A complete learning repository for Deep Learning using PyTorch, covering everything from tensors and neural networks to training pipelines and real-world projects.
This repository is built as a learning + portfolio project — every notebook here represents a concept I have learned and implemented from scratch.
Deep Learning is not about just using libraries — it is about understanding:
- How neural networks work
- How gradients flow
- How loss functions & optimizers train models
- How real training pipelines are built
This repository documents that journey in a structured, practical way.
- Python
- PyTorch
- Jupyter Notebook
- NumPy
- Matplotlib
(Updated as I learn more)
Learned what PyTorch is, why it’s used, how to import it, and the core idea of tensors.
Created tensors (scalars, vectors, matrices), used zeros, ones, random tensors, and explored shapes.
Performed tensor arithmetic, matrix multiplication, aggregations, and inspected tensor properties.
Learned reshaping, stacking, indexing, and converting between NumPy arrays and PyTorch tensors.
Set random seeds, ensured reproducibility, and moved tensors between CPU and GPU.
Built a binary classification model, prepared data, and defined loss functions and optimizers.
Implemented training loops, backpropagation, evaluation metrics, and visualized predictions.
Trained a multiclass classifier using Softmax and CrossEntropyLoss and evaluated performance.
Improved models using non-linearities and summarized the full classification pipeline.
Understood what computer vision is, how images are represented as tensors, and how PyTorch handles image data.
Built a baseline image classification model and learned the full training + evaluation flow on image data.
Improved the baseline model by adjusting architecture, epochs, and training strategy.
Learned how CNNs work, implemented convolutional layers, and understood feature extraction.
Trained a CNN model, evaluated performance, and visualized predictions on image data.
Reviewed the full computer vision pipeline and summarized learnings from Chapter 3.
(This structure grows as new topics and projects are added.)
git clone https://github.com/KavinKohli/Deep-Learning-PyTorch.git
cd Deep-Learning-PyTorch