Maswali yanayounganishwa na mwongozo · Ngumu Kiwango

Maswali ya Kizazi cha Muziki wa Kihierarkia ya MusicLM

Kagua tokeni za daraja la MusicLM, hali ya melodia, upatanishi wa maandishi-sauti ya MuLan, na nyenzo ya tathmini ya MusicCaps.

Njia za mwongozo zinazohusianaKizazi cha Muziki wa Kihierarkia cha Musiclm
Swali 1 ya 8

How does MusicLM formulate conditional music generation?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. How does MusicLM formulate conditional music generation?
    • As hierarchical sequence-to-sequence modeling over audio tokens
    • As painting pixels with a diffusion model
    • As concatenating fixed prerecorded loops
    • As simulating instrument mechanics
  2. Ni nini jukumu la ishara za semantic katika MusicLM?
    • Wanahifadhi metadata ya faili ya wimbo
    • Hunasa muundo wa hali ya juu kama vile wimbo na mdundo
    • Wanatafsiri maneno katika lugha zingine
    • Wanadhibiti kiasi cha pato
  3. Ni sehemu gani hutoa maelezo mazuri ya akustisk kama timbre na muundo?
    • Nakala ya haraka moja kwa moja
    • Hatua ya kujenga upya Griffin-Lim
    • Tokeni za akustisk za SoundStream neural codec
    • Kitabu cha sheria cha nadharia ya muziki kilichoandikwa kwa mkono
  4. Je, MusicLM inaunganisha vipi arifa ya maandishi kwa sauti ya muziki?
    • Kwa kulinganisha maneno muhimu dhidi ya vichwa vya nyimbo
    • Kwa kumwomba mwanadamu aweke lebo kila klipu
    • Kupitia API tofauti ya tafsiri
    • Kupitia MuLan, nafasi ya pamoja ya kupachika maandishi ya muziki
  5. Kwa nini muundo wa daraja, uliopangwa kwa hatua husaidia na muundo wa masafa marefu?
    • Inaruhusu mfano wa ishara mbaya, polepole wa muda mrefu bila urefu mkubwa wa mlolongo
    • Huondoa hitaji la data yoyote ya mafunzo
    • Hucheza wimbo haraka ili kutoshea zaidi
    • Inazalisha sekunde tatu tu kwa wakati mmoja
  6. What does the MusicLM paper report about its generated music duration and sample rate?
    • Every output is exactly 30 seconds at 8 kHz
    • Its experiments report 24 kHz music that can remain consistent over several minutes
    • Only one-second clips at 48 kHz
    • Uncompressed MIDI of unlimited duration
  7. What additional input can MusicLM use alongside text to guide generation?
    • A spectrogram-only image caption
    • A microphone-array steering vector
    • A hummed or whistled melody
    • A source-code repository
  8. What did the MusicLM paper release as MusicCaps?
    • A collection of 5,500 synthesizer presets
    • A speech transcription benchmark
    • A set of SoundStream model weights only
    • 5,500 music-text pairs with human-written descriptions