The difference in 5 equation in paper and code

[Tismoney]

In original paper the last rule of update node vector is: $v_i^t = (1 - z_{s, i}^t) \odot v_i^{t-1} + z_{s, I}^t \odot v_i^t$. But following the code, the rules is not the same:

SR-GNN/pytorch_code/model.py

Lines 45 to 47 in 90123c8

	inputgate = torch.sigmoid(i_i + h_i)
	newgate = torch.tanh(i_n + resetgate * h_n)
	hy = newgate + inputgate * (hidden - newgate)

Code's rule is: $v_i^t = (1 - z_{s, i}^t) \odot v_i^t + z_{s, I}^t \odot v_i^{t-1}$. The difference is swap v_i^t and v_i^{t-1}.

Is it mistake in paper or in code?

[Tismoney]

As I understand, there is not mistake conceptually. In paper $z$ learns how much information save, but in code it is responsible how much information forget. Am I right?

[yichudu]

As I understand, there is not mistake conceptually. In paper $z$ learns how much information save, but in code it is responsible how much information forget. Am I right?

the readability of the code in this repository is too bad.

[wangzglife]

I agree with you, I also found that this may be a mistake