?
    • Basibwe kurukurikirane
    • Boherejwe kuri moderi itandukanye
    • Basimbuka guhagarika bakoresheje ihuza risigaye, banyura mubidahindutse
    • Bahinduwe mumashusho
  4. Niyihe laboratoire yasohoye tekinike yubuvange?
    • OpenAI muri 2019
    • Meta muri 2015
    • Microsoft muri 2012
    • Google Ibitekerezo byimbitse muri 2024
  5. Guhuza MoD nuruvange-rw-Impuguke bitanga iki?
    • MoDE, inzira igana kubwimbitse no guhitamo abahanga
    • Icyitegererezo kidafite ibipimo
    • Umuyoboro mwiza
    • Sisitemu ya tokenizer-yonyine
  6. Ni ukubera iki top-k ihamye ihitamo kurenza icyerekezo kuri MoD?
    • Cyakora icyitegererezo gisohoka inyandiko idasanzwe
    • Igumana imiterere ya tensor ihagaze neza, ikora neza kubikoresho
    • Bikuraho ibikenewe mumahugurwa
    • Yongera ingano yamagambo
  7. What does the learned router do at a Mixture-of-Depths block?
    • Selects a new model family for every prompt
    • Predicts labels for the next training batch
    • Scores tokens so the block can select which ones receive full attention and MLP computation
    • Changes the vocabulary size at inference
  8. Why does causal generation in MoD use auxiliary predictors for routing?
    • To let a model read the entire answer before generating
    • To remove the fixed top-k compute budget
    • To convert the sequence into an image
    • To make routing decisions without looking at future tokens