ConceptAI & ML
Floating-point precision in deep learning
The bit representation used for numbers in tensors, where default float32 uses 32 bits but deep learning training can tolerate lower precisions to achieve higher performance.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Lecturers
1Connections
2The lectures that define it or assume it, and the concepts next to it.