PyTorch torch.bitwise_not() Method

The torch.bitwise_not() method performs a bitwise NOT operation on the elements of an input tensor. A bitwise is a unary operation that inverts each bit in the operand. For example, 0 becomes 1 and 1 becomes 0.

Syntax

torch.bitwise_not(input, out=None)

Parameters

Argument	Description
input (Tensor)	It represents a tensor that must be of type integer (torch.int8, torch.int16, torch.int32, torch.int64) or boolean type (torch.bool).
out (Tensor, optional)	By default, it is None. It is an output tensor to store the result.

Bitwise NOT on Integer Tensors

It inverts all bits in the binary representation of each element.

For signed integers, it uses two’s complement representation.

import torch

# Create an integer tensor
input_tensor = torch.tensor([0, 1, 2, -3], dtype=torch.int32)

bitwise_not_tensor = torch.bitwise_not(input_tensor)

print(input_tensor)
print(bitwise_not_tensor)

# Output:
# tensor([ 0,  1,  2, -3], dtype=torch.int32)
# tensor([-1, -2, -3,  2], dtype=torch.int32)

Here is the explanation:

1. 0 (binary: 00000000) becomes -1 (binary: 11111111, two’s complement).
2. 1 (binary: 00000001) becomes -2 (binary: 11111110).
3. 2 (binary: 00000010) becomes -3 (binary: 11111101).
4. -3 (binary: 11111101) becomes 2 (binary: 00000010).

Boolean Tensors

If your input is a boolean tensor (True or False), bitwise_not flips True to False and vice versa.

import torch

# Create a boolean tensor
bool_tensor = torch.tensor([True, False, True])
output_tensor = torch.bitwise_not(bool_tensor)

print(bool_tensor)
print(output_tensor)

# Output:
# tensor([ True, False,  True])
# tensor([False,  True, False])

Using the “out” Parameter

If you have a pre-allocated tensor, you can use the out parameter that allows us to store the result in a pre-allocated tensor.

import torch

input_tensor = torch.tensor([5, -6], dtype=torch.int32)

output_tensor = torch.empty_like(input_tensor)

torch.bitwise_not(input_tensor, out=output_tensor)

print(input_tensor)
print(output_tensor)

# Output:
# tensor([ 5, -6], dtype=torch.int32)
# tensor([-6,  5], dtype=torch.int32)

Unsupported Tensor Type

If you attempt to use the bitwise_not operator on a floating tensor, it will throw an error.

import torch

float_tensor = torch.tensor([1.0, 2.0])

try:
    torch.bitwise_not(float_tensor)
except RuntimeError as e:
    print("Error:", e)

# Output: Error: "bitwise_not_cpu" not implemented for 'Float'

That’s all!