[BUG] <title>MiniCPM_V_2_6_INT4 版本执行提示 tensor 异常

[Xls1994]

是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this?

• [X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions

该问题是否在FAQ中有解答？ | Is there an existing answer for this in FAQ?

• [X] 我已经搜索过FAQ | I have searched FAQ

当前行为 | Current Behavior

INT4 版本的官方示例代码，使用 transformers 执行，看起来是采样的方法有问题，详细报错如下。

Unused kwargs: ['_load_in_4bit', '_load_in_8bit', 'quant_method']. These kwargs are not used in <class 'transformers.utils.quantization_config.BitsAndBytesConfig'>. low_cpu_mem_usage was None, now set to True since model is quantized. Loading checkpoint shards: 100%|███████████████████████████████████████████████████████████| 2/2 [00:05<00:00, 2.76s/it] Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Traceback (most recent call last): File "minicpm_2_6.py", line 19, in res = model.chat( File "/home/deploy/.cache/huggingface/modules/transformers_modules/MiniCPM-V-2_6-int4/modeling_minicpmv.py", line 378, in chat res = self.generate( File "/home/deploy/.cache/huggingface/modules/transformers_modules/MiniCPM-V-2_6-int4/modeling_minicpmv.py", line 262, in generate result = self._decode(model_inputs["inputs_embeds"], tokenizer, attention_mask, decode_text=decode_text, **kwargs) File "/home/deploy/.cache/huggingface/modules/transformers_modules/MiniCPM-V-2_6-int4/modeling_minicpmv.py", line 186, in _decode output = self.llm.generate( File "/home/deploy/anaconda3/envs/yyl_env_py388/lib/python3.8/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context return func(*args, **kwargs) File "/home/deploy/anaconda3/envs/yyl_env_py388/lib/python3.8/site-packages/transformers/generation/utils.py", line 1622, in generate result = self._sample( File "/home/deploy/anaconda3/envs/yyl_env_py388/lib/python3.8/site-packages/transformers/generation/utils.py", line 2829, in _sample next_tokens = torch.multinomial(probs, num_samples=1).squeeze(1) RuntimeError: probability tensor contains either inf, nan or element < 0

期望行为 | Expected Behavior

可以正常运行出结果，之前 V2.5 版本是可以正常运行的。 环境都是同一样的，按道理不应该出现这个问题。

复现方法 | Steps To Reproduce

No response

运行环境 | Environment

- OS: centos
- Python: 3.8.9
- Transformers: 4.40.0
- PyTorch:2.1.0
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):12.1

备注 | Anything else?

No response