56
edits
Changes
→Data Parallelism[https://pytorch.org/tutorials/beginner/blitz/data_parallel_tutorial.html]
return t
== Data Parallelism[https://pytorch.org/tutorials/beginner/blitz/data_parallel_tutorial.html] ==
This section details the ways to parallelize your NN[https://pytorch.org/tutorials/beginner/blitz/data_parallel_tutorial.html].
As image recognition is graphical in nature, multiple GPUs are the best way to parallelize dataset training. <code>DataParallel</code> is a single-machine parallel model, that uses multiple GPUs. It is more convenient than a multi-machine, distributed training model.