`neural_compressor.tensorflow.quantization.utils.graph_rewriter.int8.freeze_value_without_calib`

Freeze Value without calibration Graph Rewriter.

Module Contents

Classes

FreezeValueWithoutCalibTransformer

Freeze value without calibration.

class neural_compressor.tensorflow.quantization.utils.graph_rewriter.int8.freeze_value_without_calib.FreezeValueWithoutCalibTransformer(model, max_min_data, postfix, th=0.95, device='gpu')[source]: Freeze value without calibration.