objectscope.evaluator
=====================

.. py:module:: objectscope.evaluator


Classes
-------

.. autoapisummary::

   objectscope.evaluator.Evaluator


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

.. py:class:: Evaluator(cfg, test_data_name, roi_heads_score_threshold=0.5, tasks=('bbox', ), output_dir='output/object_detector', **kwargs)

   Bases: :py:obj:`object`


   .. py:attribute:: cfg


   .. py:attribute:: test_data_name


   .. py:attribute:: roi_heads_score_threshold
      :value: 0.5



   .. py:attribute:: output_dir
      :value: 'output/object_detector'



   .. py:attribute:: evaluator


   .. py:attribute:: dataset_nm


   .. py:attribute:: metadata


   .. py:method:: get_model_paths(cfg=None)


   .. py:method:: evaluate_models(cfg=None, model_paths=None, roi_heads_score_threshold=None) -> pandas.DataFrame


   .. py:method:: get_best_model(eval_df=None, metric='AP50')


   .. py:method:: plot_evaluation_results(df: Union[pandas.DataFrame, None] = None, metric='AP50', labels={'AP50': 'Average Precision at IoU=0.5', 'model_name': 'Model Name'})


   .. py:method:: evaluate_confidence_thresholds(thresholds: list, cfg=None) -> pandas.DataFrame


   .. py:method:: get_best_threshold(threshold_df: Union[pandas.DataFrame, None] = None, metric='AP50', thresholds: Union[List, None] = None) -> dict


   .. py:method:: plot_random_samples(n=3)


   .. py:method:: __call__()