diff --git a/src/otx/algo/segmentation/backbones/litehrnet.py b/src/otx/algo/segmentation/backbones/litehrnet.py
index 8de1a58bc09..7b8ffd450e1 100644
--- a/src/otx/algo/segmentation/backbones/litehrnet.py
+++ b/src/otx/algo/segmentation/backbones/litehrnet.py
@@ -14,7 +14,6 @@
 
 import torch
 import torch.utils.checkpoint as cp
-from mmengine.utils import is_tuple_of
 from torch import nn
 from torch.nn import functional
 
@@ -153,9 +152,6 @@ def __init__(
         if len(act_cfg) != 2:
             msg = "act_cfg must be a dict or a tuple of dicts of length 2."
             raise ValueError(msg)
-        if not is_tuple_of(act_cfg, dict):
-            msg = "act_cfg must be a dict or a tuple of dicts."
-            raise TypeError(msg)
 
         self.channels = channels
         total_channel = sum(channels)
@@ -226,9 +222,6 @@ def __init__(
         if len(act_cfg) != 2:
             msg = "act_cfg must be a dict or a tuple of dicts of length 2."
             raise ValueError(msg)
-        if not is_tuple_of(act_cfg, dict):
-            msg = "act_cfg must be a dict or a tuple of dicts."
-            raise TypeError(msg)
 
         self.global_avgpool = nn.AdaptiveAvgPool2d(1)
         self.conv1 = ConvModule(
diff --git a/src/otx/algo/segmentation/heads/ham_head.py b/src/otx/algo/segmentation/heads/ham_head.py
index 8a9f7e42bed..68c7b006cba 100644
--- a/src/otx/algo/segmentation/heads/ham_head.py
+++ b/src/otx/algo/segmentation/heads/ham_head.py
@@ -9,7 +9,6 @@
 
 import torch
 import torch.nn.functional as f
-from mmengine.device import get_device
 from torch import nn
 
 from otx.algo.modules import ConvModule
@@ -223,7 +222,7 @@ def _build_bases(
         segments: int,
         channels: int,
         basis_vectors: int,
-        device: torch.device | None = None,
+        device: torch.device,
     ) -> torch.Tensor:
         """Build bases in initialization.
 
@@ -237,8 +236,6 @@ def _build_bases(
         Returns:
             torch.Tensor: Tensor of shape (batch_size * segments, channels, basis_vectors) containing the built bases.
         """
-        if device is None:
-            device = get_device()
         bases = torch.rand((batch_size * segments, channels, basis_vectors)).to(device)
 
         return f.normalize(bases, dim=1)
diff --git a/src/otx/recipe/_base_/data/mmseg_base.yaml b/src/otx/recipe/_base_/data/mmseg_base.yaml
index d6ed6336c23..55cd7c91a39 100644
--- a/src/otx/recipe/_base_/data/mmseg_base.yaml
+++ b/src/otx/recipe/_base_/data/mmseg_base.yaml
@@ -18,11 +18,14 @@ config:
           size:
             - 512
             - 512
+          scale:
+            - 0.2
+            - 1.0
           ratio:
             - 0.5
             - 2.0
           antialias: True
-      - class_path: torchvision.transforms.v2.RandomPhotometricDistort
+      - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
       - class_path: torchvision.transforms.v2.RandomHorizontalFlip
         init_args:
           p: 0.5
diff --git a/src/otx/recipe/semantic_segmentation/dino_v2.yaml b/src/otx/recipe/semantic_segmentation/dino_v2.yaml
index 05bd298200b..a33f8eb58bb 100644
--- a/src/otx/recipe/semantic_segmentation/dino_v2.yaml
+++ b/src/otx/recipe/semantic_segmentation/dino_v2.yaml
@@ -80,11 +80,14 @@ overrides:
               size:
                 - 560
                 - 560
+              scale:
+                - 0.2
+                - 1.0
               ratio:
                 - 0.5
                 - 2.0
               antialias: True
-          - class_path: torchvision.transforms.v2.RandomPhotometricDistort
+          - class_path: otx.core.data.transform_libs.torchvision.PhotoMetricDistortion
           - class_path: torchvision.transforms.v2.RandomHorizontalFlip
             init_args:
               p: 0.5