Skip to content

raise error if fsdp_cpu_offload + opt_in_bwd #2826

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

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

raise error if fsdp_cpu_offload + opt_in_bwd #2826

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
901723f
rfc
Jun 4, 2025
ab02d75
Revert "rfc"
Jun 4, 2025
1c0117b
Merge branch 'main' of https://github.com/pytorch/torchtune
Jun 6, 2025
1cc6946
Merge branch 'main' of https://github.com/pytorch/torchtune
Jun 11, 2025
08f5dc9
Merge branch 'main' of https://github.com/pytorch/torchtune
Jun 13, 2025
dd99edd
raise error if opt_in_bwd and cpu offload
Jun 13, 2025
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
5 changes: 5 additions & 0 deletions recipes/full_dpo_distributed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -177,6 +177,11 @@ def __init__(self, cfg: DictConfig) -> None:
"Gradient accumulation is not supported with optimizer in bwd."
"Please set gradient_accumulation_steps=1, or optimizer_in_bwd=False."
)
if self.fsdp_cpu_offload:
raise RuntimeError(
"CPU offload is not supported with optimizer in bwd atm."
"Please set fsdp_cpu_offload=False, or optimizer_in_bwd=False."
)

# activation checkpointing/offloading
self._enable_activation_checkpointing = cfg.get(
Expand Down
5 changes: 5 additions & 0 deletions recipes/full_finetune_distributed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -226,6 +226,11 @@ def __init__(self, cfg: DictConfig) -> None:
"Gradient accumulation is not supported with optimizer in bwd."
"Please set gradient_accumulation_steps=1, or optimizer_in_bwd=False."
)
if self.fsdp_cpu_offload:
raise RuntimeError(
"CPU offload is not supported with optimizer in bwd atm."
"Please set fsdp_cpu_offload=False, or optimizer_in_bwd=False."
)

# activation checkpointing/offloading
self._enable_activation_checkpointing = cfg.get(
Expand Down
5 changes: 5 additions & 0 deletions recipes/qat_distributed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -232,6 +232,11 @@ def __init__(self, cfg: DictConfig) -> None:
"Gradient accumulation is not supported with optimizer in bwd."
"Please set gradient_accumulation_steps=1, or optimizer_in_bwd=False."
)
if self.fsdp_cpu_offload:
raise RuntimeError(
"CPU offload is not supported with optimizer in bwd atm."
"Please set fsdp_cpu_offload=False, or optimizer_in_bwd=False."
)

self._enable_async_checkpointing = cfg.get("enable_async_checkpointing", False)
self._checkpoint_client = CheckpointClient(cfg)
Expand Down
Loading