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

Fuse QuantizedConv QuantizedDeConv with redundant Dequantize Graph Rewriter.

Module Contents

Classes

FuseConvRedundantDequantizeTransformer

Fuse _QuantizedConv/_QuantizedDeConv with the successor Dequantize Op.

class neural_compressor.adaptor.tf_utils.graph_rewriter.int8.fuse_conv_redundant_dequantize.FuseConvRedundantDequantizeTransformer(model, device='cpu')[source]: Fuse _QuantizedConv/_QuantizedDeConv with the successor Dequantize Op.