Skip to content
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

Εnable reward model offloading option #930

Merged
merged 6 commits into from
Oct 29, 2024
Merged

Εnable reward model offloading option #930

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
144e913
enable reward model offloading option
kfertakis Oct 1, 2024
82c59ab
fixed code formatting
kfertakis Oct 14, 2024
2babdd3
more formatting fixes
kfertakis Oct 15, 2024
14960f1
Merge branch 'master' into kfertakis/rlhf_reward_offload
tjruwase Oct 29, 2024
83eb5e6
Merge branch 'master' into kfertakis/rlhf_reward_offload
loadams Oct 29, 2024
9a44a43
Pre-commit formatting fix
loadams Oct 29, 2024
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
17 changes: 4 additions & 13 deletions applications/DeepSpeed-Chat/dschat/rlhf/rlhf_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -268,31 +268,22 @@ def _init_reward(self, critic_model_name_or_path):
# If critic is ZeRO-3 then we use it for everything, otherwise assume we have enough memory
zero_stage = 0

ds_config = get_eval_ds_config(offload=self.args.offload,
dtype=self.args.dtype,
stage=zero_stage)
ds_config[
'train_micro_batch_size_per_gpu'] = self.args.per_device_training_batch_size
ds_config[
'train_batch_size'] = self.args.per_device_training_batch_size * torch.distributed.get_world_size(
) * self.args.gradient_accumulation_steps

ds_eval_config = get_eval_ds_config(offload=False,
ds_config = get_eval_ds_config(offload=self.args.offload_reward_model,
dtype=self.args.dtype,
stage=zero_stage)

# We need to set train batch size and micro batch size here to pass the sanity check of DeepSpeed engine.
ds_eval_config[
ds_config[
'train_micro_batch_size_per_gpu'] = self.args.per_device_training_batch_size
ds_eval_config[
ds_config[
'train_batch_size'] = self.args.per_device_training_batch_size * torch.distributed.get_world_size(
) * self.args.gradient_accumulation_steps

# Model
reward_model = create_critic_model(
model_name_or_path=critic_model_name_or_path,
tokenizer=self.tokenizer,
ds_config=ds_eval_config,
ds_config=ds_config,
num_padding_at_beginning=self.args.num_padding_at_beginning,
rlhf_training=True,
dropout=self.args.critic_dropout,
Expand Down
4 changes: 4 additions & 0 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 @@ -246,6 +246,10 @@ def parse_args():
'--offload_reference_model',
action='store_true',
help='Enable ZeRO Offload techniques for reference model')
parser.add_argument(
'--offload_reward_model',
action='store_true',
help='Enable ZeRO Offload techniques for reward model')
parser.add_argument(
'--actor_zero_stage',
type=int,
Expand Down
Loading