DeepSpeed
fix regression in shard checkpoint loading in AutoTP Path caused by qkv_copy() is deleted and add UT case for shard checkpoint loading in AutoTP
#3457
Merged

fix regression in shard checkpoint loading in AutoTP Path caused by qkv_copy() is deleted and add UT case for shard checkpoint loading in AutoTP #3457

sywangyi
sywangyi sywangyi requested a review from RezaYazdaniAminabadi RezaYazdaniAminabadi 2 years ago
sywangyi sywangyi requested a review from jeffra jeffra 2 years ago
sywangyi sywangyi requested a review from mrwyattii mrwyattii 2 years ago
sywangyi sywangyi requested a review from awan-10 awan-10 2 years ago
sywangyi sywangyi requested a review from cmikeh2 cmikeh2 2 years ago
sywangyi sywangyi requested a review from arashb arashb 2 years ago
sywangyi sywangyi requested a review from tjruwase tjruwase 2 years ago
sywangyi
sywangyi sywangyi force pushed from e20dc1f3 to 7b011834 2 years ago
tjruwase
tjruwase commented on 2023-05-05
sywangyi sywangyi force pushed from 7e168f1d to 24d128be 2 years ago
sywangyi sywangyi force pushed from 24d128be to 72a49c66 2 years ago
sywangyi sywangyi requested a review from tjruwase tjruwase 2 years ago
sywangyi sywangyi changed the title add UT case for shard checkpoint loading in AutoTP add UT case for shard checkpoint loading in AutoTP and fix regression in shard checkpoint loading in AutoTP Path caused by qkv_copy() is deleted 2 years ago
sywangyi sywangyi changed the title add UT case for shard checkpoint loading in AutoTP and fix regression in shard checkpoint loading in AutoTP Path caused by qkv_copy() is deleted fix regression in shard checkpoint loading in AutoTP Path caused by qkv_copy() is deleted and add UT case for shard checkpoint loading in AutoTP 2 years ago
tjruwase
tjruwase approved these changes on 2023-05-09
tjruwase tjruwase added merge-queue
sywangyi add UT case for shard checkpoint loading in AutoTP
5b901634
sywangyi autoTP path also support shard loading
7b2f78e8
sywangyi sywangyi force pushed from 9724a4d2 to 7b2f78e8 2 years ago
sywangyi
tjruwase tjruwase merged b31b46c0 into master 2 years ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone