Deep Learning with PyTorch

1. 01
   Part 1Neural network basics8 topics1 project

   • Neurons, layers and activations
   • Forward pass and loss functions
   • Backpropagation in plain words
   • Gradient descent and optimizers (SGD, Adam)
   • PyTorch tensors and autograd
   • Building a model with nn.Module
   • The training loop
   • GPUs and Google Colab

   Project Recognise handwritten digits.

2. 02
   Part 2Computer vision8 topics1 project

   • Images as numbers
   • Convolutional neural networks (CNNs)
   • Pooling, padding and stride
   • Data augmentation
   • Transfer learning with pre-trained models
   • Image classification project
   • Object detection with YOLO (intro)
   • Image segmentation (intro)

   Project Classify plant diseases from leaf photos.

3. 03
   Part 3Text and transformers8 topics1 project

   • Tokenization and embeddings
   • RNNs and LSTMs (short history)
   • Attention and self-attention
   • The transformer architecture
   • Hugging Face Transformers library
   • Fine-tuning BERT for text classification
   • Text generation with small LLMs
   • LoRA: fine-tuning on a small budget

   Project Sentiment analysis of product reviews.

4. 04
   Part 4Training well8 topics1 project

   • Overfitting, dropout and weight decay
   • Batch normalization
   • Learning rate schedules
   • Early stopping and checkpoints
   • Mixed precision training
   • Tracking runs with TensorBoard or Weights & Biases
   • Debugging a model that does not learn
   • Reading a research paper

   Project Train, compare and pick the best of 3 models.

5. 05
   Part 5Deploy8 topics1 project

   • Saving and loading models
   • Exporting to ONNX
   • A Gradio demo app
   • Serving with FastAPI
   • Hugging Face Spaces deployment
   • Making models smaller and faster (quantization)
   • Portfolio project
   • Interview questions for DL roles

   Project A live image or text AI demo.

This is the planned syllabus. Small changes are possible before launch.