INNLab

INNLab copied to clipboard

I faced this problem "The algorithm failed to converge because the input matrix is ill-conditioned or has too many repeated singular values (error code: 767)" during the training of JacobianLinear...

fanyuuwang

Zhangyanbo

In `utilities.default_net()`, we have a default activation function (currently, we use `nn.SELU`). Which one works best?

Zhangyanbo

Do we need to add `.detach()` after `var` in `INN.BatchNorm1d`?

2

In `INN.BatchNorm1d`, the forward function is: ```python def forward(self, x, log_p=0, log_det_J=0): if self.compute_p: if not self.training: # if in self.eval() var = self.running_var # [dim] else: # if in...

Zhangyanbo