Why does the hash value of models change everytime I save it?

Question

I am using torch.save() to save a model file. However, everytime I save it, it changes. Why so?

netG_1 = torch.load('netG.pth')
netG_2 = torch.load('netG.pth')

torch.save(netG_1, 'netG_1.pth')
torch.save(netG_2, 'netG_2.pth')

Using md5sum *.pth :

779f0fefca47d17a0644033f9b65e594  netG_1.pth
476f502ec2d1186c349cdeba14983d09  netG_2.pth
b0ceec8ac886a11b79f73fc04f51c6f9  netG.pth

The model is an instance of this class:

https://github.com/taoxugit/AttnGAN/blob/master/code/model.py#L397

Answer 1

A class which does not define the __hash__ method will have its instances hashed according to their id . As for CPython, this means everytime the instance is saved and reloaded, it changes hash since its position in memory changed.

Here is a proof of concept.

class Foo:
    pass

instance = Foo()

print('hash:', hex(hash(instance)))
print('id:  ', hex(id(instance)))

Output

hash: 0x22f3f57608
id:   0x22f3f576080

The exact transformation is hash(o) == id(o) // 16 .