def forward(self, x):
'''
:param x: b,n,f,ssn,srn
:return:
'''
b, n, f, ssn, srn = x.shape
assert (ssn == 5 and srn == 5)
x = x.reshape(b * n, f, ssn, srn)
x1 = self.network1_embed1_relu(
self.network1_embed1(x) + self.network1_embed1_short(x))
x1 = self.network1_embed2_relu(
self.network1_embed2(x1) + self.network1_embed2_short(x1))
x1 = self.network1_embed3(x1)
x2 = self.network2_embed1_relu(
self.network2_embed1(x) + self.network2_embed1_short(x))
x2 = self.network2_embed2_relu(
self.network2_embed2(x2) + self.network2_embed2_short(x2))
x2 = self.network2_embed3(x2)
x1 = x1.reshape(b * n, 8, 25)
x2 = x2.reshape(b * n, 16, 25).permute(0, 2, 1) # b*n,25,16
x = torch.bmm(x1, x2).reshape(b * n, 128) # b*n,8,25
assert (x.shape[1] == 128)
x = x.reshape(b, n, 128)
x = l2_normalize(x, axis=2)
return x
def forward(self, x):
'''
:param x: b,n,f,ssn,srn
:return:
'''
x = l2_normalize(x, axis=2)
return x[:, :, :, 0, 0]
def forward(self, x):
x = self.conv1(self.conv0(x))
x = l2_normalize(x, axis=1)
return x