Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

Conformer Architecture adanwo

Learn how the Conformer combines convolution and self-attention for speech, including its block design and reported benchmark results.

Jẹmọ itọsọna awọn ọnaConformer faaji
Ibeere 1 ti 8

Awọn ọna meji wo ni faaji Conformer darapọ ni bulọọki kan?

Awotẹlẹ gbogbo ibeere8 ibeere

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

  1. Awọn ọna meji wo ni faaji Conformer darapọ ni bulọọki kan?
    • Convolution ati akiyesi ara ẹni
    • Loorekoore ati pooling
    • Awọn igi ipinnu ati ikojọpọ
    • Iwadi tan ina ati tokenization
  2. Kilode ti iṣakojọpọ convolution ati akiyesi ṣe wulo paapaa fun ọrọ sisọ?
    • Ifarabalẹ n ṣe alaye awọn alaye foonu agbegbe lakoko ti convolution n ṣe itọju ipo agbaye
    • Convolution n gba awọn ilana agbegbe lakoko ti akiyesi ara ẹni n gba ipo ti o gun-gun
    • Awọn mejeeji ṣe pataki fun awọn aworan, kii ṣe ohun
    • Bẹni iranlọwọ; o jẹ odasaka fun iyara
  3. Kini 'Macaron' tabi eto ipanu kan ti Àkọsílẹ Conformer?
    • Awọn ipele ifarabalẹ meji ni ayika iyipada kan
    • Layer kikọ-siwaju ẹyọkan nikan
    • Ifarabalẹ ati ariyanjiyan ti a we laarin awọn modulu ilọsiwaju-idaji-igbesẹ meji
    • Convolution tolera lori oke ti convolution
  4. Ajo wo ni o ṣafihan Conformer, ati ni ọdun wo?
    • OpenAI ni ọdun 2018
    • Meta ni ọdun 2015
    • Microsoft ni ọdun 2012
    • Google ni ọdun 2020
  5. Kini module convolution inu a Conformer ojo melo pẹlu?
    • Iyipada ọlọgbọn kan, GLU kan, iyipada ijinle, isọdọtun, ati imuṣiṣẹ Swish kan
    • Nikan kan nikan ni kikun ti sopọ Layer
    • sẹẹli LSTM loorekoore
    • A ID igbo classifier
  6. On LibriSpeech test-clean/test-other, what WER did the original Conformer paper report with an external language model?
    • 2.1% / 4.3%
    • 1.9% / 3.9%
    • 9.1% / 9.3%
    • 0.1% / 0.3%
  7. What positional information does the Conformer paper add to attention?
    • A geographic GPS coordinate for every speech frame
    • A fixed class label for each utterance
    • Relative positional encoding
    • A separate video frame index
  8. What word error rates did the Conformer paper report without an external language model on LibriSpeech test/test-other?
    • 20.1% / 40.3%
    • 0.1% / 0.3%
    • 12.1% / 14.3%
    • 2.1% / 4.3%