Forward Propagation | Deep Learning

Live Engine

Select Topic

easyForward Propagation

A network has input shape (batch_size=32, features=128), first layer weight matrix W₁ of shape (128, 64), and bias b₁ of shape (64,). An engineer writes the forward pass as: h = x @ W₁ + b₁. What is the shape of h and how does the bias broadcasting work?

Live Engine

Select Topic

easyForward Propagation

A network has input shape (batch_size=32, features=128), first layer weight matrix W₁ of shape (128, 64), and bias b₁ of shape (64,). An engineer writes the forward pass as: h = x @ W₁ + b₁. What is the shape of h and how does the bias broadcasting work?