/dnn_scratch

Implementation of Deep neural network from scratch

Primary LanguagePython

dnn_scratch

Implementation of Deep neural network from scratch. Following the course by DeepLearning.ai and NNFS here is the implmentation for building your own Neural Network using Vanilla Python.

Resources for understanding concepts in DNN

https://docs.google.com/document/d/1lM-hM5fFjPF2nZFn4iVkrCymGNPgklHgyXt1vvfqlLo/edit?usp=sharing

Features yet to be implemented.

  1. Generalised load_dataset Function
  2. Weights Initialization
  3. Normalizing Input
  4. L2 Cost Regularisation (Done)
  5. Droput Method (Need to add test cases)
  6. Other Regularisation
  7. Gradient Check
  8. Mini Batch (Done)
  9. Exponentially Weighted Average
  10. Bias Correction in Exponentially Weighted Average
  11. RMSprop
  12. Gradient Descent with Momentum (Done)
  13. Adam Optimiser
  14. Learning Rate Decay
  15. Other Activation Function Support