update

app_idefics2.py

@@ -74,16 +74,25 @@ def get_chat_history(history):
                     "content": []
                 }
             )
+            print(num_images, cur_image_idx, len(images))
             assert num_images + cur_image_idx <= len(images), f"Number of images uploaded is less than the number of <image> placeholders in the text. Please upload more images."
             if num_images > 0:
-                for sub_text in message[0].split("<image>"):
-                    if sub_text.strip():
-                        messages[-1]["content"].append({"type": "text", "text": sub_text.strip()})
-                    if cur_image_idx < len(images):
-                        messages[-1]["content"].append({"type": "image"})
-                        cur_image_idx += 1
+                split_text = message[0].split("<image>")
+                if split_text[0].strip():
+                    messages[-1]["content"].append({"type": "text", "text": split_text[0].strip()})
+                for idx in range(num_images):
+                    messages[-1]["content"].append({"type": "image"})
+                    if split_text[idx + 1].strip():
+                        messages[-1]["content"].append({"type": "text", "text": split_text[idx + 1].strip()})
             else:
                 messages[-1]["content"].append({"type": "text", "text": message[0]})
+            if message[1]:
+                messages.append(
+                    {
+                        "role": "assistant",
+                        "content": [{"type": "text", "text": message[1]}]
+                    }
+                )
         elif isinstance(message[0], tuple):
             pass
     return messages, images