update chat

app.py

@@ -234,7 +234,7 @@ def upload_img(gr_img, text_input, chat_state,chatbot):
     chatbot = chatbot + [[(path,), None]]
     llm_message = chat.upload_img(gr_img, chat_state, img_list)
     return gr.update(interactive=False), gr.update(interactive=True, placeholder='Type and press Enter'), gr.update(
-        value="Start Chatting", interactive=False), chat_state, img_list
+        value="Start Chatting", interactive=False), chat_state, img_list,chatbot
 
 
 def gradio_ask(user_message, chatbot, chat_state):
@@ -293,7 +293,7 @@ with gr.Blocks() as demo:
             text_input = gr.Textbox(label='User', placeholder='Please upload your image first', interactive=False)
 
     upload_button.click(upload_img, [image, text_input, chat_state,chatbot],
-                        [image, text_input, upload_button, chat_state, img_list])
+                        [image, text_input, upload_button, chat_state, img_list,chatbot])
 
     text_input.submit(gradio_ask, [text_input, chatbot, chat_state], [text_input, chatbot, chat_state]).then(
         gradio_answer, [chatbot, chat_state, img_list, num_beams, temperature], [chatbot, chat_state, img_list]

multimodal/open_flamingo/chat/conversation.py

@@ -318,7 +318,7 @@ class Chat:
         previsual_token_id = self.tokenizer("<|#previsual#|>", add_special_tokens=False)["input_ids"][-1]
         prebox_token_id = self.tokenizer("<|#prebox#|>", add_special_tokens=False)["input_ids"][-1]
         size = 224
-        model.eval()
+        self.model.eval()
         # "/gpfs/u/home/LMCG/LMCGljnn/scratch-shared/cdl/tmp_img/chat_vis/chat19.png"
         # image_path = input("Please enter the image path: ")
         image = img_list[0].convert("RGB")
@@ -358,8 +358,9 @@ class Chat:
         image_start_index_list = ((input_ids == media_token_id).nonzero(as_tuple=True)[-1] + 1).tolist()
         image_start_index_list = [[x] for x in image_start_index_list]
         image_nums = [1] * len(input_ids)
-        with torch.no_grad() and torch.cuda.amp.autocast(dtype=torch.float16):
-            outputs = model.generate(
+        # and torch.cuda.amp.autocast(dtype=torch.float16)
+        with torch.no_grad():
+            outputs = self.model.generate(
                 batch_images,
                 input_ids,
                 attention_mask=attention_mask,