`neural_compressor.experimental.strategy.utils.tuning_sampler`

Tuning sampler.

Module Contents

`TuningSamplerRegistry`	Class decorator used to register all TuningSampler subclasses.
`TuningOrder`	Not displayed in API Docs.
`TuningSampler`	Not displayed in API Docs.
`ModelWiseTuningSampler`	Not displayed in API Docs.
`OpTypeWiseTuningSampler`	Not displayed in API Docs.
`OpWiseTuningSampler`	Not displayed in API Docs.
`FallbackTuningSampler`	Not displayed in API Docs.
`SmoothQuantSampler`	Sampler for the hyperparameter tuning of smooth quantization.

class neural_compressor.experimental.strategy.utils.tuning_sampler.TuningSamplerRegistry[source]: Class decorator used to register all TuningSampler subclasses.

class neural_compressor.experimental.strategy.utils.tuning_sampler.TuningOrder[source]: Not displayed in API Docs.

class neural_compressor.experimental.strategy.utils.tuning_sampler.TuningSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_order_lst: List[TuningOrder], initial_op_tuning_cfg: Dict, kwargs: Dict = {})[source]

Not displayed in API Docs.

Basic class of tuning sampler.

class neural_compressor.experimental.strategy.utils.tuning_sampler.ModelWiseTuningSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_items_priority: List[str], tuning_order_lst: List[TuningOrder], op_dtype_dict: Dict[tuple, str], initial_op_tuning_cfg: Dict[tuple, neural_compressor.experimental.strategy.utils.tuning_structs.OpTuningConfig])[source]: Not displayed in API Docs.

class neural_compressor.experimental.strategy.utils.tuning_sampler.OpTypeWiseTuningSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_items_priority: List[str], tuning_order_lst: List[TuningOrder], op_dtype_dict: Dict[tuple, str], initial_op_tuning_cfg: Dict[tuple, neural_compressor.experimental.strategy.utils.tuning_structs.OpTuningConfig])[source]: Not displayed in API Docs.

class neural_compressor.experimental.strategy.utils.tuning_sampler.OpWiseTuningSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_items_priority: List[str], tuning_order_lst: List[TuningOrder], op_dtype_dict: Dict[tuple, str], initial_op_tuning_cfg: Dict)[source]: Not displayed in API Docs.

class neural_compressor.experimental.strategy.utils.tuning_sampler.FallbackTuningSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_order_lst: List[TuningOrder], initial_op_tuning_cfg: Dict[tuple, Any], op_dtypes: Dict[str, str], accumulate: bool, skip_first: bool = True)[source]: Not displayed in API Docs.

class neural_compressor.experimental.strategy.utils.tuning_sampler.SmoothQuantSampler(tuning_space: neural_compressor.experimental.strategy.utils.tuning_space.TuningSpace, tuning_order_lst: List[TuningOrder], initial_op_tuning_cfg: Dict, kwargs: Dict = {})[source]: Sampler for the hyperparameter tuning of smooth quantization.