CNN Uint8 datatype Issue - numpy

Question

I am trying to make a CNN for MNIST using Keras, But I have some problems with the code.
I mostly get this error:
TypeError: Value passed to parameter 'input' has DataType uint8 not in list of allowed values: float16, bfloat16, float32, float64
Here is my code:
import numpy as np
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras.layers import Dense, Conv2D, Dropout, MaxPooling2D
from tensorflow.keras.callbacks import TensorBoard
from tensorflow.keras.utils import to_categorical
(Train_Data, Train_Labels), (Test_Data, Test_Labels) = tf.keras.datasets.mnist.load_data()
Train_Data = Train_Data.reshape(60000,28,28,1)
Test_Data = Test_Data.reshape(10000,28,28,1)
def save(model):
model.save("CNN")
def load(name):
model = tf.keras.models.load_model(name)
model = keras.Sequential()
model.add(Conv2D(784, kernel_size=3, activation='relu'))
model.add(MaxPooling2D(pool_size=(5,5)))
model.add(Dropout(.2))
model.add(keras.layers.Flatten())
model.add(Dense(25, activation='relu'))
model.add(Dense(10, activation='softmax'))
model.compile(optimzer='adam', loss="mse", metrics=['accuracy'])
model.fit(Train_Data, Train_Labels)
I don't know what to do, help would be appreciated,

Answer 1

The original images of the MNIST data are of type uint8 (values in range [0,255]), however before training a CNN you need to normalize them. Commonly you need to normalize it to some uniform bound around zero, for example [-0.5,0.5]. You can do so by adding the lines:
Train_Data = Train_Data / 255 - 0.5
Test_Data = Train_Data / 255 - 0.5