pymllm.layers.base
==================

.. py:module:: pymllm.layers.base


Classes
-------

.. autoapisummary::

   pymllm.layers.base.MllmBaseLayer


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

.. py:class:: MllmBaseLayer

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


   .. py:attribute:: quant_method
      :type:  Optional[pymllm.layers.quantize_base.QuantizeMethodBase]
      :value: None



   .. py:method:: weight_loader(param, loaded_weight)

      Load weights into a parameter.

      This is the default implementation that directly copies the loaded weight
      into the parameter. Subclasses should override this method to implement
      custom loading logic (e.g., tensor parallelism sharding).

      :param param: The parameter to load weights into.
      :param loaded_weight: The weight tensor loaded from checkpoint.



   .. py:method:: forward(*args, **kwargs)
      :abstractmethod: