neural_compressor.tensorflow.keras.layers.depthwise_conv2d

Initialize custom depthwise conv2d layers for Keras quantization.

Classes

QDepthwiseConv2D

The custom quantized DepthwiseConv2D layer.

Functions

initialize_int8_depthwise_conv2d(fp32_layer, q_config)

Initialize int8 depthwise conv2d.

Module Contents

class neural_compressor.tensorflow.keras.layers.depthwise_conv2d.QDepthwiseConv2D(kernel_size, strides=(1, 1), padding='valid', depth_multiplier=1, data_format=None, dilation_rate=(1, 1), activation=None, use_bias=True, depthwise_initializer='glorot_uniform', bias_initializer='zeros', depthwise_regularizer=None, bias_regularizer=None, activity_regularizer=None, depthwise_constraint=None, bias_constraint=None, act_min_value=None, act_max_value=None, weight_min_value=None, weight_max_value=None, granularity='per_tensor', quant_status='calib', quant_mode='SCALED', quant_T='s8', quant_round_mode='HALF_AWAY_FROM_ZERO', quant_narrow_range=False, quant_axis=None, **kwargs)[source]

The custom quantized DepthwiseConv2D layer.

neural_compressor.tensorflow.keras.layers.depthwise_conv2d.initialize_int8_depthwise_conv2d(fp32_layer, q_config)[source]

Initialize int8 depthwise conv2d.