Update app.py

app.py

@@ -236,25 +236,40 @@ def clear_response(history):
 
 def add_message(history, message):
     # history=[]
-    global our_chatbot
+    global our_chatbot, chat_image_num
     if len(history) == 0:
         our_chatbot = InferenceDemo(
             args, model_path, tokenizer, model, image_processor, context_len
         )
-
+    
     if len(message["files"]) <= 1:
         for x in message["files"]:
             history.append(((x,), None))
+            chat_image_num += 1
+        if chat_image_num > 1:
+            history = []
+            our_chatbot = InferenceDemo(
+                args, model_path, tokenizer, model, image_processor, context_len
+            )
+        for x in message["files"]:
+            history.append(((x,), None))
+            chat_image_num += 1
+        if message["text"] is not None:
+            history.append((message["text"], None))
+        return history, gr.MultimodalTextbox(value=None, interactive=False)
     else:
-        return None, gr.MultimodalTextbox(value=None, interactive=False)
-    # elif len(message["files"]) > 1:
-    #     history.append((tuple(message["files"]), None))
-    # if len(message["files"]) > 1:
-    #     return
-    if message["text"] is not None:
-        history.append((message["text"], None))
+        # for x in message["files"]:
+        history.append(((message["files"],), None))
+        chat_image_num += len(message["files"])
+        if message["text"] is not None:
+            history.append((message["text"], None))
+            
+        return history, gr.MultimodalTextbox(value=None, interactive=False)
+    
+    # if message["text"] is not None:
+    #     history.append((message["text"], None))
         
-    return history, gr.MultimodalTextbox(value=None, interactive=False)
+    # return history, gr.MultimodalTextbox(value=None, interactive=False)
 
 
 @spaces.GPU