Bug Report for handwritten-digit-classifier

[bewarb]

Even with the given solution you cannot pass the local test or the submission grader. Using self.eval() or nn.Identity() gets you slightly closer at:

[[0.5522,0.4798,0.5047,0.4956,0.5159,0.5456,0.516,0.4954,0.4085,0.5162]]

Solution Code Output:

Input:

images=torch.randn(1, 28 * 28)

Your Output:

[[0.607,0.4426,0.494,0.4634,0.5263,0.5707,0.5564,0.4835,0.3663,0.5077]]

Expected output:

[[0.565,0.4828,0.5203,0.471,0.5244,0.5394,0.5588,0.4864,0.4046,0.5196]]