Skip to content

[Bugfix] DeepSeek V4: skip expert tensor when no mapping succeeds (AMD + NVIDIA) #44030

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
dparikh79 wants to merge 1 commit into
base: main
Choose a base branch
from
+8 −0
Open

[Bugfix] DeepSeek V4: skip expert tensor when no mapping succeeds (AMD + NVIDIA) #44030

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: 4 additions & 0 deletions vllm/models/deepseek_v4/amd/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -822,6 +822,7 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
and loaded_weight.dtype == torch.float8_e8m0fnu
):
loaded_weight = loaded_weight.view(torch.uint8)
success = False
for mapping in expert_mapping:
param_name, weight_name, expert_id, expert_shard_id = mapping
if weight_name not in name:
Expand All @@ -847,6 +848,9 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
if success:
name = name_mapped
break
if not success:
# Skip when no mapping bound name_mapped (#42769).
continue
loaded_params.add(name_mapped)
continue
elif "attn_sink" in name:
Expand Down
4 changes: 4 additions & 0 deletions vllm/models/deepseek_v4/nvidia/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1154,6 +1154,7 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
and loaded_weight.dtype == torch.float8_e8m0fnu
):
loaded_weight = loaded_weight.view(torch.uint8)
success = False
for mapping in expert_mapping:
param_name, weight_name, expert_id, expert_shard_id = mapping
if weight_name not in name:
Expand All @@ -1179,6 +1180,9 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
if success:
name = name_mapped
break
if not success:
# Skip when no mapping bound name_mapped (#42769).
continue
loaded_params.add(name_mapped)
continue
elif "attn_sink" in name:
Expand Down
Loading