Ikibazo gihuza ibibazo · Hagati Urwego

T5 na Inyandiko-Kuri-Kwimura Ikibazo

Ikizamini cya T5 ibanzirizasuzuma, span ruswa, encoder-decoder kwitondera, kandi itandukaniro riri hagati yo guhuza amabwiriza nibisubizo bifunze-ibitabo.

Inzira zijyanye no kuyoboraT5 Kandi Inyandiko Kuri Kwimura
Ikibazo 1 Bya 8

Ni ikihe gitekerezo gisobanura inyuma ya T5?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ikihe gitekerezo gisobanura inyuma ya T5?
    • Ikora gusa imashini
    • Itanga amashusho ahereye ku bisobanuro
    • Buri gikorwa cya NLP cyateguwe nkinyandiko yinjiza kubisohoka
    • Isimbuza amahugurwa yose hamwe no kwiga gushimangira
  2. Ni ubuhe bwubatsi bwa Transformer T5 ikoresha?
    • Kodegisi yuzuye
    • Encoder-gusa, nka BERT
    • Decoder-gusa, nka GPT
    • Nta kwitabwaho na gato
  3. Niyihe ntego yo kwitegura T5 ikoresha cyane cyane?
    • Guhanura-interuro gusa
    • Guteganya ibiciro byimigabane
    • Igishushanyo
    • Ruswa ya ruswa, guhisha ibice bihuza byasimbujwe ibimenyetso bya sentinel
  4. Nuwuhe murambo T5 wateguwe mbere?
    • Wikipedia gusa
    • C4 Ikomeye Isukuye Corpus
    • Dataset yinkomoko yinkomoko
    • Inyandiko-mvugo
  5. Nigute T5 igereranya umurimo wo gutondeka nko gusesengura amarangamutima?
    • Isohora ibishoboka gusa vector
    • Isubiza indangamuntu
    • Itanga ijambo risanzwe ijambo, nka 'positif' cyangwa 'negative'
    • Ntishobora gukora ibyiciro
  6. Niki FLAN-T5 ugereranije na T5 yumwimerere?
    • Icyitegererezo gito
    • Amabwiriza-yahinduwe urubyaro rushyira mubikorwa bitagaragara
    • Sisitemu yo kumenya imvugo
    • Igikoresho kidafitanye isano
  7. When T5 generates a summary, what does decoder cross-attention let it consult?
    • The encoder's representations of the input article
    • Only previously generated summary tokens
    • Only the unprocessed task label
    • The training optimizer's stored gradients
  8. In a closed-book T5 question-answering setup, where does the answer's factual information come from at inference?
    • A document retrieval step required for every question
    • A newly supplied answer passage required in the prompt
    • A live database lookup built into span corruption
    • Knowledge encoded in the model's learned parameters