Two dimensional data regression using pytorch

"""
View more, visit my tutorial page: https://morvanzhou.github.io/tutorials/
My Youtube Channel: https://www.youtube.com/user/MorvanZhou


Dependencies:
torch: 0.1.11
matplotlib
"""
import torch
from torch.autograd import Variable
import torch.nn.functional as F
import matplotlib.pyplot as plt
import numpy as np


torch.manual_seed(1)    # reproducible


# x = torch.unsqueeze(torch.linspace(-1, 1, 100), dim=1)  # x data (tensor), shape=(100, 1)
# y = x.pow(2) + 0.2*torch.rand(x.size())                 # noisy y data (tensor), shape=(100, 1)


x = np.array([[0.7060,0.8235,0.4387],[0.0318,0.6948,0.3816],[0.2769,0.3171,0.7655],[0.0462,0.9502,0.7952]],dtype=np.float32)
y = np.array([[1.8447],[1.2161],[1.4053],[1.7721]],dtype=np.float32)


# torch can only train on Variable, so convert them to Variable
x, y = Variable(torch.from_numpy(x)), Variable(torch.from_numpy(y))


# plt.scatter(x.data.numpy()[:,0], y.data.numpy())
# plt.show()




class Net(torch.nn.Module):
    def __init__(self, n_feature, n_hidden1, n_hidden2, n_output):
        super(Net, self).__init__()
        self.hidden1 = torch.nn.Linear(n_feature, n_hidden1)   # hidden layer
        self.hidden2 = torch.nn.Linear(n_hidden1, n_hidden2)
        self.predict = torch.nn.Linear(n_hidden2, n_output)   # output layer


    def forward(self, x):
        x = F.relu(self.hidden1(x))      # activation function for hidden layer
        x = F.relu(self.hidden2(x))
        x = self.predict(x)             # linear output
        return x


net = Net(n_feature=3, n_hidden1=5, n_hidden2=5, n_output=1)     # define the network
print(net)  # net architecture


optimizer = torch.optim.SGD(net.parameters(), lr=0.01)
loss_func = torch.nn.MSELoss()  # this is for regression mean squared loss


plt.ion()   # something about plotting


for t in range(10000):
    prediction = net(x)     # input x and predict based on x


    loss = loss_func(prediction, y)     # must be (1. nn output, 2. target)


    optimizer.zero_grad()   # clear gradients for next train
    loss.backward()         # backpropagation, compute gradients
    optimizer.step()        # apply gradients
    print(loss.data.numpy())
    if t % 5 == 0:
        # plot and show learning process
        plt.cla()
        plt.scatter(x.data.numpy()[:,1], y.data.numpy())
        plt.plot(x.data.numpy()[:,1], prediction.data.numpy(), 'r-', lw=5)
        plt.text(0.5, 0, 'Loss=%.4f' % loss.data[0], fontdict={'size': 20, 'color':  'red'})
        plt.pause(0.1)


plt.ioff()
plt.show()