CORE-301
The Transformer architecture
- Area
- COREShared trunk
- Level
- 300Advanced
- Designed workload
- 8 h
- Status
- Syllabus published
A design budget, set per level in the curriculum registry. Not a measurement: no module has been recorded.
Syllabus
- Content
- Attention, embeddings, positional encoding, encoder/decoder
- Hands-on
- Implement attention from scratch and visualise it
Where this module appears
Listed in 1 of the 10 programmes and in 4 of the 5 tracks. It is always the same object: opening CORE-301 from any of them lands here, with this syllabus and this status.
Programmes
Tracks
- Agentic AI300reused
- Deep Learning300reused
- Computer Vision300reused
- NLP & LLMs300reused
“own” is the track the module belongs to; “reused” is a track that pulls it in from elsewhere in the curriculum. Both point at this same object.