Update modeling_gpt2vision.py

modeling_gpt2vision.py

@@ -66,7 +66,8 @@ class GPT2Vision(PreTrainedModel):
             max_length=768,
             return_tensors="pt",
         ).to(device)
-        
+
+        print("text_inputs",text_inputs)
         # Adjust attention mask to account for image tokens and the extra <image> token
         batch_size = text_inputs.input_ids.shape[0]
         img_attention = torch.ones((batch_size, self.img_tokens + 1), dtype=torch.long, device=device)