Update app.py

app.py

@@ -10,13 +10,13 @@ model_id = "xtuner/llava-llama-3-8b-v1_1-transformers"
 
 print("Loading model...")
 processor = AutoProcessor.from_pretrained(model_id)
+
+# Adjusted model loading to use Accelerate's `device_map`
 model = LlavaForConditionalGeneration.from_pretrained(
     model_id,
     torch_dtype=torch.float16,
-    low_cpu_mem_usage=True
+    device_map="auto"  # Uses the Accelerate library for efficient memory usage
 )
-model.to("cuda" if torch.cuda.is_available() else "cpu")
-model.generation_config.eos_token_id = 128009
 print("Model loaded successfully!")
 
 PLACEHOLDER = """
@@ -50,7 +50,7 @@ def bot_streaming(message, history):
         # Prepare inputs
         image = Image.open(image)
         prompt = f"<|start_header_id|>user<|end_header_id|>\n\n<image>\n{message['text']}<|eot_id|>"
-        inputs = processor(prompt, image, return_tensors="pt").to(device=model.device, dtype=torch.float16)
+        inputs = processor(prompt, image, return_tensors="pt").to(model.device, dtype=torch.float16)
 
         # Stream text generation
         streamer = TextIteratorStreamer(processor, skip_special_tokens=True, skip_prompt=True)