Maswali yanayounganishwa na mwongozo · Ngumu Kiwango

Maswali ya AudioLM

Fahamu jinsi AudioLM inavyoonyesha muendelezo wa sauti kwa kutumia tokeni za kisemantiki na akustika, uwakilishi wa kodeki na mafunzo ya sauti pekee.

Njia za mwongozo zinazohusianaSauti ya sauti
Swali 1 ya 8

AudioLM hutumia wazo gani la msingi kutoa sauti?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. AudioLM hutumia wazo gani la msingi kutoa sauti?
    • Inachukulia sauti kama mfuatano wa ishara na kutabiri inayofuata, kama modeli ya lugha
    • Inatoa sauti kutoka kwa muziki wa laha
    • Ni sampuli za kelele za nasibu bila kujifunza
    • Inarekodi na kurudia maonyesho ya wanadamu
  2. Ni nini jukumu la ishara za 'semantic' katika AudioLM?
    • Wanakamata sauti nzuri na kurekodi kelele
    • Hunasa muundo wa muda mrefu kama vile fonetiki na melodi
    • Wanahifadhi nakala ya maandishi
    • Wanadhibiti umbizo la faili towe
  3. Ni codec gani ya neural hutoa tokeni za akustisk za AudioLM?
    • MP3
    • WaveNet
    • SoundStream
    • Whisper
  4. Je, AudioLM inahitaji nini kama data ya mafunzo?
    • Sauti imeoanishwa na manukuu ya kina
    • Faili za MIDI pekee
    • Picha za mawimbi
    • Sauti mbichi pekee, isiyo na manukuu
  5. After AudioLM predicts semantic tokens, how does its hierarchical generation add acoustic detail?
    • It predicts coarse acoustic tokens, then fine acoustic tokens
    • It predicts fine acoustic tokens, then coarse acoustic tokens
    • It predicts coarse acoustic tokens before semantic tokens
    • It predicts acoustic tokens without a semantic-token stage
  6. Ni mfumo gani wa baadaye uliojengwa moja kwa moja kwenye mbinu ya msingi ya tokeni ya AudioLM?
    • Stable Diffusion
    • MusicLM
    • GPT-2
    • DALL-E
  7. Which model does AudioLM use to produce its semantic token sequence?
    • A text-only GPT decoder trained on transcripts
    • The SoundStream waveform decoder
    • A self-supervised w2v-BERT representation model
    • A hand-authored MIDI score
  8. What audio types did the AudioLM paper demonstrate continuing from an audio prompt?
    • Only environmental noise and no music
    • Spoken text transcripts and MIDI files only
    • Images and silent video clips
    • Speech and piano music