Skip to content

Add the getter and setter of skip_fp8_weight_update_tensor #3015

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
xrennvidia wants to merge 4 commits into
base: main
Choose a base branch
from
+16 −9
Open

Add the getter and setter of skip_fp8_weight_update_tensor #3015

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
f332287
add the getter and setter of skip_fp8_weight_update_tensor
xrennvidia May 20, 2026
a5e984b
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] May 20, 2026
25e01dd
Update transformer_engine/pytorch/quantization.py
xrennvidia May 20, 2026
b265f25
Merge branch 'main' into xren/fix_skip_fp8_weight_update
xrennvidia May 22, 2026
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
11 changes: 2 additions & 9 deletions transformer_engine/pytorch/graph.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -324,12 +324,7 @@ def _make_graphed_callables(

if cache_quantized_params:
# Initialize flag that controls FP8 weight updates
qstate = FP8GlobalStateManager.quantization_state
if qstate.skip_fp8_weight_update_tensor is None:
qstate.skip_fp8_weight_update_tensor = torch.empty(
1, dtype=torch.float32, device="cuda"
)
qstate.skip_fp8_weight_update_tensor.fill_(False)
FP8GlobalStateManager.set_skip_fp8_weight_update_tensor(False)

# Check callables
for c in callables:
Expand Down Expand Up @@ -841,9 +836,7 @@ def forward(ctx, skip_fp8_weight_update, cuda_graph_stream, cuda_graph_event, *i
# Set flag for whether to update FP8 weight updates
ctx.is_first_module = FP8GlobalStateManager.is_first_fp8_module()
if ctx.is_first_module and skip_fp8_weight_update is not None:
FP8GlobalStateManager.quantization_state.skip_fp8_weight_update_tensor.fill_(
skip_fp8_weight_update
)
FP8GlobalStateManager.set_skip_fp8_weight_update_tensor(skip_fp8_weight_update)
ctx.cuda_graph_stream = cuda_graph_stream
ctx.cuda_graph_event = cuda_graph_event
# Copy values from new tensors into static tensors
Expand Down
14 changes: 14 additions & 0 deletions transformer_engine/pytorch/quantization.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -409,6 +409,20 @@ class FP8GlobalStateManager:

quantization_state = FP8GlobalState()

@classmethod
def set_skip_fp8_weight_update_tensor(cls, skip: bool) -> None:
"""Set the skip fp8 weight update tensor"""
if cls.quantization_state.skip_fp8_weight_update_tensor is None:
cls.quantization_state.skip_fp8_weight_update_tensor = torch.empty(
1, dtype=torch.float32, device="cuda"
)
cls.quantization_state.skip_fp8_weight_update_tensor.fill_(skip)

@classmethod
def get_skip_fp8_weight_update_tensor(cls) -> Optional[torch.Tensor]:
"""Get the skip fp8 weight update tensor"""
return cls.quantization_state.skip_fp8_weight_update_tensor

@classmethod
def reset(cls) -> None:
"""Reset the global state"""
Expand Down
Loading