Learning & Building in Pytorch
- Tensors Basics: View Notebook - Learning about Tensors in PyTorch
- Autograd: View Notebook - Learning about Autograd and how they help in ML during Forward Pass, Backward Pass.
- Training Pipeline: View Notebook - Building a Single Neuron Neural Network with Training Pipeline for Learning.
- NN Modules: View Notebook - Building NN with help of NN Modules.
- Dataset & DataLoader: View Notebook - Theory on Dataset & DataLoader and Mini Batch Gradient Descent also Improving our code.
- ANN (Artificial Neural Network) / MLP (Multi Layer Perceptron): View Notebook - Buidling ANN without GPU with and without GPU of MNIST Fashion Dataset.
- ANN Optimization: View Notebook - Optimizing ANN to reduce Overfitting using Dropouts, Batch Normalization and L2 Regularization.
- Hyperparameter Tuning & Bayesian Optimization (Optuna terms): View PDF - Grid Search CV and Random Search CV are techniques used for hyperparameter tuning but Optuna uses Bayesian Optimization by default.
- Optuna Basics: View Notebook - Applying hyperparameter tuning using Optuna Bayesian Optimization.
- Hyperparameter Tuning Using Optuna: View Notebook - Hyperparameter Tuning of ANN with different parameters using Optuna to get the best accuracy for model.
- Convolutional Neural Network: View Notebook - Simple CNN for Image Classification.
- Transfer Learning: View Notebook - Adding Transfer Learning in CNN with pretrained VGG-16 model.
- Recurrent Neural Network (RNNs): View Notebook - RNN that predicts answers from set of 100 questions dataset with variations.
- Long Term Short Term (LSTM): View Notebook, View PDF - Simple next word predictor LSTM.