neural_compressor.ux.utils.workload.config

Configuration module.

Module Contents

Classes

Config

Configuration class.

class neural_compressor.ux.utils.workload.config.Config(data: Dict[str, Any] = {})

Bases: neural_compressor.ux.utils.json_serializer.JsonSerializer

Configuration class.

initialize(data: Dict[str, Any] = {}) → None

Initialize config from dict.

remove_dataloader() → None

Remove datalader.

remove_accuracy_metric() → None

Remove accuracy metric in config.

set_evaluation_dataloader(dataloader: dict) → None

Udpate dataloader in evaluation config.

set_evaluation_dataset_path(dataset_path: str) → None

Update dataset path in evaluation config.

get_performance_configs() → neural_compressor.ux.utils.workload.evaluation.Configs | None

Get evaluation.performance.configs.

set_performance_cores_per_instance(cores_per_instance: int) → None

Update cores_per_instance evaluation config.

get_performance_cores_per_instance() → int | None

Get cores_per_instance evaluation config.

set_performance_num_of_instance(num_of_instance: int) → None

Update num_of_instance evaluation config.

get_performance_num_of_instance() → int | None

Get num_of_instance evaluation config.

set_accuracy_and_performance_batch_sizes(batch_size: int) → None

Update batch_size evaluation configs.

set_quantization_dataloader(dataloader: dict) → None

Update dataloader in quantization config.

set_quantization_dataset_path(dataset_path: str) → None

Update dataset path in quantization config.

set_quantization_batch_size(batch_size: str) → None

Update batch_size in quantization config.

set_workspace(path: str) → None

Update tuning workspace path in config.

set_accuracy_goal(accuracy_goal: float) → None

Update accuracy goal in config.

set_accuracy_metric(metric: dict) → None

Update accuracy metric in config.

set_transform(transform: List[Dict[str, Any]]) → None

Set transforms metrics in config.

static process_transform(config: collections.OrderedDict, transform: List[Dict[str, Any]]) → None

Process transformation.

set_quantization_approach(approach: str) → None

Update quantization approach in config.

set_inputs(inputs: List[str]) → None

Update inputs in config.

set_outputs(outputs: List[str]) → None

Update outputs in config.

set_quantization_sampling_size(sampling_size: str) → None

Update quantization sampling size in config.

set_performance_warmup(warmup: int) → None

Update performance evaluation warmup iteration number.

set_performance_iterations(iterations: int) → None

Update performance evaluation iteration number.

set_optimization_precision(framework: str, precision: str) → None

Update graph optimization precision.

load(path: str) → None

Load configuration from file.

dump(yaml_path: str) → None

Dump configuration to file.