ontolearn.nero_architectures
============================

.. py:module:: ontolearn.nero_architectures


Classes
-------

.. autoapisummary::

   ontolearn.nero_architectures.MAB
   ontolearn.nero_architectures.SAB
   ontolearn.nero_architectures.ISAB
   ontolearn.nero_architectures.PMA
   ontolearn.nero_architectures.SetTransformer
   ontolearn.nero_architectures.DeepSet
   ontolearn.nero_architectures.SetTransformerNet


Module Contents
---------------

.. py:class:: MAB(dim_Q, dim_K, dim_V, num_heads, ln=False)

   Bases: :py:obj:`torch.nn.Module`


   Multi-head Attention Block.


   .. py:attribute:: dim_V


   .. py:attribute:: num_heads


   .. py:attribute:: fc_q


   .. py:attribute:: fc_k


   .. py:attribute:: fc_v


   .. py:attribute:: fc_o


   .. py:method:: forward(Q, K)


.. py:class:: SAB(dim_in, dim_out, num_heads, ln=False)

   Bases: :py:obj:`torch.nn.Module`


   Self-Attention Block.


   .. py:attribute:: mab


   .. py:method:: forward(X)


.. py:class:: ISAB(dim_in, dim_out, num_heads, num_inds, ln=False)

   Bases: :py:obj:`torch.nn.Module`


   Induced Self-Attention Block.


   .. py:attribute:: I


   .. py:attribute:: mab0


   .. py:attribute:: mab1


   .. py:method:: forward(X)


.. py:class:: PMA(dim, num_heads, num_seeds, ln=False)

   Bases: :py:obj:`torch.nn.Module`


   Pooling by Multihead Attention.


   .. py:attribute:: S


   .. py:attribute:: mab


   .. py:method:: forward(X)


.. py:class:: SetTransformer(dim_input, num_outputs, dim_output, num_inds=32, dim_hidden=128, num_heads=4, ln=False)

   Bases: :py:obj:`torch.nn.Module`


   Set Transformer architecture.


   .. py:attribute:: enc


   .. py:attribute:: dec


   .. py:method:: forward(X)


.. py:class:: DeepSet(num_instances: int, num_embedding_dim: int, num_outputs: int)

   Bases: :py:obj:`torch.nn.Module`


   DeepSet neural architecture for set-based learning.


   .. py:attribute:: name
      :value: 'DeepSet'



   .. py:attribute:: num_instances


   .. py:attribute:: num_embedding_dim


   .. py:attribute:: num_outputs


   .. py:attribute:: embeddings


   .. py:attribute:: fc0


   .. py:attribute:: fc1


   .. py:method:: forward(xpos, xneg)


   .. py:method:: positive_expression_embeddings(tensor_idx_individuals: torch.LongTensor)


   .. py:method:: negative_expression_embeddings(tensor_idx_individuals: torch.LongTensor)


.. py:class:: SetTransformerNet(num_instances: int, num_embedding_dim: int, num_outputs: int)

   Bases: :py:obj:`torch.nn.Module`


   Set Transformer based architecture.


   .. py:attribute:: name
      :value: 'ST'



   .. py:attribute:: num_instances


   .. py:attribute:: num_embedding_dim


   .. py:attribute:: num_outputs


   .. py:attribute:: embeddings


   .. py:attribute:: set_transformer_negative


   .. py:attribute:: set_transformer_positive


   .. py:method:: forward(xpos, xneg)