A data scientist implements a convolutional neural network in SAS DLPy for image classification. After adding Conv2D and pooling layers, they add a Flatten layer. What is the purpose of the Flatten layer?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because a Flatten layer reshapes the multi-dimensional output of convolutional/pooling layers (height × width × channels) into a single 1D feature vector, enabling the output to be passed to Dense (fully connected) layers that expect 1D input. A describes Batch Normalization.
Full explanation below image
Full Explanation
B is correct because a Flatten layer reshapes the multi-dimensional output of convolutional/pooling layers (height × width × channels) into a single 1D feature vector, enabling the output to be passed to Dense (fully connected) layers that expect 1D input. A describes Batch Normalization. C describes Global Average Pooling. D describes Dropout layers.