Classification Visualization plots a transposed image

[prateekgargX]

Tutorial: 2
Introduction to PyTorch

Bug Description
the function to visualize classification boundary, visualize_classification(model, data, label) plots a transposed image. To fix this, meshgrid indexing has to be changed from 'ij' to 'xy'

To Reproduce
Run the notebook on a dataset which is not symmetric in x1 and x2. For example moons dataset on sklearn.

    def generate_continuous_xor(self):

        X,y = make_moons(n_samples=self.size,random_state=42)
        X = np.float32(X)
        y = np.float32(y)
        mean = np.mean(X, axis=0)
        std = np.std(X, axis=0)
        X = (X - mean) / std

        data = torch.from_numpy(X) #torch.randint(low=0, high=2, size=(self.size, 2), dtype=torch.float32)
        label = torch.from_numpy(y).to(torch.long)
        # To make it slightly more challenging, we add a bit of gaussian noise to the data points.
        data += self.std * torch.randn(data.shape)

        self.data = data
        self.label = label

Screenshots
Earlier:

New:

Runtime environment:

• Local computer and Google Colab
• both on CPU only or GPU

This hasn't been noticed till now since original dataset was symmetric in x1 and x2