karpathy / nn-zero-to-hero

Neural Networks: Zero to Hero
MIT License
10.9k stars 1.33k forks source link

Correct loss calc from SSE to MSE in backprop loop #20

Open sfuller14 opened 1 year ago

sfuller14 commented 1 year ago

SSE is used for loss rather than MSE (~1:54:30 in the video lecture)

Using len(ys) for n -- MSE = SSE / n