fxia22/pointnet.pytorch

Architecture: Very small issue

timothylimyl opened this issue · 2 comments

timothylimyl commented

Hi Fei Xia,

I was wondering whether this line (14-15) of code in the PointNet Architecture is missing an extra layer as of the official paper release:

class STN3d(nn.Module):
    def __init__(self):
        super(STN3d, self).__init__()
        self.conv1 = torch.nn.Conv1d(3, 64, 1)
        *---HERE-> = torch.nn.Conv1d(64,64,1)*
        self.conv2 = torch.nn.Conv1d(64, 128, 1)

Not a very big deal but just want to clarify on this.

Cheers,
Tim

chrispduck commented

I believe this layer is missing - see line 33 of original TensorFlow implementation here

janotjir commented

Yes, I think it is missing too. But if I got it right, instead in STN3d (T-Net) architecture, which consists of MLP(64, 128, 1024), I would add it to the class PointNetfeat

  • 👍2