Open robotsp opened 5 months ago
@zhuohan123 Hi Zhuohan,
Thanks for your cool work on pipeline parallelism. May I ask is Terapipe implemented on 1F1B schedule? Does it integrate on Megatron-LM framework?
Thanks!
Hi, maybe Seq1F1B is what you need: Arxiv link: https://arxiv.org/abs/2406.03488 Github: https://github.com/MayDomine/Seq1F1B
@zhuohan123 Hi Zhuohan,
Thanks for your cool work on pipeline parallelism. May I ask is Terapipe implemented on 1F1B schedule? Does it integrate on Megatron-LM framework?
Thanks!