Maswali yanayounganishwa na mwongozo · Kati Kiwango

Maswali ya Kawaida ya Mitandao ya Neural

Kagua hali zinazojirudia, pembejeo zilizoagizwa, changamoto za upinde rangi, milango, usawa wa mafunzo, masasisho ya nambari na vikomo vya kumbukumbu za mafunzo.

Njia za mwongozo zinazohusianaMitandao ya Neural ya Kawaida
Swali 1 ya 8

Je, ni taarifa gani ya mara kwa mara ambayo RNN ya kawaida ya unidirectional inachanganya na kila ingizo jipya?

Hakiki kila swali8 maswali

Majibu na maelezo huonekana unapojibu maswali.

  1. Je, ni taarifa gani ya mara kwa mara ambayo RNN ya kawaida ya unidirectional inachanganya na kila ingizo jipya?
    • Only the previous printed prediction with no internal state
    • A fresh unrelated dataset at every step
    • The previous hidden state
    • No parameters or intermediate information
  2. Je, hali iliyofichwa ya RNN inapaswa kuelewekaje?
    • An evolving internal representation that may summarize earlier inputs
    • A guaranteed lossless archive of every earlier input
    • The network's entire training dataset on disk
    • Only a final class label with no intermediate role
  3. Ni ugumu gani wa mafunzo unaweza kutokea kutokana na kueneza gradient za RNN kupitia hatua nyingi za wakati?
    • A guarantee that every hidden state grows without bound
    • Vanishing or exploding gradient magnitudes
    • An automatic change from speech to image inputs
    • Permanent deletion of the training objective
  4. Je, LSTM na GRU huboresha vipi kwenye RNN wazi?
    • Wanaondoa kumbukumbu zote kutoka kwa mtandao
    • Wanafanya kazi kwa nambari tu, sio maandishi
    • Wanafundisha bila data yoyote
    • Wanatumia milango kudhibiti ni taarifa gani ya kuweka, kusasisha au kutupa
  5. Je, ni aina gani ya data ambazo RNN zimeundwa hasa?
    • Data mfuatano kama vile maandishi, hotuba na mfululizo wa saa
    • Picha za utulivu bila agizo
    • Nambari zisizopangwa bila mpangilio
    • Mipango ya ujenzi wa pande tatu pekee
  6. What training-computation difference does the original transformer architecture offer compared with a conventional sequential RNN?
    • Guaranteed fully parallel autoregressive generation
    • A guarantee that no training examples are needed
    • More parallel computation across training positions and direct attention interactions
    • A model with no learned parameters
  7. In the stated linear recurrent illustration h-new equals one-half h-previous plus x, what is h-new for h-previous two and x three?
    • Tano
    • Moja
    • Saba
    • Nne
  8. Why does a fixed-size RNN inference state not establish constant memory for full-sequence training?
    • Training never uses intermediate values
    • Backpropagation through time may store or recompute information across many steps
    • Every model deletes its parameters before learning
    • A fixed state guarantees every historical input is retained exactly