Which initializer will be used if I specify the init params in Conv2D and then call the initialize()?

below is my code:

import mxnet as mx
from mxnet import nd
from mxnet.gluon import nn

class ConvBlock(nn.HybridBlock):
    def __init__(self, **kwargs):
        super(ConvBlock, self).__init__(**kwargs)
        self.feature = nn.HybridSequential()
        self.feature.add(nn.Conv2D(16, 3, 1, 1, weight_initializer=mx.init.Xavier()))
        self.feature.add(nn.BatchNormal())
        self.feature.add(nn.ReLU())
    def hybrid_forward(F, x):
        return self.feature(x)

if __name__ == '__main__':
    net = ConvBlock()
    net.initialize()   # what initializer is used actually to initialize the net parameters ?

My question is what’s the actual initializer used during initialization of net.feature[0], i.e. Conv2D(...) ?