The video “Using a convolutional network” shows a model with a multi-dimensional Embedding layer followed by a 1 dimensional Convolutional layer. How does the 1 dimensional Convolutional layer calculate the convolutions across the multi-dimensional output of the Embedding layer?
Or in other words, what happens to the third dimension (16 in the video) of the output shape of the Embedding layer?
Sir you can refer this link for more clear explanation
What I feel is output shape of embedding layer will be of form (input_length, output_dim), where input_length = max_length and output_dim = embedding_dim
Conv considers the last dimension as the number of channels and will output the last dimension equal to the number of filters you specify when configuring the filter.
1 Like