| 1 |
1D_tensors |
Basic operations on 1D tensors |
| 2 |
2D_tensors |
Basic operations on 2D tensors |
| 3 |
Derivatives |
Derivatives in Pytorch |
| 4 |
Toy_dataset |
Creating a toy dataset in Pytorch, compose and perform transformations on it |
| 5 |
Datasets_and_transforms |
Build an image dataset object and perform pre-build transformations using torchvision.transforms on it |
| 6 |
MNIST_data_&_transforms |
How to use pre-built MNIST dataset and perform transformations on it |
| 7 |
Regression_prediction |
Make predictions for multiple 1D inputs using linear class |
| 8 |
1D_Linear_regression_1_parameter |
Create linear regression model using 1 parameter, cost/criterion function using MSE, and plot parameters as well as loss values |
| 9 |
1D_Linear_regression_2_parameters |
1D Linear regression model using 2 parameters (w and b). Visualize the data space and the parameter space during training via batch gradient descent |
| 10 |
Stochastic_gradient_descent |
1D Linear regression using stochastic gradient descent |
| 11 |
Mini_batch_gradient_descent |
1D Linear regression using mini-batch gradient descent. This code also includes comparison between batch, stochastic and mini-batch gradient descent with different batch sizes |
| 12 |
Mini_batch_gradient_descent2 |
1D Linear regression using PyTorch build-in functions |
| 13 |
Models_with_different_LR |
1D Linear regression with different learning rates and view results such as training and validation losses at different LR |
| 14 |
Multiple_linear_regression_prediction |
Multiple linear regression prediction (preparing forward propagation with 1xn tensor input) |
| 15 |
Multiple_linear_regression_training |
Multiple linear regression training with input of 1xn tensor |
| 16 |
Multi_target_linear_regression |
Multiple target linear regression prediction (forward propagation) |
| 17 |
training_multiple_output_linear_regression |
pytorch build in functions to train multiple target linear regression |
| 18 |
logistic_regression_prediction |
Prediction using sigmoid/logistic function |
| 19 |
logistic_regression_prediction |
Illustration of poor performance of logistic regression via bad parameters initialization |
| 20 |
Softmax_in_1D |
Building a Softmax classifier in 1D |
| 21 |
predicting_MNIST_using_Softmax |
Classify handwritten digits from the MNIST database by using Softmax classifier and visualize parameters learned for each class following model training |
| 22 |
simpleNN_1hiddenlayer |
Simple Neural Network with 1 hidden layer |
| 23 |
NN_more_hidden_neurons |
Neural Networks with 1 hidden layer (more neurons) |
| 24 |
Neural_networks |
Building a neural network with 1 hidden layer to classify noisy XOR data |
| 25 |
1_layer_neural_network_MNIST |
Neural networks with 1 hidden layer to classify MNIST data |
| 26 |
Activation_function |
How to apply different Activation functions in Neural Network |
| 27 |
Different_activations_on_neural_network |
Apply different activation functions in Neural Network on the MNIST dataset |
| 28 |
Deep_Neural_Networks |
Deep Neural Networks with 2 hidden layers on the MNIST dataset |
| 29 |
Deeper_Neural_Networks |
Deep Neural Networks with 3 hidden layers using nn.ModuleList() |
| 30 |
Dropout_prediction |
Deep Neural Networks with dropout for classification |
| 31 |
Dropout_regression |
Using dropout in regression |
| 32 |
Weight_initialization |
Performance of neural networks with constant (w = 1) vs. default weight initialization |
| 33 |
Xavier_initialization |
Performance of neural networks with uniform, default, and Xavier initialization |
| 34 |
He_initialization |
Performance of neural networks with uniform, default, and He initialization |
| 35 |
MomentumwithPolynomialFunctions |
Use of momentum in the model optimization |
| 36 |
MomentumwithPolynomialFunctions |
Neural networks model optimization with different momentum values |
| 37 |
BachNorm |
Comparison of neural networks with and without batch normalization |
| 38 |
Convolution |
Convolution on an image and estimate the output size using kernel of K size |
| 39 |
MaxPooling |
Application of activation function and max pooling |
| 40 |
Multiple_Channel_Convolution |
Convolutions using multiple input and output channels |
| 41 |
ConvolutionalNeuralNetworkexample |
Example of convolutional neural network |
| 42 |
CNN_Small_Image |
Build CNN using small MNIST images, visualize learned parameters and plot testing image activations after each layer. The code also plot the mis-classified samples |
| 43 |
CNN_Small_Image_batch |
Compare a CNN using batch normalization with a regular CNN to classify handwritten digits from the MNIST database |