PyTorch Tensor Operations song & examples

1. tensor.add(tensor2) – Element-wise addition (can also use +).
2. tensor.sub(tensor2) – Element-wise subtraction (can also use -).
3. tensor.mul(tensor2) – Element-wise multiplication (can also use *).
4. tensor.div(tensor2) – Element-wise division (can also use /).
5. torch.mm(tensor1, tensor2) – Matrix multiplication.
6. tensor.sum() – Returns the sum of all elements.
7. tensor.mean() – Returns the mean of all elements.
8. tensor.max() – Returns the maximum value.
9. tensor.min() – Returns the minimum value.
10. torch.cat(tensors, dim) – Concatenates a sequence of tensors along a given dimension.
11. torch.stack(tensors) – Stacks tensors along a new dimension.

Here are examples for the basic Tensor Operations in PyTorch:

1. tensor.add(tensor2) (Element-wise addition)

import torch

tensor1 = torch.tensor([1, 2, 3])
tensor2 = torch.tensor([4, 5, 6])
result = tensor1.add(tensor2)
print(result)

Output:

tensor([5, 7, 9])

Alternatively, you can use +:

result = tensor1 + tensor2
print(result)

---

2. tensor.sub(tensor2) (Element-wise subtraction)

result = tensor1.sub(tensor2)
print(result)

Output:

tensor([-3, -3, -3])

Alternatively, you can use -:

result = tensor1 - tensor2
print(result)

---

3. tensor.mul(tensor2) (Element-wise multiplication)

result = tensor1.mul(tensor2)
print(result)

Output:

tensor([ 4, 10, 18])

Alternatively, you can use *:

result = tensor1 * tensor2
print(result)

---

4. tensor.div(tensor2) (Element-wise division)

result = tensor1.div(tensor2)
print(result)

Output:

tensor([0.2500, 0.4000, 0.5000])

Alternatively, you can use /:

result = tensor1 / tensor2
print(result)

---

5. torch.mm(tensor1, tensor2) (Matrix multiplication)

tensor1 = torch.tensor([[1, 2], [3, 4]])
tensor2 = torch.tensor([[5, 6], [7, 8]])
result = torch.mm(tensor1, tensor2)
print(result)

Output:

tensor([[19, 22],
        [43, 50]])

---

6. tensor.sum() (Sum of all elements)

tensor = torch.tensor([1, 2, 3])
result = tensor.sum()
print(result)

Output:

tensor(6)

---

7. tensor.mean() (Mean of all elements)

tensor = torch.tensor([1.0, 2.0, 3.0])
result = tensor.mean()
print(result)

Output:

tensor(2.)

---

8. tensor.max() (Maximum value)

tensor = torch.tensor([1, 2, 3])
result = tensor.max()
print(result)

Output:

tensor(3)

---

9. tensor.min() (Minimum value)

tensor = torch.tensor([1, 2, 3])
result = tensor.min()
print(result)

Output:

tensor(1)

---

10. torch.cat(tensors, dim) (Concatenation along a given dimension)

tensor1 = torch.tensor([[1, 2], [3, 4]])
tensor2 = torch.tensor([[5, 6], [7, 8]])
result = torch.cat((tensor1, tensor2), dim=0)  # Concatenate along rows
print(result)

Output:

tensor([[1, 2],
        [3, 4],
        [5, 6],
        [7, 8]])

Concatenating along columns (dim=1):

result = torch.cat((tensor1, tensor2), dim=1)
print(result)

Output:

tensor([[1, 2, 5, 6],
        [3, 4, 7, 8]])

---

11. torch.stack(tensors) (Stack tensors along a new dimension)

tensor1 = torch.tensor([1, 2])
tensor2 = torch.tensor([3, 4])
result = torch.stack((tensor1, tensor2), dim=0)  # Stacks tensors into a 2D array
print(result)

Output:

tensor([[1, 2],
        [3, 4]])

Alternatively, stack along a new dimension:

result = torch.stack((tensor1, tensor2), dim=1)
print(result)

Output:

tensor([[1, 3],
        [2, 4]])

---

These examples demonstrate basic tensor operations in PyTorch, which are fundamental for tensor manipulation, especially in deep learning.