It is known that the total parameter of a certain layer of the fully connected neural network is 330, The number of neurons in the previous layer and this layer May be?

Question

Options

• A32 with 10
• B10 with 33
• C33 with 10
• D9 with 33

• A
  11% (5)
• B
  73% (33)
• D
  16% (7)

Explanation

In a fully connected layer, the number of weight parameters equals n_prev × n_curr (previous layer neurons × current layer neurons). Options B (10 × 33 = 330) and C (33 × 10 = 330) both satisfy this equation, making them correct - note that B and C are mathematically equivalent, just with the layer sizes swapped in the description.

Option A (32 × 10 = 320) falls short by 10, and Option D (9 × 33 = 297) falls short by 33 - neither product reaches 330.

Memory tip: Think of the weight matrix as a grid - rows × columns = total weights. You just need to find which pair of factors multiplies to the target number (here, 330 = 10 × 33). If bias terms were included, the formula would be n_prev × n_curr + n_curr; since only weights are counted here, simply factor the total parameter count to find valid layer size pairs.

No community discussion yet for this question.