mirror of https://github.com/hpcaitech/ColossalAI
e6d50ec107
* support existing sharded and unsharded parameters in zero * add unitest for moe-zero model init * polish moe gradient handler |
||
---|---|---|
.. | ||
__init__.py | ||
_base_ophook.py | ||
_memtracer_ophook.py | ||
_shard_grad_ophook.py | ||
_shard_param_ophook.py | ||
zero_hook.py |