neural_compressor.torch.quantization.autotune
=============================================

.. py:module:: neural_compressor.torch.quantization.autotune

.. autoapi-nested-parse::

   Intel Neural Compressor Pytorch quantization AutoTune API.



Functions
---------

.. autoapisummary::

   neural_compressor.torch.quantization.autotune.get_rtn_double_quant_config_set
   neural_compressor.torch.quantization.autotune.get_all_config_set
   neural_compressor.torch.quantization.autotune.autotune


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

.. py:function:: get_rtn_double_quant_config_set() -> List[neural_compressor.torch.quantization.config.RTNConfig]

   Generate RTN double quant config set.

   :returns: a set of quant config
   :rtype: List[RTNConfig]


.. py:function:: get_all_config_set() -> Union[neural_compressor.common.base_config.BaseConfig, List[neural_compressor.common.base_config.BaseConfig]]

   Generate all quant config set.

   :returns: a set of quant config
   :rtype: Union[BaseConfig, List[BaseConfig]]


.. py:function:: autotune(model: torch.nn.Module, tune_config: neural_compressor.common.base_tuning.TuningConfig, eval_fn: Callable, eval_args=None, run_fn=None, run_args=None, example_inputs=None)

   The main entry of auto-tune.

   :param model: _description_
   :type model: torch.nn.Module
   :param tune_config: _description_
   :type tune_config: TuningConfig
   :param eval_fn: for evaluation of quantized models.
   :type eval_fn: Callable
   :param eval_args: arguments used by eval_fn. Defaults to None.
   :type eval_args: tuple, optional
   :param run_fn: for calibration to quantize model. Defaults to None.
   :type run_fn: Callable, optional
   :param run_args: arguments used by run_fn. Defaults to None.
   :type run_args: tuple, optional
   :param example_inputs: used to trace torch model. Defaults to None.
   :type example_inputs: tensor/tuple/dict, optional

   :returns: The quantized model.