Universal Checkpoint for Sequence Parallelism #305
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
samadejacobs
requested review from
jeffra,
tjruwase,
ShadenSmith,
conglongli,
awan-10,
eltonzheng,
minjiaz,
RezaYazdaniAminabadi,
duli2012,
mrwyattii,
arashb,
xiaoxiawu-microsoft and
GuanhuaWang
as code owners
tjruwase
removed request for
arashb,
ShadenSmith,
jeffra,
duli2012,
conglongli,
awan-10,
GuanhuaWang,
mrwyattii,
eltonzheng,
minjiaz and
RezaYazdaniAminabadi
tjruwase
requested review from
lekurile
and removed request for
xiaoxiawu-microsoft
tjruwase
reviewed
Dec 11, 2023
tjruwase
reviewed
Dec 11, 2023
examples_deepspeed/universal_checkpointing/run_universal_fp16.sh
Outdated
Show resolved
Hide resolved
tjruwase
reviewed
Dec 11, 2023
zdaiot
pushed a commit
to zdaiot/Megatron-DeepSpeed
that referenced
this pull request
Jan 20, 2024
* Extend universal checkpoint support for ds sequence parallelism (SP) and ZeRO stage 2 * Extend universal checkpoint support for ds sequence parallelism (SP) and ZeRO stage 2 * Extend README and batch scripts discussion * Extend README and batch scripts discussion * Extend README and batch scripts discussion * Remove debug statement * Script no pipeline parallel for ZeRO stage 2
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR extends the universal checkpoint to support DS sequence parallelism and training scenarios where pipeline parallelism is not enabled.
The attached Tensorboard chart show a training scenario (validation curve) where a GPT model is pre-trained with data parallelism (4 GPUs), and checkpoints are saved at the 100th and 200th iterations. The checkpoint at the 100th iteration is later loaded for continual pre-training with different configurations (more GPU resources, data parallelism = 4 GPUs, sequence parallelism = 2 GPUs).