Pass raw images without ToTensor()

1

Hello I want to pass raw images(0-255) to my CNN.

I am not using ToTensor().

img1 = self.loader(os.path.join(self.root1, path1))
img1 = torch.from_numpy(np.asarray(img1)).float()

I got a warning.
UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program.

What it mean, does it will affect CNN in any manner?

2

No, it shouldn’t affect the model training as long as you don’t try to write to the underlying numpy array.
You could create a copy of the numpy array to get rid of this warning, since from_numpy will share the same data between the tensor and np.array.

3

I did this

img1 = self.loader(os.path.join(self.root1, path1))
img1 = np.asarray(img1)
img1 = torch.from_numpy((img1)).float()

Still got the warning

4

np.asarray is also using the reference, if no copy is needed.
From the docs:

Array interpretation of a . No copy is performed if the input is already an ndarray with matching dtype and order. If a is a subclass of ndarray, a base class ndarray is returned.

5 
img1 = self.loader(os.path.join(self.root1, path1))
img1 = torch.from_numpy(1.0 * np.asarray(img1)).float()