Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo cya WaveNet

Ongera usubiremo ibyerekezo bya WaveNet byikora, impanvu yagutse, n'impamvu yambere yo gutoranya yatinze.

Inzira zijyanye no kuyoboraWavenet
Ikibazo 1 Bya 8

Nigute WaveNet itanga amajwi?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nigute WaveNet itanga amajwi?
    • Icyitegererezo kimwe kibisi icyarimwe, giteganijwe kubiteganijwe mbere
    • Muguhitamo amashusho mumasomero ahamye
    • Muguhindura spekrogramu muburyo bwanditse
    • Ingero zose icyarimwe nta gutumiza
  2. Ni ubuhe buryo bw'ingenzi bushya bwo kwemeza muri WaveNet?
    • Igice kinini cyo guhuriza hamwe
    • Impamvu zagutse
    • Amarembo asubirwamo gusa
    • Byuzuye
  3. How did the original WaveNet represent quantized waveform samples for prediction?
    • As a single continuous average for the entire clip
    • As speaker IDs only
    • As a categorical distribution over 256 µ-law-quantized values
    • As a sequence of image patches
  4. Ni ikihe kibazo gikomeye cyagaragaye cya WaveNet y'umwimerere?
    • Byashoboraga gusa guceceka
    • Ntabwo byasabye amakuru yamahugurwa
    • Cyakoraga kumashusho gusa
    • Icyitegererezo gikurikiranye cyatumye ibisekuruza bitinda
  5. In the speech-synthesis experiments described in the WaveNet paper, what does the model generate?
    • Raw speech waveforms sample by sample, conditioned on linguistic features
    • A mel-spectrogram that is itself the final audio
    • Only phoneme labels without audio
    • A MIDI score
  6. Nuwuhe muryango watangije WaveNet?
    • NASA
    • DeepMind
    • OpenAI
    • Adobe
  7. What does causal convolution prevent in WaveNet?
    • Using preceding samples as context
    • Generating any audio waveform
    • Using future audio samples to predict the current sample
    • Conditioning on text or speaker features
  8. What activation structure does the original WaveNet use in its gated convolutional layers?
    • A ReLU output concatenated with image patches
    • A softmax over speakers only
    • A linear layer without nonlinearities
    • A tanh-filter output multiplied elementwise by a sigmoid gate