ResNetBlock.md

from tensorflow import keras
import numpy as np
from pyradox import modules

inputs = keras.Input(shape=(28, 28, 1))
x = keras.layers.ZeroPadding2D(2)(inputs)         # padding to increase dimenstions to 32x32
x = keras.layers.Conv2D(3, 1, padding='same')(x)  # increasing the number of channels to 3
x = modules.ResNetBlock(filters=32)(x)
x = keras.layers.GlobalAvgPool2D()(x)
outputs = keras.layers.Dense(10, activation="softmax")(x)

model = keras.models.Model(inputs=inputs, outputs=outputs) 

model.summary()
keras.utils.plot_model(model, show_shapes=True, expand_nested=True)

Model: "model"
__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_1 (InputLayer)            [(None, 28, 28, 1)]  0                                            
__________________________________________________________________________________________________
zero_padding2d (ZeroPadding2D)  (None, 32, 32, 1)    0           input_1[0][0]                    
__________________________________________________________________________________________________
conv2d (Conv2D)                 (None, 32, 32, 3)    6           zero_padding2d[0][0]             
__________________________________________________________________________________________________
conv2d_2 (Conv2D)               (None, 32, 32, 32)   128         conv2d[0][0]                     
__________________________________________________________________________________________________
batch_normalization_1 (BatchNor (None, 32, 32, 32)   128         conv2d_2[0][0]                   
__________________________________________________________________________________________________
activation (Activation)         (None, 32, 32, 32)   0           batch_normalization_1[0][0]      
__________________________________________________________________________________________________
conv2d_3 (Conv2D)               (None, 32, 32, 32)   9248        activation[0][0]                 
__________________________________________________________________________________________________
batch_normalization_2 (BatchNor (None, 32, 32, 32)   128         conv2d_3[0][0]                   
__________________________________________________________________________________________________
activation_1 (Activation)       (None, 32, 32, 32)   0           batch_normalization_2[0][0]      
__________________________________________________________________________________________________
conv2d_1 (Conv2D)               (None, 32, 32, 128)  512         conv2d[0][0]                     
__________________________________________________________________________________________________
conv2d_4 (Conv2D)               (None, 32, 32, 128)  4224        activation_1[0][0]               
__________________________________________________________________________________________________
batch_normalization (BatchNorma (None, 32, 32, 128)  512         conv2d_1[0][0]                   
__________________________________________________________________________________________________
batch_normalization_3 (BatchNor (None, 32, 32, 128)  512         conv2d_4[0][0]                   
__________________________________________________________________________________________________
add (Add)                       (None, 32, 32, 128)  0           batch_normalization[0][0]        
                                                                 batch_normalization_3[0][0]      
__________________________________________________________________________________________________
activation_2 (Activation)       (None, 32, 32, 128)  0           add[0][0]                        
__________________________________________________________________________________________________
global_average_pooling2d (Globa (None, 128)          0           activation_2[0][0]               
__________________________________________________________________________________________________
dense (Dense)                   (None, 10)           1290        global_average_pooling2d[0][0]   
==================================================================================================
Total params: 16,688
Trainable params: 16,048
Non-trainable params: 640
__________________________________________________________________________________________________