Update README.md

README.md

@@ -3698,6 +3698,14 @@ The `GME` models support three types of input: **text**, **image**, and **image-
 
 ```python
 from transformers import AutoModel
+from transformers.utils.versions import require_version
+
+
+require_version(
+    "transformers<4.52.0",
+    "GmeQwen2VL has some issues with transformers>=4.52.0, please downgrade: pip install transformers==4.51.3"
+)
+
 
 t2i_prompt = 'Find an image that matches the given text.'
 texts = [