`neural_compressor.adaptor.tf_utils.quantize_graph.qdq.fuse_qdq_conv`

Quantize Conv2D/Conv3D/DepthwiseConv2dNative.

Module Contents

Classes

FuseNodeStartWithConv2d

Quantize Conv2D/Conv3D/DepthwiseConv2dNative to int8 op.

class neural_compressor.adaptor.tf_utils.quantize_graph.qdq.fuse_qdq_conv.FuseNodeStartWithConv2d(**kwargs)[source]: Quantize Conv2D/Conv3D/DepthwiseConv2dNative to int8 op.