github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-26 04:03:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

[hotfix] Add layer norm gradients all-reduce for sequence parallel (#4926)

Browse Source 
* [hotfix] Add layer norm gradients all-reduce for sequence parallel. (#4915)

* Add layer norm gradients all-reduce for sequence parallel.

* skip pipeline inference test

* [hotfix] fixing polices of sequence parallel (#4922)

* Add layer norm gradients all-reduce for sequence parallel.

* fix parameter passing when calling get_autopolicy

---------

Co-authored-by: littsk <1214689160@qq.com>

* Hotfix/add grad all reduce for sequence parallel (#4927)

* Add layer norm gradients all-reduce for sequence parallel.


* fix parameter passing when calling get_autopolicy

* fix bug using wrong variables

---------

Co-authored-by: littsk <1214689160@qq.com>

* fix policy initialization

* fix bloom and chatglm policices

* polish code of handling layernorm

* fix moe module

* polish code of class initializing

---------

Co-authored-by: Zhongkai Zhao <kanezz620@gmail.com>
This commit is contained in:
littsk
2023-11-03 13:32:43 +08:00
committed by GitHub
parent d99b2c961a
commit 1a3315e336
30 changed files with 1120 additions and 552 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
colossalai/shardformer/policies/vit.py
View File

@@ -159,9 +159,6 @@ class ViTPolicy(Policy):
# ViTModel
class ViTModelPolicy(ViTPolicy):
def __init__(self) -> None:
super().__init__()
def module_policy(self):
from transformers.models.vit.modeling_vit import ViTModel
@@ -227,9 +224,6 @@ class ViTForImageClassificationPolicy(ViTPolicy):
# ViTForMaskedImageModeling
class ViTForMaskedImageModelingPolicy(ViTPolicy):
def __init__(self) -> None:
super().__init__()
def module_policy(self):
from transformers.models.vit.modeling_vit import ViTForMaskedImageModeling, ViTModel