Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Is-beddellada Kala-guurka ah ee qoto-dheer

Codso kala qaybsanaanta si qoto dheer iyo micnaha leh, is barbar dhig kharashyada isdhaafsiga, xisaabi tusaale yar, oo aqoonso nashqada qaabka moobilka oo hufan.

Wadooyinka haga ee la xidhiidhaKacdoonnada kala sooci kara ee qotodheer
Su'aal 1 ee 8

Waa maxay labada tillaabo ee kacaanka si qoto dheer u kala go'a u kala qaybiyo convolution-ga?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay labada tillaabo ee kacaanka si qoto dheer u kala go'a u kala qaybiyo convolution-ga?
    • Tallaabo isku-dubbarid oo ay ku xigto tallaabo caadiya
    • Laba isugeyn oo 3x3 ah oo leh ReLU dhexdooda
    • Shaandheeye meel qotodheer u qoto dheer halkii kanaal, ka dibna isku dhafka kanaalka 1x1 micnaha leh
    • Isku-dhafan cod-bixiye iyo fur-fure
  2. Talaabada qoto dheer, sidee loo daaweeyaa kanaalada wax gelinta?
    • Kanaal walba si madaxbanaan ayaa loo sifeeyaa kernel u gaar ah
    • Dhammaan kanaalada marka hore waa la soo koobay
    • Kanaalada si aan kala sooc lahayn ayaa loo shaandheeyay ka hor inta aan la shaandhayn
    • Kaliya kanaalka ugu horreeya ayaa la isku duubay
  3. For many output channels, what operation-count reduction does the guide associate with a 3×3 depthwise separable convolution?
    • Qiyaastii 2 jeer ayaa ka yar
    • Qiyaastii 100 jeer ka yar
    • Waxay isticmaashaa xisaabin badan, ma yara
    • Qiyaastii 8 ilaa 9 jeer ka yar
  4. Waa maxay doorka rog-rogista (1x1) ee nashqadan?
    • Si loo dhimo xallinta meelaha bannaan
    • In la isku daro macluumaadka kanaalada goob kasta
    • Si loogu daro codaynta booska
    • Si meesha looga saaro shuruudaha eexda
  5. Qaab dhismeedka caanka ah ee si qoto dheer u faafiyey iskudhafyada kala soocida ee aaladaha mobilada?
    • MobileNet
    • ResNet
    • AlexNet
    • LSTM
  6. An implementation applies only the depthwise spatial filters and omits the pointwise layer. Which capability is missing from that block?
    • Spatial filtering within each channel
    • Separate kernels for the input channels
    • Learned combination of information across channels
    • Local spatial feature extraction
  7. Using the guide’s formulas with M = 4 input channels and N = 16 outputs, how many multiply-adds per location do standard and separable 3×3 convolutions require?
    • Standard 100; separable 576
    • Standard 576; separable 100
    • Standard 576; separable 36
    • Standard 64; separable 100
  8. In the guide’s cost ratio 1/N + 1/9 for a 3×3 separable convolution, what happens as the output-channel count N becomes very large?
    • The ratio approaches 9
    • The ratio approaches zero
    • The ratio becomes independent of kernel size for every kernel
    • The ratio approaches 1/9