APPLICATION OF SUPERVISED LEARNING
DEEP LEARNING
|
Question
[CLICK ON ANY CHOICE TO KNOW THE RIGHT ANSWER]
|
True or False:Dot-product attention with softmax activation can be thought of as a soft form of dictionary lookup over matrices
|
|
True
|
|
|
False
|
|
|
Either A or B
|
|
|
None of the above
|
Explanation:
Detailed explanation-1: -In summary, self-attention allows a transformer model to attend to different parts of the same input sequence, while attention allows a transformer model to attend to different parts of another sequence.
Detailed explanation-2: -Attention is an interface connecting the encoder and decoder that provides the decoder with information from every encoder hidden state. With this framework, the model is able to selectively focus on valuable parts of the input sequence and hence, learn the association between them.
There is 1 question to complete.