Add new fusion argument from onnxruntime v1.16.2

optimum/onnxruntime/configuration.py

@@ -708,6 +708,8 @@ class OptimizationConfig:
             Do not fuse GroupNorm. Only works for model_type=unet.
         disable_packed_kv (`bool`, defaults to `True`):
             Do not use packed kv in cross attention. Only works for model_type=unet.
+        disable_rotary_embeddings (`bool`, defaults to `False`):
+            Whether to disable Rotary Embedding fusion.
     """
 
     optimization_level: int = 1
@@ -752,6 +754,9 @@ class OptimizationConfig:
     disable_group_norm_fusion: bool = True
     disable_packed_kv: bool = True
 
+    # ONNX Runtime 1.16.2 arguments
+    disable_rotary_embeddings: bool = False
+
     def __post_init__(self):
         def deprecate_renamed_attribute(old_name, new_name, mapping_func=None):
             if getattr(self, old_name, None) is not None:
@@ -801,6 +806,7 @@ class Box:
             "use_raw_attention_mask": "use_raw_attention_mask",
             "enable_gemm_fast_gelu_fusion": "enable_gemm_fast_gelu",
             "use_multi_head_attention": "use_multi_head_attention",
+            "disable_rotary_embeddings": "disable_rotary_embeddings",
         }
         for attr_name, fusion_attr_name in attribute_map.items():
             setattr(args, fusion_attr_name, getattr(self, attr_name))