Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cya RNN-Transducer

Sobanukirwa uburyo moderi ya RNN-Transducer ihuza amakadiri ya acoustic namateka yo gusohora kugirango akurikirane imirimo yo kwandukura.

Inzira zijyanye no kuyoboraRnn Moderi ya Transducer
Ikibazo 1 Bya 8

Ni izihe mbogamizi za CTC RNN-Transducer ikemura?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni izihe mbogamizi za CTC RNN-Transducer ikemura?
    • CTC isabwa-kwigenga kwifata, irinda kwerekana ibyasohotse-token biterwa
    • CTC ntishobora gutunganya amajwi na gato
    • CTC isaba GPU gukora
    • CTC ikora amashusho gusa
  2. Nibihe bintu bitatu byingenzi bigize RNN-Transducer?
    • Encoder, decoder, nivangura
    • Encoder (umuyoboro wandikirwa), umuyoboro wo guhanura, hamwe numuyoboro uhuriweho
    • Tokenizer, gushiramo, na softmax gusa
    • Imitwe ibiri yitonderwa hamwe na router
  3. Ni uruhe ruhare umuyoboro uhanura ugira muri RNN-T?
    • Ihagarika amajwi yerekana amajwi
    • Ipima ingano ya mikoro
    • Ikora nkicyitegererezo cyururimi, itondekanya ibimenyetso byasohotse mbere
    • Ihitamo igipimo cyicyitegererezo
  4. Kuki RNN-T ishyigikira umuvuduko muke gutemba muburyo busanzwe?
    • Itunganya imvugo yose icyarimwe mbere yo kohereza ikintu icyo aricyo cyose
    • Irengagiza ibihe byose
    • Ikora gusa kuri dosiye zanditswe mbere
    • Guhuza kwayo ni monotonic n'ibumoso-iburyo, bitera imbere binyuze mumajwi hamwe nubusa
  5. Muri decode ya RNN-T, gusohora ikimenyetso cyambaye iki?
    • Iterambere kumwanya ukurikira wamajwi
    • Kurangiza inyandiko yose
    • Gusiba ijambo ryabanje
    • Hindura urundi rurimi
  6. Ni ubuhe bwoko bwa encoder bugenda busimbuza LSTM muri sisitemu igezweho ya sisitemu ya RNN-T?
    • Ibiti
    • Kode ya kodegisi
    • Umufuka wuzuye-w-amagambo
    • Icyitegererezo cya Markov
  7. What two histories condition the RNN-T joint network’s output distribution?
    • Only the future transcript and a speaker photo
    • Only the last acoustic frame, without output context
    • The encoder representation of acoustic input and the prediction-network representation of prior output labels
    • The optimizer state and learning rate
  8. What alignment behavior does an RNN-T allow before advancing to the next audio frame?
    • It must emit exactly one label for every frame
    • It advances time only after a word-final token
    • It generates all output before reading any audio
    • It can emit zero or more nonblank labels at the current frame, then a blank advances time