diff --git a/tensorrt_llm/_torch/auto_deploy/transform/library/fuse_mamba_a_log.py b/tensorrt_llm/_torch/auto_deploy/transform/library/fuse_mamba_a_log.py
index 5e967e1f35e..977aeeabc5c 100644
--- a/tensorrt_llm/_torch/auto_deploy/transform/library/fuse_mamba_a_log.py
+++ b/tensorrt_llm/_torch/auto_deploy/transform/library/fuse_mamba_a_log.py
@@ -31,31 +31,11 @@
 
 from ...models.factory import ModelFactory
 from ...shim.interface import CachedSequenceInterface
+from ...utils._graph import del_attr_by_name, get_attr_by_name, set_attr_by_name
 from ...utils.logger import ad_logger
 from ...utils.pattern_matcher import ADPatternMatcherPass
 from ..interface import BaseTransform, SharedConfig, TransformInfo, TransformRegistry
 
-
-def _get_attr_by_name(obj, name):
-    for part in name.split("."):
-        obj = getattr(obj, part)
-    return obj
-
-
-def _set_attr_by_name(obj, name, value):
-    parts = name.split(".")
-    for part in parts[:-1]:
-        obj = getattr(obj, part)
-    setattr(obj, parts[-1], value)
-
-
-def _del_attr_by_name(obj, name):
-    parts = name.split(".")
-    for part in parts[:-1]:
-        obj = getattr(obj, part)
-    delattr(obj, parts[-1])
-
-
 _PATTERN_INPUT_NAME = "a_log_like"
 
 
@@ -82,13 +62,13 @@ def _ensure_a_fused_param(gm: GraphModule, param_name: str) -> Optional[str]:
 
     new_param_name = param_name.replace("A_log", "A_fused")
     try:
-        _get_attr_by_name(gm, new_param_name)
+        get_attr_by_name(gm, new_param_name)
         return new_param_name
     except AttributeError:
         pass
 
     try:
-        a_log = _get_attr_by_name(gm, param_name)
+        a_log = get_attr_by_name(gm, param_name)
     except AttributeError:
         ad_logger.warning(f"Could not find attribute {param_name} in gm.")
         return None
@@ -96,7 +76,7 @@ def _ensure_a_fused_param(gm: GraphModule, param_name: str) -> Optional[str]:
     with torch.no_grad():
         a_fused = -torch.exp(a_log.float())
 
-    _set_attr_by_name(
+    set_attr_by_name(
         gm,
         new_param_name,
         nn.Parameter(a_fused, requires_grad=False),
@@ -120,7 +100,7 @@ def _maybe_remove(name: str) -> None:
         if not name.endswith("A_log") or name in used_a_log_targets:
             return
         try:
-            _del_attr_by_name(gm, name)
+            del_attr_by_name(gm, name)
             removed = True
         except AttributeError:
             ad_logger.warning(f"Failed to delete unused parameter {name} from GraphModule.")
diff --git a/tensorrt_llm/_torch/auto_deploy/transform/library/sharding.py b/tensorrt_llm/_torch/auto_deploy/transform/library/sharding.py
index 3a5d606aebe..deb82a43e66 100644
--- a/tensorrt_llm/_torch/auto_deploy/transform/library/sharding.py
+++ b/tensorrt_llm/_torch/auto_deploy/transform/library/sharding.py
@@ -33,6 +33,7 @@
 from ...custom_ops.trtllm_dist import is_trtllm_op_available
 from ...models.factory import ModelFactory, ShardingConfigSource
 from ...shim.interface import CachedSequenceInterface
+from ...utils._graph import del_attr_by_name
 from ...utils.logger import ad_logger
 from ...utils.node_utils import (
     LayerSubgraph,
@@ -1467,7 +1468,12 @@ def get_partition(lst, world_size, rank):
     for expert in (
         w_up_list_to_remove + w_down_list_to_remove + w_gate_list_to_remove + scales_to_remove
     ):
-        delattr(gm, expert.target)
+        try:
+            del_attr_by_name(gm, expert.target)
+        except AttributeError:
+            ad_logger.warning(
+                f"Failed to delete unused parameter {expert.target} from GraphModule."
+            )
 
 
 def _slice_expert_dim(gm: GraphModule, tensor_node: Node, lo: int, hi: int) -> Node:
diff --git a/tensorrt_llm/_torch/auto_deploy/utils/_graph.py b/tensorrt_llm/_torch/auto_deploy/utils/_graph.py
index cd61bd52f1e..04db9bd80d8 100644
--- a/tensorrt_llm/_torch/auto_deploy/utils/_graph.py
+++ b/tensorrt_llm/_torch/auto_deploy/utils/_graph.py
@@ -401,3 +401,55 @@ def get_lm_head_weights(model: nn.Module) -> torch.Tensor:
     gm, output_node = get_output_node(model)
     lm_head_node = get_lm_head_node(gm, output_node)
     return get_weight_tensor(gm, lm_head_node)
+
+
+def get_attr_by_name(obj, name):
+    """Get an attribute specified by a dot-separated path on an object.
+
+    Args:
+        obj: The root object from which to resolve the attribute path.
+        name (str): Dot-separated attribute path (e.g., "a.b.c").
+
+    Returns:
+        The value of the resolved attribute.
+
+    Raises:
+        AttributeError: If any component in the path does not exist.
+    """
+    for part in name.split("."):
+        obj = getattr(obj, part)
+    return obj
+
+
+def set_attr_by_name(obj, name, value):
+    """Set an attribute specified by a dot-separated path on an object.
+
+    Args:
+        obj: The root object on which to set the attribute.
+        name (str): Dot-separated attribute path (e.g., "a.b.c").
+        value: The value to assign to the target attribute.
+
+    Raises:
+        AttributeError: If any intermediate component in the path does not exist.
+    """
+    parts = name.split(".")
+    for part in parts[:-1]:
+        obj = getattr(obj, part)
+    setattr(obj, parts[-1], value)
+
+
+def del_attr_by_name(obj, name):
+    """Delete an attribute specified by a dot-separated path from an object.
+
+    Args:
+        obj: The root object from which to delete the attribute.
+        name (str): Dot-separated attribute path (e.g., "a.b.c").
+
+    Raises:
+        AttributeError: If any intermediate component in the path does not exist
+            or if the final attribute does not exist.
+    """
+    parts = name.split(".")
+    for part in parts[:-1]:
+        obj = getattr(obj, part)
+    delattr(obj, parts[-1])