Update utils/model.py

utils/model.py

@@ -429,6 +429,7 @@ class OwlViTForClassification(nn.Module):
             position_ids = position_ids.to(device_)
             position_embedding = position_ids.repeat(position_ids.size(0), 1, 1)
             text_inputs_parts["position_ids"] = position_ids
+            print(f"position_embedding : {position_embedding.shape}")
             print(f"pos + emb: {(txt_embeds + position_embedding).shape}")
             text_embeds_parts = self.owlvit.text_model.get_text_features(**text_inputs_parts)