nnpd.nre.training
=================

.. py:module:: nnpd.nre.training

.. autoapi-nested-parse::

   Readable BCE training, including exact global-batch sharding for DDP.



Functions
---------

.. autoapisummary::

   nnpd.nre.training.predict_logits
   nnpd.nre.training.classification
   nnpd.nre.training.fit_binary


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

.. py:function:: predict_logits(model, x, theta, batch_size: int, device) -> numpy.ndarray

.. py:function:: classification(labels: numpy.ndarray, logits: numpy.ndarray) -> dict

.. py:function:: fit_binary(context, model) -> dict

   No sample padding or dropping: every training row contributes once per epoch.

   A batch is split by rank. Local loss SUM * world/global_count compensates
   for DDP's gradient averaging even for unequal or empty last shards.
   An empty shard runs a zero-weight dummy forward to participate in collectives.
   The configured batch size is GLOBAL, not per GPU.