torch.eq() problem -> **Why are the last two rows like this? They are all **True**, but the labels are0, 0, 1, 0, 1`.**

[CYBki]

`y_pred_probs = torch.sigmoid(y_logits)
y_pred_probs

tensor([[0.4916],
[0.4923],
[0.5223],
[0.4827],
[0.5736]], device='cuda:0')`

`torch.round(y_pred_probs)

tensor([[0.],
[0.],
[1.],
[0.],
[1.]], device='cuda:0')`

`# Find the predicted labels
y_preds = torch.round(y_pred_probs)

In full (logits -> pred probs -> pred labels)

y_pred_labels = torch.round(torch.sigmoid(model_0(X_test.to(device))[:5]))

Check for equality

print(torch.eq(y_preds.squeeze(), y_pred_labels.squeeze()))

Get rid of extra dimension

y_preds.squeeze()
tensor([True, True, True, True, True], device='cuda:0')
tensor([0., 0., 1., 0., 1.], device='cuda:0')
**Why are the last two rows like this? They are all **True**, but the labels are0, 0, 1, 0, 1`.**