Allow cuda tf32 matmul to optimize

app.py

@@ -35,6 +35,7 @@ MODEL = os.getenv(
 
 torch.backends.cudnn.deterministic = True
 torch.backends.cudnn.benchmark = False
+torch.backends.cuda.matmul.allow_tf32 = True
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")