Using custom layers in Functional API

I am trying to write a custom layer using subclassing in Keras and then use it in functional API. I followed the advise in the documentation - https://www.tensorflow.org/guide/keras/custom_layers_and_models#you_can_optionally_enable_serialization_on_your_layers but its not working for me. Here is my code snippet-

from tensorflow import keras
from tensorflow.keras import layers

class conv2D_Block(keras.layers.Layer):
    def __init__(self, filters, kernel_size=(3,3), strides=1, padding='same', pool_size=(2,2), 
**kwargs):
        super().__init__(**kwargs)
        
        self.filters = filters
        self.kernel_size = kernel_size 
        self.strides = strides 
        self.padding = padding 
        self.pool_size = pool_size 
        

        self.convlayer = layers.Conv2D(filters, kernel_size, strides = strides, padding= 
padding)  
        self.pool = layers.MaxPool2D(pool_size = pool_size)
        self.BatchNormalization = layers.BatchNormalization()

    def call(self, x):
        x = self.convlayer(x)
        x = keras.activations.relu(x)
        x = self.pool(x)
        x = self.BatchNormalization(x)
        return x

    def get_config(self):
        config = super(conv2D_Block, self).get_config()
        #config.update({filters:self.filters, kernel_size: self.kernel_size, 
        #strides:self.strides, padding:
        #              self.padding, pool_size: self.pool_size})
        config.update({convlayer: self.convlayer, pool: self.pool, BatchNormalization: 
        self.BatchNormalization})
        return config

layer1 = conv2D_Block(16)
config1 = layer1.get_config()

inputs = keras.Input(shape = (28,28,1))
inputs = conv2D_Block.from_config(config1)(inputs)
flatconv = layers.Flatten()(inputs)
output = layers.Dense(10,activation='softmax')(flatconv)
model = keras.Model(inputs = inputs, outputs = output)
model.compile(optimizer = 'adam', loss='sparsecategoricalcrossentropy', metrics = 'accuracy')

I understand that in the example given in the link, attributes of the model are simple (integer etc) while the example I am using has keras object attributes. I tried with minor changes to the code (using filter etc as attributes) but none of it works. For the above code, I am getting the error File C:\Users\Asus-ROG\Desktop\Machine learning\keras-convlayer.py, line 60, in inputs = conv2D_Block.from_config(config1)(inputs)

File C:\Users\Asus-ROG\anaconda3\envs\tf_gp\lib\site- 
packages\tensorflow\python\keras\engine\base_layer.py, line 697, in from_config
return cls(**config)

TypeError: __init__() missing 1 required positional argument: 'filters'

I have tried the obvious fixes(I hope) as you can see in the commented parts of the code. Can someone please help me out?