The results are different each time

[saltedfisssh]

Even when I fix the random seed during reconstruction, the resulting point cloud files and evaluation metrics still vary. Could you help me understand what might be causing this discrepancy? Below is the code used to fix the seed:

def setup_seed(seed):
     torch.manual_seed(seed)
     torch.cuda.manual_seed_all(seed)
     np.random.seed(seed)
     random.seed(seed)
     torch.backends.cudnn.deterministic = True