MNIST-RECOGNIZER-CNN

Here i implemented a CNN model which will recognize the MNIST handwritten data set with a accuracy of 99.67 %

1) Dataset

• The mnist dataset is loaded using keras API datasets
• Preprocessing the dataset
• converting to a single channel image
• applied one hot encoding to the label values
• normalizing the pixel values to be in the range of 0 to 1
• Chaninging the datatype of the pixel to float 32

2) Model architecture

• Sequentail api is used
• single Convolution and Flattern layers is added
• Batch normalization is applied after convolution and flatten layer

3) Model configuration

• SGD optimizer is used
• lost: Categorical cross entropy
• The model will be evaluated using five-fold cross-validation. The value of k=5 was chosen to provide a baseline for both repeated evaluation and to not be so large as to require a long running time. Each test set will be 20% of the training dataset, or about 12,000 examples, close to the size of the actual test set for this problem.