Fix parameter order in GPT-OSS weight loading function for non-MXFP4 weights (#29506)

Signed-off-by: Filipp Fisin <[email protected]>
Co-authored-by: Cyrus Leung <[email protected]>

Author: qGentry

vllm/model_executor/models/gpt_oss.py

@@ -647,8 +647,8 @@ def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
             )
         else:
             return self._load_weights_other(
-                ep_rank_start,
                 ep_rank_end,
+                ep_rank_start,
                 heads_per_rank,
                 head_start,
                 weights,