Skip to content

saving hidden_states.clone() #7705

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
Jul 5, 2025
Merged

saving hidden_states.clone() #7705

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
4 changes: 2 additions & 2 deletions python/sglang/srt/layers/logits_processor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -436,8 +436,8 @@ def _get_logits(
if self.do_tensor_parallel_all_gather_dp_attn:
logits_metadata.compute_dp_attention_metadata(hidden_states)
hidden_states, local_hidden_states = (
logits_metadata.gathered_buffer,
hidden_states.clone(),
torch.empty_like(logits_metadata.gathered_buffer),
hidden_states,
)
dp_gather_replicate(hidden_states, local_hidden_states, logits_metadata)

Expand Down
5 changes: 0 additions & 5 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 @@ -1780,11 +1780,6 @@ def forward(
hidden_states, residual, forward_batch
)

if self.enable_dp_attention and self.speculative_algorithm.is_eagle():
# NOTE: this line resolves the degradation of MTP reception rate for non-zero DP ranks.
# See discussion here (https://github.com/sgl-project/sglang/pull/6081#discussion_r2147452251).
hidden_states = hidden_states.clone()

return hidden_states, residual

def op_comm_prepare_attn(
Expand Down
Loading