flexible-cnn

In this project I have tried to create a flexible version of convolutional neural network using PyTorch, where all that one needs to do is just plug-in the required parameters for the model and have the model ready without bothering about the underlying implementation details.

Files

• main.py - this is where you modify the network parameters.
• models.py - PyTorch implementation for the CNN and Data class.
• helper.py - contains various utility functions that deal with training the model and plotting the performance curves (makes use of GPU if available).

Experiment

Configuration

• Modify the network parameter in main.py according to your requirements.

Run

python main.py