Update inference.py

inference.py

@@ -40,9 +40,9 @@ def load_model(model_path):
     vl_chat_processor: VLChatProcessor = VLChatProcessor.from_pretrained(model_path)
     tokenizer = vl_chat_processor.tokenizer
     vl_gpt: MultiModalityCausalLM = AutoModelForCausalLM.from_pretrained(
-        model_path, trust_remote_code=True, load_in_8bit=True,low_cpu_mem_usage=True
+        model_path, trust_remote_code=True, load_in_8bit=True,low_cpu_mem_usage=True,device_map="auto"
     )
-    vl_gpt = vl_gpt.cuda().eval()
+    vl_gpt = vl_gpt.eval()
     return tokenizer, vl_gpt, vl_chat_processor