eq, view_as

• eq

 

It  stands for "equal" - it performs element-wise equality comparison between tensors.

tensor1.eq(tensor2)  # Returns True/False for each element

• Compares each element in tensor1 with corresponding element in tensor2
• Returns a boolean tensor with True where elements are equal, False otherwise

---

For example,

 

predicted = torch.tensor([1, 2, 3, 4, 5])
targets = torch.tensor([1, 0, 3, 0, 5])

result = predicted.eq(targets)
# → tensor([True, False, True, False, True])

correct = result.sum()
# → 3 (three correct predictions)

---

• view_as

 

It reshapes a tensor to have the same shape as another tensor.

tensor1.view_as(tensor2)  # Reshape tensor1 to match tensor2's shape

• Automatically reshapes tensor1 to have the same dimensions as tensor2
• Equivalent to tensor1.view(tensor2.shape)
• Does NOT copy data, just changes how it's viewed

---

For example,

 

predicted = torch.tensor([1, 2, 3, 4])      # Shape: (4,)
targets = torch.tensor([[1], [2], [3], [4]])  # Shape: (4, 1)

# Reshape targets to match predicted
targets_reshaped = targets.view_as(predicted)  # Shape: (4,)
# → tensor([1, 2, 3, 4])

# Now comparison works properly
result = predicted.eq(targets_reshaped)