Just enough neural network: a policy is a function with knobs

What happens inside the 33-to-8 network that will drive the robot: neurons, layers, tanh, the forward pass by hand, counting parameters, and gradients. Then a tiny network learns the hand-written trot in plain numpy.

50 min

Questa pagina non è ancora tradotta, quindi è mostrata in inglese.