cmo_digits

Handwritten Digit recogniser done as part of the Computational Methods and Optimisation Course at Plaksha

Training

You can import the Network class and train it with different hyperparameters and training data or you can use train.py.

Check the usage of train.py:

• python3 train.py --help

Results

SGD

Network Hyperparameters and Properties

• number of epochs: 30
• mini batch size: 10
• learning rate: 3.0
• activation function: sigmoid

Results:

Predicted 9537 out of 10000 numbers correctly.

Resources & References

• Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.
• Nielsen, M. (2015). Neural Networks and Deep Learning. Determination Press. https://neuralnetworksanddeeplearning.com/
• Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
• Hasanpour, S. H., Rouhani, M., Fayyaz, M., & Sabokrou, M. (2016). Lets keep it simple, using simple architectures to outperform deeper and more complex architectures. arXiv preprint arXiv:1608.06037.
• He, K., Zhang, X., Ren, S., & Sun, J. (2015). Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In Proceedings of the IEEE international conference on computer vision (pp. 1026-1034).