Update utils/model.py

utils/model.py

@@ -420,7 +420,7 @@ class OwlViTForClassification(nn.Module):
             #DEUBUG:
             print(f"text_inputs_parts - input_ids: {text_inputs_parts['input_ids'].shape}. attention_mask : {text_inputs_parts['attention_mask'].shape}")
             print(f"position_embedding: {self.owlvit.position_embedding()}")
-            text_embeds_parts = self.owlvit.get_text_features(**text_inputs_parts)
+            text_embeds_parts = self.owlvit.text_model.text_model.get_text_features(**text_inputs_parts)
             
             # # Embed images and text queries
             query_mask, text_embeds_parts = self._get_text_query_mask(text_inputs_parts, text_embeds_parts, batch_size)