XQuantConfig.rst.txt

orphan:

XQuant Configuration

.. autoclass:: model_compression_toolkit.xquant.common.xquant_config.XQuantConfig
    :members:

Note

The following parameters are only used in the xquant_report_troubleshoot_pytorch_experimental function.

• quantize_reported_dir
• threshold_quantize_error
• is_detect_under_threshold_quantize_error
• threshold_degrade_layer_ratio
• threshold_zscore_outlier_removal
• threshold_ratio_unbalanced_concatenation
• threshold_bitwidth_mixed_precision_with_model_output_loss_objective