Update optimum/intel/openvino/modeling_visual_language.py

Co-authored-by: Vladimir Zlobin <vladimir.zlobin@intel.com>

Author: eaidova

optimum/intel/openvino/modeling_visual_language.py

@@ -585,8 +585,8 @@ def forward(
             attention_mask=attention_mask,
             position_ids=position_ids,
             past_key_values=past_key_values,
-            image_bound=None,
-            tgt_sizes=None,
+            image_bound=image_bound,
+            tgt_sizes=tgt_sizes,
             **kwargs,
         )
         return self.language_model.forward(