How can I get the Hidden Embeddings from the 2nd last fully connected layer for t-SNE visualization?

how can we do this in pytorch?
After training the model, get the learned representations from the 2nd last FC layer for t-SNE visualization. The Model is defined below

class VisualNet(nn.Module):
    def __init__(self):
        super(VisualNet, self).__init__()
        self.conv1 = nn.Conv2d()  #Ignore the function values
        slef.conv2 = nn.Conv1d()
        slef.conv3 = nn.Conv1d()
        slef.conv4 = nn.Conv1d(512,512)

       self.fc1   = nn.Linear(512,256)
        self.fc2   = nn.Linear(256, 8)

    def forward(self, x):
        x = self.conv1(self.conv2(self.conv3(self.conv4(x)))
        x_fc1 = self.fc1(x)
        x = self.fc2(x_fc1)        

        return x

model = VisualNet()

After finishing Training, How can I get x_fc1 embeddings
for t-SNE ??

Assuming you are looking for the output activations of self.fc1 (unsure what embeddings would mean in this context), you could use forward hooks as described here.