A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 13, 02:00 UTC
Fix gpt oss weight loading with EP + bf16 (#28765)
Signed-off-by: ashors1 <ashors@nvidia.com>
Anna Shors committed
Nov 16, 2025 at 05:12 UTC
8d259fad6cd5a93bef04d00640e132e84c0c9b20
1 file changed
+1
-1
vllm/model_executor/models/gpt_oss.py
+1
-1
@@ -494,8 +494,8 @@ class GptOssModel(nn.Module):
494
495
def _load_weights_other(
496
self,
497
- ep_rank_start: int,
497
ep_rank_end: int,
498
+ ep_rank_start: int,
499
heads_per_rank: int,
500
head_start: int,
501
weights: Iterable[tuple[str, torch.Tensor]],