Guide-yakabatana mibvunzo · Zvakaoma Nhanho

VALL-E uye Codec Mutauro Models Quiz

Ongorora nzira yeVALL-E's codec-token language-model, chizvarwa chayo chechiratidzo, uye mamiriro arikuseri kwezvishoma-kupfura mhedzisiro yezwi.

Nzira dzekutungamira dzinoenderanaVall E Uye Codec Mutauro Models
Mubvunzo 1 ye 8

Ndeipi pfungwa yepakati inosiyanisa VALL-E kubva kumavambo mavara-kune-kutaura masisitimu?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. Ndeipi pfungwa yepakati inosiyanisa VALL-E kubva kumavambo mavara-kune-kutaura masisitimu?
    • Inofanotaura neural codec tokens semutauro wekufananidzira basa pane kufanotaura spectrogram
    • Inoshandisa chete nemaoko masaini kugadzirisa
    • Zvinoda maawa ekuteerera kubva kune inotarirwa mutauri
    • Inogadzira zvinyorwa kubva mukutaura
  2. VALL-E inoda mareferensi akawanda sei kuti igadzire izwi remutauri mutsva?
    • Inenge awa imwe chete
    • Anenge masekonzi matatu
    • Hapana referensi redhiyo zvachose
    • Inenge maminitsi makumi matatu
  3. Ndeipi neural codec inoshandiswa neVALL-E kugadzira maodhiyo tokeni?
    • MP3
    • Opus
    • EnCodec
    • WAV
  4. Sei VALL-E ichishandisa zvese zviri zviviri autoregressive uye isiri-autoregressive nhanho?
    • Kuturikira mutauro mumitauro yakawanda
    • Kuparadzanisa mimhanzi kubva pakutaura
    • Kudzvanya mapindiro emavara
    • Nekuti macodec tokens akaturikidzana, saka inofanotaura coarse tokens zvakateerana uye ma tokeni akanaka anoenderana.
  5. Zvichida kuti ingani data rekutaura rakadzidziswa VALL-E, richigonesa kusimba zero-shot cloning?
    • Anenge maawa 60 000
    • Anenge 100 maawa
    • Anenge maawa 1 000
    • Anenge maawa gumi
  6. Kunze kwetimbre, chii chimwe chinogona kuchengetedza VALL-E kubva papfupi referenzi clip?
    • Mazwi chaiwo emutauri zvachose
    • Iyo acoustic nharaunda yekurekodha
    • Vhidhiyo yakaperekedzwa nayo
    • Iyo yekutanga faira fomati
  7. What does VALL-E condition on to imitate a target speaker in its demonstration setup?
    • A photograph and a written biography only
    • A long labeled transcript from every target-speaker recording
    • A short audio prompt from that speaker plus the text to speak
    • A speaker ID with no audio example
  8. What text representation does the VALL-E paper use as input to its text-to-code model?
    • A raw waveform of the target utterance
    • A speaker’s face embedding
    • A list of MIDI notes
    • A phoneme sequence