Can I use captum for keypoint detection?

[waynehong666]

Is it possible use captum library for keypoint detection?

[NarineK]

Thank you for the question @waynehong666 ! Can you elaborate more question ? Do you have any example models ?

[waynehong666]

For example, https://www.kaggle.com/c/facial-keypoints-detection problem uses the following cnn model.

class CNN(nn.Module):
def init(self):
super(CNN, self).init()
self.conv1 = nn.Conv2d(in_channels=1, out_channels=4, kernel_size=5) # (b,1,96,96) to (b,4,92,92)
self.conv1_bn = nn.BatchNorm2d(4)
self.conv2 = nn.Conv2d(in_channels=4, out_channels=64, kernel_size=3) # (b,4,46,46) to (b,64,44,44)
self.conv2_bn = nn.BatchNorm2d(64)
self.conv3 = nn.Conv2d(in_channels=64, out_channels=128, kernel_size=3) # (b,64,22,22) to (b,128,20,20)
self.conv3_bn = nn.BatchNorm2d(128)
self.conv4 = nn.Conv2d(in_channels=128, out_channels=256, kernel_size=3) # (b,128,10,10) to (b,256,8,8)
self.conv4_bn = nn.BatchNorm2d(256)
self.fc1 = nn.Linear(25644, 1024)
self.fc2 = nn.Linear(1024, 256)
self.fc3 = nn.Linear(256, 30)
self.dp1 = nn.Dropout(p=0.4)

def forward(self, x, verbose=False):
    # apply conv1, relu and maxpool2d
    x = self.conv1_bn(self.conv1(x))
    x = F.relu(x)
    x = F.max_pool2d(x, kernel_size=2)
    x = self.dp1(x)
    
    # apply conv2, relu and maxpool2d
    x = self.conv2_bn(self.conv2(x))
    x = F.relu(x)
    x = F.max_pool2d(x, kernel_size=2)
    x = self.dp1(x)
    
    # apply conv3, relu and maxpool2d
    x = self.conv3_bn(self.conv3(x))
    x = F.relu(x)
    x = F.max_pool2d(x, kernel_size=2)
    x = self.dp1(x)
    
    # apply conv4, relu and maxpool2d
    x = self.conv4_bn(self.conv4(x))
    x = F.relu(x)
    x = F.max_pool2d(x, kernel_size=2)
    
    # apply dropout
    x = self.dp1(x)
    
    x = x.view(-1, 256*4*4)
    
    # now use FC layer with relu
    x = self.fc1(x)
    x = F.relu(x)
    x = self.dp1(x)
    x = self.fc2(x)
    x = F.relu(x)
    x = self.dp1(x)
    x = self.fc3(x)
    return x

[sycophant-stone]

waiting for more info of Captum

[wwdok]

Is there any news ? I also want this feature !