Update app.py

app.py

@@ -11,7 +11,7 @@ torch_dtype = torch.bfloat16 if torch.cuda.is_available() else torch.float32
 os.environ['HF_HOME'] = '/tmp/cache'
 
 # Load model and tokenizer (using cerebras BTLM-3B-8K)
-model_name = "cerebras/btlm-3b-8k"
+model_name = "cerebras/btlm-3b-8k-chat"
 
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(