AI & ML Engineering · Unit 10 · Deep Learning
NLP and Transformers
How machines read text, and the architecture behind every modern language model.
- 10.1
Text Preprocessing
Cleaning text, and classical representations like bag of words and TF-IDF.
- 10.2
Tokenisation
How text becomes tokens, and why subword tokenisers won.
- 10.3
Word Embeddings
Representing meaning as vectors, from Word2Vec to contextual embeddings.
- 10.4
The Attention Mechanism
Self-attention: how a model decides which words matter to each other.
- 10.5
Transformer Architecture
Encoders, decoders, positional encoding and multi-head attention.
- 10.6
Pretrained Models and Hugging Face
Loading, running and adapting pretrained models.
Unit project
A transformer fine-tuned for a text classification task, compared against a classical baseline.
Paid learners get the full brief, submit it from their dashboard and have it reviewed. It goes in your portfolio.
Open every unit
Enrol once to open every unit, its videos and a reviewed project for your portfolio.