Pipeline Parallelism: Distributed Training via Model Partitioning

2 points| ml_basics | 2 years ago |siboehm.com