fixed the kwargs passing to build_decode_specialization

Signed-off-by: Onkar Chougule <ochougul@qti.qualcomm.com>

Author: ochougul

QEfficient/transformers/models/modeling_auto.py

@@ -2834,6 +2834,7 @@ def build_decode_specialization(
         kv_cache_batch_size: Optional[int] = None,
         full_batch_size: Optional[int] = None,
         num_speculative_tokens: Optional[int] = None,
+        **kwargs,
     ):
         """
         Builds a dictionary representing a compilation specialization for the decode phase.