Q-Former Query Tokens

Problem

Count the learned query embedding parameters of a Q-Former: num_queries * d_model, since each query token is a free vector.

The signature is qformer_query_params(num_queries, d_model). Return an integer.

Hint

A small set of queries compresses a variable-length image into fixed-length features.

case 1

in: [32,768]

expected: 24576

case 2

in: [1,1]

expected: 1

case 3

in: [64,1024]

expected: 65536

case 4

in: [0,512]

expected: 0

Starter codePython

def qformer_query_params(num_queries, d_model):
    # Your code here
    pass

4 test cases · run your code against them in the editor