diff --git a/backends/gaudi/server/text_generation_server/models/__init__.py b/backends/gaudi/server/text_generation_server/models/__init__.py
index ddae0a96..6ca7b567 100644
--- a/backends/gaudi/server/text_generation_server/models/__init__.py
+++ b/backends/gaudi/server/text_generation_server/models/__init__.py
@@ -38,6 +38,7 @@ __all__ = [
 ]
 from text_generation_server.models.globals import ATTENTION
 
+VLM_BATCH_TYPES = set()
 FLASH_ATT_ERROR_MESSAGE = "{} requires Flash Attention enabled models."
 
 FLASH_ATTENTION = False
@@ -155,6 +156,9 @@ if FLASH_ATTENTION:
     }
 
 
+__all__.append(VLM_BATCH_TYPES)
+
+
 class ModelType(enum.Enum):
     DEEPSEEK_V2 = {
         "type": "deepseek_v2",