WaveNet adanwo

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

1. Bawo ni WaveNet ṣe ṣe agbejade ohun?
   • Apeere aise kan ni akoko kan, ni ilodi si awọn ayẹwo iṣaaju
   • Nipa yiyan awọn agekuru lati ile-ikawe ti o wa titi
   • Nipa yiyipada spectrograms taara si ọrọ
   • Gbogbo awọn ayẹwo nigbakanna pẹlu ko si ibere
2. Kini isọdọtun convolutional bọtini ni WaveNet?
   • Max pooling fẹlẹfẹlẹ
   • Diated fa convolutions
   • Awọn ẹnu-ọna loorekoore nikan
   • Ni kikun ti sopọ fẹlẹfẹlẹ
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. Kini apadabọ ilowo pataki ti WaveNet atilẹba?
   • O le gbe ipalọlọ nikan
   • Ko nilo data ikẹkọ
   • O ṣiṣẹ nikan lori awọn aworan
   • Iṣapẹẹrẹ lesese ṣe iran lọra
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. Ẹgbẹ wo ni o ṣafihan 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