https://github.com/kevingastelum/mypytorchdeepdive

My Pytorch Notebook
https://github.com/kevingastelum/mypytorchdeepdive

Last synced: 1 day ago
JSON representation

My Pytorch Notebook

• Host: GitHub
• URL: https://github.com/kevingastelum/mypytorchdeepdive
• Owner: KevinGastelum
• Created: 2022-11-06T22:26:45.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2023-04-28T16:52:56.000Z (over 1 year ago)
• Last Synced: 2024-04-18T03:00:06.524Z (7 months ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 663 KB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Pytorch for Deep Learning

## [1. The Fundamentals [Click to View Code]](https://github.com/KevinGastelum/PytorchDeepLearning/blob/main/1_PytorchFundamentals.ipynb)

 * Intro to Tensors

 * Tensor Datatypes

 * Tensor Operations (Addition, Subtraction, Multiplication, Division)

 * Tensor Errors (Shape Error, DataType Error, Device Error, NumPy Error)

 * Tensor Aggregation (aRange, Min, Max, Mean, Sum, etc)

 * Tensor Shaping, Stacking, Squeezing and Unsqueezing

 * Tensor Indexing

 

 ## [2. PyTorch Workflow [Click to View Code]](https://github.com/KevinGastelum/PytorchDeepLearning/blob/main/2_PytorchWorkflow.ipynb)

 * Prepping and Loading data

 * Building our Model

 * Training Our Model

 * Making Predictions

 * Saving and Loading our Model

 * Putting it all together

 * Building a Pytorch Linear Model

 * Training Deep dive (Forward Pass, Loss Function, Optimizer, Training and Test loop)

 

 ## [3. Neural Network Classification [Click to View Code]](https://github.com/KevinGastelum/PytorchDeepLearning/blob/main/3_PytorchClassification.ipynb)

 * Prepping and Loading Classification data (make_circles Dataset)

     * Turn Data into Tensors, Create Train/Test Split

 * Building Classification Model 

     * nnModule, Hidden Layers, Hidden Neurons, forward pass

 * **loss function (BCELoss/BCEWithLogitsLoss) and Optimizer (SGD/Adam)** 

 * Training the Model

 * Using **activation function (Sigmoid)** to convert raw logits into Predictions

 * Build Train and Test loop

 * Making Predictions to evaluate our current Model

 * Improving our Model - **Hyperparameter Tuning** (Hidden Layers, Hidden Neurons, Activation Function, learning rate, loss function)

 ### Neural Network Part 2 Multi-Class Classification

 * Prepping and Loading Multi-Class Classification data (make_blobs Dataset)

     * Turn Data into Tensors, Create Train/Test split

 * Building a Multi-Class Classification Model

     * nnModule, Hidden Layers, Hidden Neurons, Forward pass

 * **loss function(CrossEntropyLoss) and optimizer(SGD)** for Multi-class Model

 * Using **activation function (softmax)** to convert raw logits into Predictions

 * Build Train and Test Loop for Multi-class model 

 * Multi-class model Predictions