Skip to content

[chat] enable zero-offload support #553

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
May 26, 2023
Merged

[chat] enable zero-offload support #553

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 0 additions & 2 deletions applications/DeepSpeed-Chat/training/step1_supervised_finetuning/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -198,8 +198,6 @@ def main():
# If passed along, set the training seed now.
set_random_seed(args.seed)

assert not args.offload, "zero-offload is not currently supported but coming soon!"

torch.distributed.barrier()

tokenizer = load_hf_tokenizer(args.model_name_or_path, fast_tokenizer=True)
Expand Down
2 changes: 0 additions & 2 deletions applications/DeepSpeed-Chat/training/step2_reward_model_finetuning/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -187,8 +187,6 @@ def main():

args.global_rank = torch.distributed.get_rank()

assert not args.offload, "zero-offload is not currently supported but coming soon!"

ds_config = get_train_ds_config(offload=args.offload,
stage=args.zero_stage)
ds_config[
Expand Down
2 changes: 0 additions & 2 deletions applications/DeepSpeed-Chat/training/step3_rlhf_finetuning/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -365,8 +365,6 @@ def main():

args.global_rank = torch.distributed.get_rank()

assert not args.offload, "zero-offload is not currently supported but coming soon!"

unsupervised_training_enabled = args.unsupervised_dataset_name and args.unsupervised_dataset_config_name
if unsupervised_training_enabled:
# if we enable unsupervised training, we need to double the batch size for actor model
Expand Down