@openstem shared this note
Software · L5 · Systems for Distributed Machine Learning
Data/model parallelism, parameter servers, synchronous vs. asynchronous SGD, communication-efficient gradient aggregation, and pipeline parallelism.
Software
Updated 0
0
Read only
More in Software
Browse Software- Software · L2 · Git & Dev Workflow: Resolving a Merge Conflictnote
- Software · L2 · MongoDB: Replica Sets & Failovernote
- Software · L5 · Compiler Construction Theorynote
- Software · L1 · Operating Systems: The Process Lifecyclenote
- Software · L5 · Type Theory and the λ-Calculusnote
- Software · L2 · PostgreSQL: Replication & High Availabilitynote