SubBoxCNNPooler

One dimension CNN encoding multi-kernel layer. Input embeddings are convoluted using linear kernels each parametrized with a (window) size of kernel_size[kernel_i] The output of the kernels are concatenated together, max-pooled and finally projected to a size of output_size.

Parameters

PARAMETER	DESCRIPTION
pipeline	Pipeline instance TYPE: Pipeline DEFAULT: None
name	Name of the component TYPE: str DEFAULT: 'sub-box-cnn-pooler'
output_size	Size of the output embeddings Defaults to the input_size TYPE: Optional[int] DEFAULT: None
out_channels	Number of channels TYPE: Optional[int] DEFAULT: None
kernel_sizes	Window size of each kernel TYPE: Sequence[int] DEFAULT: (3, 4, 5)
activation	Activation function to use TYPE: ActivationFunction DEFAULT: 'relu'

---