未验证 提交 fd53b56a 编写于 作者: R Rahul Prasad 提交者: GitHub

fix typo (#72)

上级 bf2689a9
......@@ -34,7 +34,7 @@ model scientists also need to set up the state-of-the-art training techniques
such as distributed training, mixed precision, gradient accumulation, and
checkpointing. Yet still, scientists may not achieve the desired system
performance and convergence rate. Large model sizes are even more challenging:
a large model easily runs out of memory with pure data paralelism and it is
a large model easily runs out of memory with pure data parallelism and it is
difficult to use model parallelism. DeepSpeed addresses these challenges to
accelerate model development *and* training.
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册