Skip to content

Remove unnecessary torch.full in DeepSeek #5601

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 2 commits into from
Apr 23, 2025
Merged

Remove unnecessary torch.full in DeepSeek #5601

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
61d576a
more
fzyzcjy Apr 21, 2025
4a7ec4d
Merge branch 'main' into feat/deepep_dispatch
zhyncs Apr 23, 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
13 changes: 7 additions & 6 deletions python/sglang/srt/models/deepseek_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -323,12 +323,6 @@ def forward_deepep(
self, hidden_states: torch.Tensor, forward_mode: ForwardMode
) -> torch.Tensor:
shared_output = None
topk_idx = torch.full(
(0, self.top_k), -1, dtype=torch.int, device=hidden_states.device
)
topk_weights = torch.empty(
(0, self.top_k), dtype=torch.float32, device=hidden_states.device
)
if (
forward_mode is not None
and not forward_mode.is_idle()
Expand All @@ -348,6 +342,13 @@ def forward_deepep(
correction_bias=self.correction_bias,
routed_scaling_factor=self.routed_scaling_factor,
)
else:
topk_idx = torch.full(
(0, self.top_k), -1, dtype=torch.int, device=hidden_states.device
)
topk_weights = torch.empty(
(0, self.top_k), dtype=torch.float32, device=hidden_states.device
)
if self.ep_size > 1:
# TODO(ch-wan): allow users to set num_max_dispatch_tokens_per_rank value
(
Expand Down
Loading