Fix double lm_head mapping: NVFP4 checkpoint already uses correct names

The checkpoint has lm_head.weight and model.embed_tokens.weight
already — the suffix mappings head.weight→lm_head.weight and
embed.weight→embed_tokens.weight were incorrectly applying to keys
that already had the right prefix, producing lm_lm_head.weight.

vllm/patches/deepseek_v4.py

@@ -1645,8 +1645,8 @@ def _make_deepseek_v4_nvfp4_weights_mapper() -> WeightsMapper:
         },
         orig_to_new_regex=expert_rename_regex,
         orig_to_new_suffix={
-            "head.weight": "lm_head.weight",
-            "embed.weight": "embed_tokens.weight",
+            # NVFP4 checkpoint already uses lm_head.weight and
+            # model.embed_tokens.weight — no suffix renames needed.
             ".ffn.gate.bias": ".ffn.gate.e_score_correction_bias",
         },
         orig_to_new_substr={