probability tensor contains either `inf`, `nan` or element < 0
在DCU直通VM中进行测试，报错如下：
Traceback (most recent call last):
File "/data/llama/llama_inference_pytorch/llama_infer.py", line 68, in <module>
result = lm_generation.generate(args, prompts)
File "/data/llama/llama_inference_pytorch/generate.py", line 110, in generate
next_token = torch.multinomial(scores, num_samples=1).squeeze(1)
RuntimeError: probability tensor contains either `inf`, `nan` or element < 0