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

Freeze value with calibration Graph Rewriter.

Module Contents

Classes

FreezeValueTransformer

Freeze Value with calibration.

class neural_compressor.tensorflow.quantization.utils.graph_rewriter.int8.freeze_value.FreezeValueTransformer(model, max_min_data, postfix, tensor_data=None, th=1, device='gpu', itex_mode=False)[source]: Freeze Value with calibration.