`neural_compressor.adaptor.tf_utils.graph_rewriter.int8.freeze_fake_quant`

Freeze FakeQuant op Graph Rewriter.

Module Contents

Classes

FreezeFakeQuantOpOptimizer

Freeze fake_quant op to the following Quantize op and prioring Dequantize op.

class neural_compressor.adaptor.tf_utils.graph_rewriter.int8.freeze_fake_quant.FreezeFakeQuantOpOptimizer(model)[source]: Freeze fake_quant op to the following Quantize op and prioring Dequantize op.