Making large AI models cheaper, faster and more accessible
[shardformer] add gpt2 test and layer class refactor (#4041)
* add gpt2 test and layer class refactor * add dropout in gpt2 policy
F
FoolPlayer committed
4021b9a8a2dd3a9155bba04c0ed2cd7362fa437f
Parent: d857f3d
Committed by Frank Lee <somerlee.9@gmail.com>
on 7/4/2023, 8:05:01 AM