Update app.py

app.py

@@ -29,17 +29,15 @@ model.generation_config.eos_token_id = 128009
 
 @spaces.GPU
 def bot_streaming(message, history):
-  print(message)
-  if message["files"]:
-    image = message["files"][-1]["path"]
-  else:
-    # if there's no image uploaded for this turn, look for images in the past turns
-    # kept inside tuples, take the last one
-    for hist in history:
-      if type(hist[0])==tuple:
-        image = hist[0][0]
-
-
+    print(message)
+    if message["files"]:
+        image = message["files"][-1]["path"]
+    else:
+        # if there's no image uploaded for this turn, look for images in the past turns
+        # kept inside tuples, take the last one
+        for hist in history:
+          if type(hist[0])==tuple:
+            image = hist[0][0]
     try:
         if image is None:
             # Handle the case where image is None
@@ -56,20 +54,20 @@ def bot_streaming(message, history):
     streamer = TextIteratorStreamer(processor, **{"skip_special_tokens": True})
     generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=1024)
     generated_text = ""
-    
+
     thread = Thread(target=model.generate, kwargs=generation_kwargs)
     thread.start()
     
     text_prompt =f"<|start_header_id|>user<|end_header_id|>\n\n{message['text']}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
     print(f"text_prompt: {text_prompt}")
-    
+
     buffer = ""
     for new_text in streamer:
     
         buffer += new_text
         
         generated_text_without_prompt = buffer[len(text_prompt):]
-        time.sleep(0.04)
+        time.sleep(0.08)
         yield generated_text_without_prompt
 
 
@@ -78,5 +76,5 @@ demo = gr.ChatInterface(fn=bot_streaming, css=CSS, fill_height=True, title="LLaV
                         description="Try [LLaVA Llama-3-8B](https://huggingface.co/xtuner/llava-llama-3-8b-v1_1-transformers). Upload an image and start chatting about it, or simply try one of the examples below. If you don't upload an image, you will receive an error.",
                         stop_btn="Stop Generation", multimodal=True)
 
-demo.queue(default_concurrency_limit=20, max_size=20, api_open=False)
+demo.queue(api_open=False)
 demo.launch(show_api=False, share=False)