Machine Learning at Scale

Data Parallelism

Parallelization strategy where data is partitioned across multiple machines, each training an identical copy of the model with different batches.

← Terug