Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Mimi ya Kutiririsha Kodeki ya Sauti

Kagua tokeni za Mimi za 12.5 Hz, vitabu vya msimbo vya kisemantiki na akustika, RVQ, utiririshaji wa chanzo, na jinsi kodeki yake inavyotumia mazungumzo ya Moshi.

Njia za mwongozo zinazohusianaCodec ya Sauti ya Mimi Streaming
Swali 1 ya 8

Ni maabara gani ilitoa Mimi na mtindo wa sauti wa Moshi?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Ni maabara gani ilitoa Mimi na mtindo wa sauti wa Moshi?
    • Kyutai
    • OpenAI
    • DeepMind
    • AI ya Meta
  2. Takriban ni tokeni ngapi kwa sekunde ambazo Mimi hutoa kwa hotuba?
    • Takriban 200
    • Takriban 12.5
    • Takriban 1,000
    • Takriban 16,000
  3. What information does Mimi’s first semantic codebook aim to represent?
    • Only the waveform’s stereo channel count
    • A transcript copied verbatim from a separate recognizer
    • Semantic speech information distilled from WavLM representations
    • The exact speaker identity with no acoustic tokens
  4. Kwa nini Mimi anaelezewa kama 'kutiririsha' au 'sababu'?
    • Inafanya kazi tu kwenye faili za muziki zilizotiririshwa
    • Inahitaji mkondo wa wingu kufanya kazi
    • Inatiririsha manukuu ya maandishi
    • Inatoa tokeni sauti inapowasili badala ya kusubiri klipu kamili
  5. Je, Mimi hutumia mbinu gani ya kuhesabu kusimba sauti?
    • Ukadiriaji wa vekta iliyobaki (RVQ)
    • Ukandamizaji wa kuzuia JPEG
    • Usimbaji wa urefu wa kukimbia
    • Uchambuzi wa sehemu kuu
  6. What does Mimi’s discrete audio-token representation let Moshi do?
    • Treat audio tokens as ordinary words with identical meanings
    • Model audio continuations as token sequences and decode them back into speech
    • Avoid using an audio decoder
    • Recover the original recording session stems
  7. What is a reported frame size for Mimi’s streaming codec in the Moshi paper?
    • 8 seconds
    • 1 millisecond
    • 80 milliseconds
    • Dakika 10
  8. Which encoder/decoder design does the Moshi report describe for Mimi?
    • A text-only BERT encoder with no audio decoder
    • A fixed lookup table with no learned encoder
    • A bidirectional non-causal decoder that waits for the entire conversation
    • Neural encoder and decoder Transformers with causal masking for streaming