Skip to content
AI-grafen
Sign in
☰ Menu

The goal E University

Training neural networks for real

Optimisers, regularisation, normalisation, initialisation, dataloaders and transfer learning — everything that separates a training run that works from one that does not.

Knowledge nodes
86
From zero
about 65 h
Labs
6
See what you already know — no account

The diagnostic removes what you already know, so your path is usually much shorter.

What you can do afterwards

Labs along the way

You write the code. Tests you cannot see decide whether it holds up.

Lab: a neural network in pure NumPy — forward, backprop, gradient checkDa sandbox · about 75 minLab: dot product, norm and cosine similarityDin the browser · about 40 minLab: gradient descent from scratchDin the browser · about 50 minLab: matrix multiplication and one layer of a neural networkDin the browser · about 45 minLab: semantic search with embeddingsDa sandbox · about 45 minLab: tensors, autograd and your first training loopDa sandbox · about 50 min

The whole path

Everything the goal builds on, grouped by level and in the order it builds on itself. Show on the map

AExplorer4 knowledge nodes
  1. Patterns and categories
  2. Sequences and precise instructions
  3. Integer arithmetic and order of operations
  4. What is a computer?
BInvestigator10 knowledge nodes
  1. Data in everyday life
  2. Charts and tables
  3. Algorithmic thinking
  4. Rule-based systems and machine learning
  5. Binary numbers and bits
  6. Fractions, decimals, and percentages
  7. Negative numbers in everyday life and AI
  8. Training data, features and labels
  9. Classification: how a model sorts information
  10. Neural networks: inspiration from the brain
CBuilder12 knowledge nodes
  1. Functions and coordinate systems
  2. Programming logic — variables, conditions, loops
  3. Python — the basics
  4. Python — lists, loops and dictionaries
  5. Python — strings and text processing
  6. Python — files, CSV and JSON
  7. Statistics — mean, median and spread
  8. Probability — the basics
  9. Linear regression: fitting a straight line to data
  10. Neural networks — the intuition
  11. Variables and algebraic expressions
  12. Powers and roots
DAI developer31 knowledge nodes
  1. Derivatives and optimisation
  2. Time complexity and big-O notation
  3. Python — functions, scope and exceptions
  4. Python — modules, packages and virtual environments
  5. Git — version control
  6. Probability distributions
  7. The normal distribution and standardisation
  8. Gradient descent
  9. Logarithms
  10. Vectors
  11. Matrices and matrix multiplication
  12. Linear regression with several features
  13. Logistic regression and decision trees
  14. Decision trees
  15. Neural networks — the forward pass with matrices
  16. Activation functions
  17. Backpropagation
  18. Embeddings — words as vectors
  19. NumPy — arrays and vectorisation
  20. Overfitting and generalisation
  21. Pandas — tables in Python
  22. Data quality: missing values, duplicates, errors
  23. Feature engineering
  24. Embeddings for categorical variables
  25. The perceptron
  26. PyTorch — tensors and autograd
  27. Dataloaders, batches and epochs
  28. Training, validation and test
  29. Cross-validation
  30. Train a neural network in PyTorch
  31. Visualisation with Matplotlib
EUniversity27 knowledge nodes
  1. Numerical stability and floating point
  2. Ensembles: random forest and boosting
  3. Linear maps
  4. Eigenvalues and eigenvectors
  5. Matrix factorisation and low-rank approximation
  6. Vanishing and exploding gradients
  7. Gradient clipping
  8. Weight initialisation
  9. Parallelism and why GPUs
  10. How autograd works inside
  11. Residual connections
  12. Sequence models before transformers
  13. Singular value decomposition (SVD)
  14. PCA — principal component analysis
  15. Autoencoders
  16. Deep learning on tabular data
  17. Hyperparameter search
  18. Batch and layer normalisation
  19. Checkpoints, saving and restarting
  20. GPU memory, gradient accumulation and batches
  21. Optimisers: momentum, Adam, scheduling
  22. Learning rate schedules and warm-up
  23. Regularisation: dropout, weight decay, early stopping
  24. Dropout in detail
  25. Transfer learning
  26. Training diagnostics
  27. Hyperparameters for deep networks
FAI engineering2 knowledge nodes
  1. Floating-point formats: fp32, fp16, bf16
  2. Mixed precision training