Jiarui Fang c9e7d9582d [zero] polish shard strategy (#310) ...

* init shard param from shape tuple

* add more unitest for shard param

* add set_payload method for ShardedParam

* [zero] add shareded tensor class

* polish code

* add shard stratgy

* move shard and gather logic to shard strategy from shard tensor.

* polish code

2022-03-11 15:50:28 +08:00

..

amp

added buffer sync to naive amp model wrapper (#291)

2022-03-11 15:50:28 +08:00

builder

add pytorch hooks (#179)

2022-01-25 22:20:54 +08:00

communication

moved env variables to global variables; (#215)

2022-02-15 11:31:13 +08:00

context

moved env variables to global variables; (#215)

2022-02-15 11:31:13 +08:00

engine

fix sharded param hook and unit test

2022-03-11 15:50:28 +08:00

kernel

Optimized MoE layer and fixed some bugs;

2022-03-11 15:50:28 +08:00

logging

fixed mkdir conflict and align yapf config with flake (#220)

2022-02-15 11:31:13 +08:00

nn

Added TPExpert for special situation

2022-03-11 15:50:28 +08:00

registry

add pytorch hooks (#179)

2022-01-25 22:20:54 +08:00

trainer

moved env variables to global variables; (#215)

2022-02-15 11:31:13 +08:00

utils

[profiler] primary memory tracer

2022-03-11 15:50:28 +08:00

zero

[zero] polish shard strategy (#310)

2022-03-11 15:50:28 +08:00

__init__.py

Develop/experiments (#59)

2021-12-09 15:08:29 +08:00

constants.py

moved env variables to global variables; (#215)

2022-02-15 11:31:13 +08:00

core.py

Develop/experiments (#59)

2021-12-09 15:08:29 +08:00

global_variables.py

Optimized MoE layer and fixed some bugs;

2022-03-11 15:50:28 +08:00

initialize.py

added buffer sync to naive amp model wrapper (#291)

2022-03-11 15:50:28 +08:00