Wenwen Qu 21624f6f81 fix(moe): remove norm&gate force sync (#448) ... * add zero broadcast_sync * delete old sync logic * fix merged error * refactor code * remove some unused function (is norm/gate group)		2023-11-01 11:29:55 +08:00
..
legacy	feat(ckpt): fix checkpoint bugs and add feature enhancements. (#259)	2023-09-05 17:40:48 +08:00
__init__.py	feat(numa): bind numa if possible (#320)	2023-09-25 19:34:52 +08:00
initialize_tensor.py	feat(model): implement uniform_init for tensor. (#252)	2023-09-01 01:12:53 +08:00
initialize_trainer.py	docs(*): add documentation and reST files for readthedocs (#272)	2023-09-06 15:36:03 +08:00
launch.py	fix(moe): remove norm&gate force sync (#448)	2023-11-01 11:29:55 +08:00