This project is a collection of Python scripts that demonstrate various functionalities related to machine learning, specifically deep learning using PyTorch. The scripts cover a range of topics from data preprocessing, model creation, training, testing, and visualization of results.
-
nn_seq.py
: This script defines a convolutional neural network model using PyTorch's Sequential API. -
read_data.py
: This script reads image data from a directory and creates a PyTorch Dataset object. -
rename_dataset.py
: This script renames images in a dataset and writes the labels to a text file. -
test.py
: This script loads a pre-trained model and uses it to make predictions on a single image. -
test_tensorboard.py
: This script demonstrates how to use TensorBoard for visualization of model training. -
train.py
: This script trains a model on the CIFAR-10 dataset and saves the model after each epoch. -
train_gpu_1.py
andtrain_gpu_2.py
: These scripts are variations oftrain.py
that demonstrate how to use a GPU for training if available. -
transforms.py
andtransformsV2.py
: These scripts demonstrate how to use PyTorch's transforms for data augmentation. -
nn_relu.py
: This script demonstrates the use of ReLU activation function in a neural network. -
nn_module.py
: This script demonstrates the basic structure of a PyTorch Module. -
nn_maxpool.py
: This script demonstrates the use of MaxPooling in a convolutional neural network. -
nn_optim.py
: This script demonstrates how to use PyTorch's SGD optimizer. -
nn_loss_network.py
: This script demonstrates how to calculate the loss of a network using CrossEntropyLoss. -
nn_linear.py
: This script demonstrates the use of a Linear layer in a neural network. -
nn_loss.py
: This script demonstrates how to calculate different types of loss functions in PyTorch. -
nn_conv2d.py
: This script demonstrates the use of a Conv2D layer in a convolutional neural network. -
nn_conv.py
: This script demonstrates how to perform a 2D convolution operation using PyTorch's functional API.
- Python 3.6 or above
- PyTorch 1.0 or above
- torchvision
- PIL
- TensorBoard
Each script can be run independently. For example, to run nn_seq.py
, use the following command:
python nn_seq.py
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.