MNIST_Classification_usnig_PyTorch

This project focuses on CNN training for a classification task using PyTorch on the MNIST dataset while trying different approaches with the alternating between different :

Activation Functions:

• RELU
• Sigmoid
• Tanh

Pooling methods:

• max pooling ( 2,2)
• average pooling (2,2)

Optimizers:

• Adam
• SGD
• RMSProp

Drpoout probability:

• [0.1: 0.5 ]

Data augmentation approaches:

• Rotation
• flipping
• resizing

with visualizing results for each trial

Dependencies

this project requires installation of the following libraries:

• torch

• torchvision

• torchsummary

  pip install torch torchvision torchsummary

you can trace the results here