Write a NN from Scratch you can use python and numpy of course but no library to calculate the backpropagation.

Implement the simple architecture shown on image 1, forward and backward pass, plot the loss function over time. The bias are all 0.3.

output will w11 wool input 11 ji k1 01 wil2 wi123 12 wizji wi212 w20 k2 ol wi3j1 13 B k3 03 w kal w33 Relu Relu Softmax 0.1 0.2 0.3 Wixjx = 0.3 0.2 0.7 0.4 0.3 0.9 0.2 0.3 0.5 Wjxkx = 0.3 0.5 0.7 0.6 0.4 0.8 0.1 0.4 0.8 Wkxox = 0.3 0.7 0.2 0.5 0.2 0.9 Input=[0.1 0.2 0.7). Desired target=[1.0 0.0 0.0]