Softmax Regression Implementation from Scratch

English Version pytorch
44

I’m having hard time interpreting the last part of the code.

Why do we use 'wrong = preds.type(y.dtype) != y’ instead of ‘wrong = preds != y’, which generate the exact same result.
And why does the first one work? The y.dtype just generates the type of y.

45

Answer to the funniest question:
Is it always a good idea to return the most likely label? For example, would you do this for medical diagnosis? How would you try to address this?

@d2l.add_to_class(SoftmaxRegression)
def show_top_dimensions(self, X, dims=1, labels=None):
    y_hat = self.forward(X)
    topk = torch.topk(y_hat, dims)
    for i in range(dims):
        rank = float(topk[0][0][i])
        ind = int(topk[1][0][i])
        print(ind, labels([ind]), rank) 
    print()
    
g = iter(data.val_dataloader())
X, y = next(g)
cols = 2
for i in range(cols):
    model.show_top_dimensions(X[i], 3, data.text_labels)
data.visualize((X, y), ncols=cols)

image