[BUG/Help] <title>运行示例代码出现TypeError报错

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

运行的就是示例代码如下，出现了下面的TypeError 报错 `from transformers import AutoTokenizer, AutoModel import torch model_path="/home/songxinyue/new/models/chatglm2-6b" tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True) device = torch.device("cuda:2" if torch.cuda.is_available() else "cpu")

model = AutoModel.from_pretrained(model_path, trust_remote_code=True).float()

.half()

Move the model to the device

#model.to(device) model = model.eval() response, history = model.chat(tokenizer, "hello", history=[]) print(response) response, history = model.chat(tokenizer, "What should I do if I can't sleep at night?", history=history) print(response)`

`--------------------------------------------------------------------------- TypeError Traceback (most recent call last) Cell In[1], line 13 9 # .half() 10 # Move the model to the device 11 #model.to(device) 12 model = model.eval() ---> 13 response, history = model.chat(tokenizer, "hello", history=[]) 14 print(response) 15 response, history = model.chat(tokenizer, "What should I do if I can't sleep at night?", history=history)

File ~/new/environment/anaconda3/envs/sxy/lib/python3.10/site-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.call..decorate_context(*args, **kwargs) 24 @functools.wraps(func) 25 def decorate_context(*args, **kwargs): 26 with self.clone(): ---> 27 return func(*args, **kwargs)

File ~/.cache/huggingface/modules/transformers_modules/chatglm2-6b/modeling_chatglm.py:1033, in ChatGLMForConditionalGeneration.chat(self, tokenizer, query, history, max_length, num_beams, do_sample, top_p, temperature, logits_processor, **kwargs) 1030 gen_kwargs = {"max_length": max_length, "num_beams": num_beams, "do_sample": do_sample, "top_p": top_p, 1031 "temperature": temperature, "logits_processor": logits_processor, **kwargs} 1032 inputs = self.build_inputs(tokenizer, query, history=history) -> 1033 outputs = self.generate(**inputs, **gen_kwargs) 1034 outputs = outputs.tolist()[0][len(inputs["input_ids"][0]):] 1035 response = tokenizer.decode(outputs)

File ~/new/environment/anaconda3/envs/sxy/lib/python3.10/site-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.call..decorate_context(*args, **kwargs) 24 @functools.wraps(func) 25 def decorate_context(*args, **kwargs): 26 with self.clone(): ---> 27 return func(*args, **kwargs)

File ~/new/environment/anaconda3/envs/sxy/lib/python3.10/site-packages/transformers/generation/utils.py:2015, in GenerationMixin.generate(self, inputs, generation_config, logits_processor, stopping_criteria, prefix_allowed_tokens_fn, synced_gpus, assistant_model, streamer, negative_prompt_ids, negative_prompt_attention_mask, **kwargs) 2007 input_ids, model_kwargs = self._expand_inputs_for_generation( 2008 input_ids=input_ids, 2009 expand_size=generation_config.num_return_sequences, 2010 is_encoder_decoder=self.config.is_encoder_decoder, 2011 **model_kwargs, 2012 ) 2014 # 12. run sample (it degenerates to greedy search when generation_config.do_sample=False) -> 2015 result = self._sample( 2016 input_ids, 2017 logits_processor=prepared_logits_processor, 2018 stopping_criteria=prepared_stopping_criteria, 2019 generation_config=generation_config, 2020 synced_gpus=synced_gpus, 2021 streamer=streamer, 2022 **model_kwargs, 2023 ) 2025 elif generation_mode in (GenerationMode.BEAM_SAMPLE, GenerationMode.BEAM_SEARCH): 2026 # 11. prepare beam search scorer 2027 beam_scorer = BeamSearchScorer( 2028 batch_size=batch_size, 2029 num_beams=generation_config.num_beams, (...) 2034 max_length=generation_config.max_length, 2035 )

File ~/new/environment/anaconda3/envs/sxy/lib/python3.10/site-packages/transformers/generation/utils.py:3014, in GenerationMixin._sample(self, input_ids, logits_processor, stopping_criteria, generation_config, synced_gpus, streamer, **model_kwargs) 3012 if streamer is not None: 3013 streamer.put(next_tokens.cpu()) -> 3014 model_kwargs = self._update_model_kwargs_for_generation( 3015 outputs, 3016 model_kwargs, 3017 is_encoder_decoder=self.config.is_encoder_decoder, 3018 ) 3020 unfinished_sequences = unfinished_sequences & ~stopping_criteria(input_ids, scores) 3021 this_peer_finished = unfinished_sequences.max() == 0

File ~/.cache/huggingface/modules/transformers_modules/chatglm2-6b/modeling_chatglm.py:871, in ChatGLMForConditionalGeneration._update_model_kwargs_for_generation(self, outputs, model_kwargs, is_encoder_decoder, standardize_cache_format) 863 def _update_model_kwargs_for_generation( 864 self, 865 outputs: ModelOutput, (...) 869 ) -> Dict[str, Any]: 870 # update past_key_values --> 871 model_kwargs["past_key_values"] = self._extract_past_from_model_output( 872 outputs, standardize_cache_format=standardize_cache_format 873 ) 875 # update attention mask 876 if "attention_mask" in model_kwargs:

TypeError: GenerationMixin._extract_past_from_model_output() got an unexpected keyword argument 'standardize_cache_format'`

Expected Behavior

No response

Steps To Reproduce

The code that is running is just the few lines of sample code above, and this TypeError error occurred. I suspect it's a problem with the environment, but I don't know which package specifically needs to be modified.

Environment

- OS:linux
- Python:3.10.14
- Transformers:4.28.1
- PyTorch:1.12.1+cu113
- CUDA Support (`python -c "import torch; print(torch.cuda.is_available())"`) :
True

Anything else?

No response