Update app.py

app.py

@@ -230,9 +230,6 @@ def clear_response(history):
     return history, question
 
 
-# def print_like_dislike(x: gr.LikeData):
-#     print(x.index, x.value, x.liked)
-
 
 def add_message(history, message):
     # history=[]
@@ -259,19 +256,14 @@ def add_message(history, message):
             history.append((message["text"], None))
         return history, gr.MultimodalTextbox(value=None, interactive=False)
     else:
-        # for x in message["files"]:
-        history.append(((message["files"],), None))
-        chat_image_num += len(message["files"])
+        for x in message["files"]:
+            history.append(((x,), 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 None, gr.MultimodalTextbox(value=None, interactive=False)
+    
 
 @spaces.GPU
 def bot(history, temperature, top_p, max_output_tokens):
@@ -285,6 +277,7 @@ def bot(history, temperature, top_p, max_output_tokens):
     text_this_term = ""
     # import pdb;pdb.set_trace()
     num_new_images = 0
+    
     for i, message in enumerate(history[:-1]):
         if type(message[0]) is tuple:
             # print("### message[0]",message[0])
@@ -294,7 +287,6 @@ def bot(history, temperature, top_p, max_output_tokens):
             # else:
             images_this_term.append(message[0][0])
             if is_valid_video_filename(message[0][0]):
-                # 不接受视频
                 raise ValueError("Video is not supported")
                 num_new_images += our_chatbot.num_frames
             elif is_valid_image_filename(message[0][0]):
@@ -374,7 +366,7 @@ def bot(history, temperature, top_p, max_output_tokens):
     input_ids = tokenizer_image_token(
             prompt, our_chatbot.tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt"
         ).unsqueeze(0).to(our_chatbot.model.device)
-    # print("### input_id",input_ids)
+    
     stop_str = (
         our_chatbot.conversation.sep
         if our_chatbot.conversation.sep_style != SeparatorStyle.TWO
@@ -384,9 +376,7 @@ def bot(history, temperature, top_p, max_output_tokens):
     stopping_criteria = KeywordsStoppingCriteria(
         keywords, our_chatbot.tokenizer, input_ids
     )
-    # streamer = TextStreamer(
-    #     our_chatbot.tokenizer, skip_prompt=True, skip_special_tokens=True
-    # )
+  
     streamer = TextIteratorStreamer(
         our_chatbot.tokenizer, skip_prompt=True, skip_special_tokens=True
     )
@@ -394,22 +384,6 @@ def bot(history, temperature, top_p, max_output_tokens):
     print(input_ids.device)
     print(image_tensor.device)
     
-    # with torch.inference_mode():
-    #     output_ids = our_chatbot.model.generate(
-    #         input_ids,
-    #         images=image_tensor,
-    #         do_sample=True,
-    #         temperature=0.7,
-    #         top_p=1.0,
-    #         max_new_tokens=4096,
-    #         streamer=streamer,
-    #         use_cache=False,
-    #         stopping_criteria=[stopping_criteria],
-    #     )
-
-    # outputs = our_chatbot.tokenizer.decode(output_ids[0]).strip()
-    # if outputs.endswith(stop_str):
-    #     outputs = outputs[: -len(stop_str)]
     # our_chatbot.conversation.messages[-1][-1] = outputs
 
     # history[-1] = [text, outputs]