Update app.py

app.py

@@ -324,7 +324,16 @@ title_markdown = """
 # 🌋 LLaVA: Large Language and Vision Assistant
 [[Project Page]](https://llava-vl.github.io) [[Paper]](https://arxiv.org/abs/2304.08485) [[Code]](https://github.com/haotian-liu/LLaVA) [[Model]](https://github.com/haotian-liu/LLaVA/blob/main/docs/MODEL_ZOO.md)
 
-ONLY WORKS WITH GPU! 8-bit is used by default to load 13B model in A10G-Large (24G). Set environment `bits` to 4 to load it in T4-medium (15G).
+ONLY WORKS WITH GPU!
+
+You can load the model with 8-bit or 4-bit quantization to make it fit in smaller hardwares. Setting the environment variable `bits` to control the quantization.
+
+Recommended configurations:
+| Hardware           | Bits           |
+|--------------------|----------------|
+| A10G-Large (24G)   | 8 (default)    |
+| T4-Medium (15G)    | 4              |
+| A100-Large (40G)   | 16             |
 """
 
 tos_markdown = """
@@ -593,13 +602,13 @@ if __name__ == "__main__":
     model_path = "liuhaotian/llava-v1.5-13b"
     bits = int(os.getenv("bits", 8))
 
-    preload_models(model_path)
+    # preload_models(model_path)
 
     controller_proc = start_controller()
     worker_proc = start_worker(model_path, bits=bits)
 
     # Wait for worker and controller to start
-    time.sleep(10)
+    # time.sleep(10)
 
     try:
         start_demo(args)