CODEWITHSUNDEEP
pythonpytorch
canonball
canonball

Reputation: 515

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

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

Upvotes: 2

Views: 779

Answers (1)

Olivier Melançon
Olivier Melançon

Reputation: 22324

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.

Upvotes: 5

Related Questions