fix clip_vision trouble

src/configuration_medclip.py

@@ -2,7 +2,7 @@ import copy
 
 from transformers.configuration_utils import PretrainedConfig
 from transformers.utils import logging
-
+from transformers import AutoConfig, CLIPVisionConfig
 
 logger = logging.get_logger(__name__)
 
@@ -69,12 +69,12 @@ class MedCLIPConfig(PretrainedConfig):
         text_model_type = text_config.pop("model_type")
         vision_model_type = vision_config.pop("model_type")
 
-        from transformers import AutoConfig
-
         self.text_config = AutoConfig.for_model(text_model_type, **text_config)
 
         if vision_model_type == "clip":
             self.vision_config = AutoConfig.for_model(vision_model_type, **vision_config).vision_config
+        elif vision_model_type == "clip_vision_model":
+            self.vision_config = CLIPVisionConfig(**vision_config)
         else:
             self.vision_config = AutoConfig.for_model(vision_model_type, **vision_config)