Consider the vision transformer model to classify an image. The image size is $(512512).$ The image is split into a total of $88=64$ patches.

Each patch will then be vectorized and mapped to an embedding vector. The embedding dimension is $12.$ In the transformer; the number

of heads is $1,$ number of blocks is $1,$ output dimension is $2.$

$($a$)$ What is the size of each patch?

$($b$)$ What is the dimension of the embedding matrix?

$($c$)$ What is the dimension of the query matrix?

$($d$)$ What is the dimension of the attention matrix?