Skip to content

[NVIDIA] Fix Llama4 Scout FP4 functionality issues #21499

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 30, 2025
+219 −70
Merged

[NVIDIA] Fix Llama4 Scout FP4 functionality issues #21499

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
15 changes: 14 additions & 1 deletion vllm/model_executor/layers/fused_moe/layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -874,6 +874,14 @@ def _load_per_tensor_weight_scale(self, shard_id: str,
elif shard_id == "w2":
param_data[expert_id] = loaded_weight

def _load_w13_weight_scale(self, shard_dim: int,
loaded_weight: torch.Tensor,
param: torch.Tensor, tp_rank: int):
shard_size = param.shape[shard_dim]
loaded_weight = loaded_weight.narrow(shard_dim, shard_size * tp_rank,
shard_size)
param.copy_(loaded_weight)

def _load_model_weight_or_group_weight_scale(self,
shard_dim: int,
expert_data: torch.Tensor,
Expand Down Expand Up @@ -1123,7 +1131,12 @@ def weight_loader(self,
"weight_scale_2" in weight_name if uses_weight_scale_2 else
"weight_scale" in weight_name) or "input_scale" in weight_name

if per_tensor_conditions:
if "w13_weight_scale" in weight_name:
self._load_w13_weight_scale(shard_dim=shard_dim,
loaded_weight=loaded_weight,
param=param,
tp_rank=self.tp_rank)
elif per_tensor_conditions:
self._load_per_tensor_weight_scale(
shard_id=shard_id,
param=param,
Expand Down
2 changes: 0 additions & 2 deletions vllm/model_executor/layers/quantization/modelopt.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -778,8 +778,6 @@ def process_weights_after_loading(self, layer: Module) -> None:
# Swizzle the weight blockscale.
# contracting dimension is input dimension
# block_size = 16;
assert (layer.weight_scale.shape[1] % 16 == 0), (
"Expected weight_scale.dim(1) to be divisible by 16")
assert (layer.weight_scale.dtype == torch.float8_e4m3fn), (
"Weight Block scale must be represented as FP8-E4M3")
swizzled_weight_scale = swizzle_blockscale(layer.weight_scale)
Expand Down
Loading