Update app.py

app.py

@@ -186,7 +186,7 @@ def load_model():
     #     trust_remote_code=True).to(torch.bfloat16).cuda())
     model = AutoModelForCausalLM.from_pretrained(
         model_name_or_path,
-        trust_remote_code=True).to("cpu")
+        trust_remote_code=True).to(torch.float32)
     tokenizer = AutoTokenizer.from_pretrained(model_name_or_path,
                                               trust_remote_code=True)
     return model, tokenizer