ontolearn.nces_trainer
======================

.. py:module:: ontolearn.nces_trainer

.. autoapi-nested-parse::

   Trainer for NCES instances



Classes
-------

.. autoapisummary::

   ontolearn.nces_trainer.NCESTrainer


Functions
---------

.. autoapisummary::

   ontolearn.nces_trainer.before_pad


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

.. py:function:: before_pad(arg)

.. py:class:: NCESTrainer(synthesizer, epochs=300, batch_size=128, learning_rate=0.0001, tmax=20, eta_min=1e-05, clip_value=5.0, num_workers=8, storage_path='./')

   Trainer for neural class expression synthesizers, i.e., NCES, NCES2, ROCES.


   .. py:attribute:: synthesizer


   .. py:attribute:: epochs
      :value: 300



   .. py:attribute:: batch_size
      :value: 128



   .. py:attribute:: learning_rate
      :value: 0.0001



   .. py:attribute:: tmax
      :value: 20



   .. py:attribute:: eta_min
      :value: 1e-05



   .. py:attribute:: clip_value
      :value: 5.0



   .. py:attribute:: num_workers
      :value: 8



   .. py:attribute:: storage_path
      :value: './'



   .. py:method:: compute_accuracy(prediction, target)
      :staticmethod:



   .. py:method:: get_optimizer(model, emb_model=None, optimizer='Adam')


   .. py:method:: get_data_idxs()


   .. py:method:: get_er_vocab()


   .. py:method:: show_num_trainable_params(synthesizer)
      :staticmethod:



   .. py:method:: collate_batch(batch)


   .. py:method:: map_to_token(idx_array)


   .. py:method:: train_step(batch, model, emb_model, optimizer, device, triples_dataloader=None)


   .. py:method:: train(data, shuffle_examples=False, example_sizes=None, save_model=True, optimizer='Adam', record_runtime=True)