ConceptAI & ML
Distributed data parallel
A PyTorch module that implements data parallelism across multiple GPUs by running separate processes and synchronizing gradients.
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.
Gradient accumulation1PyTorch1
Lecturers
1Connections
2The lectures that define it or assume it, and the concepts next to it.