Update app.py

app.py

@@ -4,7 +4,7 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
 
 tokenizer = AutoTokenizer.from_pretrained("BirdL/DeepSeek-Coder-V2-Lite-Instruct-FlashAttnPatch", trust_remote_code=True)
-model = AutoModelForCausalLM.from_pretrained("BirdL/DeepSeek-Coder-V2-Lite-Instruct-FlashAttnPatch", trust_remote_code=True, torch_dtype=torch.bfloat16).cuda()    
+model = AutoModelForCausalLM.from_pretrained("BirdL/DeepSeek-Coder-V2-Lite-Instruct-FlashAttnPatch", trust_remote_code=True, device_map="auto", load_in_8bit=True).cuda()    
 
 @spaces.GPU
 def respond(message, history):