从transformers 自动下载模型，response, history = model.chat(tokenizer, "你好", history=[])报错too many values to unpack (expected 2)

[LiuXingEmail]

kaggle上使用官方提供的代码从 transformers下载模型，调用 ChatGLM 生成对话，在response, history = model.chat(tokenizer, "你好", history=[])对话阶段出现以下错误

ValueError Traceback (most recent call last)
Cell In[6], line 5
3 model = AutoModel.from_pretrained("THUDM/chatglm3-6b", trust_remote_code=True, device='cuda')
4 model = model.eval()
----> 5 response, history = model.chat(tokenizer, "你好", history=[])
6 print(response)

File /opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py:115, in context_decorator..decorate_context(*args, **kwargs)
112 @functools.wraps(func)
113 def decorate_context(*args, **kwargs):
114 with ctx_factory():
--> 115 return func(*args, **kwargs)

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:1042, in ChatGLMForConditionalGeneration.chat(self, tokenizer, query, history, role, max_length, num_beams, do_sample, top_p, temperature, logits_processor, **kwargs)
1039 inputs = inputs.to(self.device)
1040 eos_token_id = [tokenizer.eos_token_id, tokenizer.get_command("<|user|>"),
1041 tokenizer.get_command("<|observation|>")]
-> 1042 outputs = self.generate(**inputs, **gen_kwargs, eos_token_id=eos_token_id)
1043 outputs = outputs.tolist()[0][len(inputs["input_ids"][0]):-1]
1044 response = tokenizer.decode(outputs)

File /opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py:115, in context_decorator..decorate_context(*args, **kwargs)
112 @functools.wraps(func)
113 def decorate_context(*args, **kwargs):
114 with ctx_factory():
--> 115 return func(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1914, 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)
1906 input_ids, model_kwargs = self._expand_inputs_for_generation(
1907 input_ids=input_ids,
1908 expand_size=generation_config.num_return_sequences,
1909 is_encoder_decoder=self.config.is_encoder_decoder,
1910 **model_kwargs,
1911 )
1913 # 13. run sample (it degenerates to greedy search when generation_config.do_sample=False)
-> 1914 result = self._sample(
1915 input_ids,
1916 logits_processor=prepared_logits_processor,
1917 logits_warper=prepared_logits_warper,
1918 stopping_criteria=prepared_stopping_criteria,
1919 generation_config=generation_config,
1920 synced_gpus=synced_gpus,
1921 streamer=streamer,
1922 **model_kwargs,
1923 )
1925 elif generation_mode in (GenerationMode.BEAM_SAMPLE, GenerationMode.BEAM_SEARCH):
1926 # 11. prepare logits warper
1927 prepared_logits_warper = (
1928 self._get_logits_warper(generation_config, device=input_ids.device)
1929 if generation_config.do_sample
1930 else None
1931 )

File /opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:2651, in GenerationMixin._sample(self, input_ids, logits_processor, stopping_criteria, generation_config, synced_gpus, streamer, logits_warper, **model_kwargs)
2648 model_inputs = self.prepare_inputs_for_generation(input_ids, **model_kwargs)
2650 # forward pass to get next token
-> 2651 outputs = self(
2652 **model_inputs,
2653 return_dict=True,
2654 output_attentions=output_attentions,
2655 output_hidden_states=output_hidden_states,
2656 )
2658 if synced_gpus and this_peer_finished:
2659 continue # don't waste resources running the code we don't need

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1518, in Module._wrapped_call_impl(self, *args, **kwargs)
1516 return self._compiled_call_impl(*args, **kwargs) # type: ignore[misc]
1517 else:
-> 1518 return self._call_impl(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1527, in Module._call_impl(self, *args, **kwargs)
1522 # If we don't have any hooks, we want to skip the rest of the logic in
1523 # this function, and just call forward.
1524 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
1525 or _global_backward_pre_hooks or _global_backward_hooks
1526 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1527 return forward_call(*args, **kwargs)
1529 try:
1530 result = None

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:941, in ChatGLMForConditionalGeneration.forward(self, input_ids, position_ids, attention_mask, past_key_values, inputs_embeds, labels, use_cache, output_attentions, output_hidden_states, return_dict, return_last_logit)
938 use_cache = use_cache if use_cache is not None else self.config.use_cache
939 return_dict = return_dict if return_dict is not None else self.config.use_return_dict
--> 941 transformer_outputs = self.transformer(
942 input_ids=input_ids,
943 position_ids=position_ids,
944 attention_mask=attention_mask,
945 past_key_values=past_key_values,
946 inputs_embeds=inputs_embeds,
947 use_cache=use_cache,
948 output_hidden_states=output_hidden_states,
949 return_dict=return_dict,
950 )
952 hidden_states = transformer_outputs[0]
953 if return_last_logit:

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1518, in Module._wrapped_call_impl(self, *args, **kwargs)
1516 return self._compiled_call_impl(*args, **kwargs) # type: ignore[misc]
1517 else:
-> 1518 return self._call_impl(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1527, in Module._call_impl(self, *args, **kwargs)
1522 # If we don't have any hooks, we want to skip the rest of the logic in
1523 # this function, and just call forward.
1524 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
1525 or _global_backward_pre_hooks or _global_backward_hooks
1526 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1527 return forward_call(*args, **kwargs)
1529 try:
1530 result = None

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:834, in ChatGLMModel.forward(self, input_ids, position_ids, attention_mask, full_attention_mask, past_key_values, inputs_embeds, use_cache, output_hidden_states, return_dict)
831 rotary_pos_emb = rotary_pos_emb.transpose(0, 1).contiguous()
833 # Run encoder.
--> 834 hidden_states, presents, all_hidden_states, all_self_attentions = self.encoder(
835 inputs_embeds, full_attention_mask, rotary_pos_emb=rotary_pos_emb,
836 kv_caches=past_key_values, use_cache=use_cache, output_hidden_states=output_hidden_states
837 )
839 if not return_dict:
840 return tuple(v for v in [hidden_states, presents, all_hidden_states, all_self_attentions] if v is not None)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1518, in Module._wrapped_call_impl(self, *args, **kwargs)
1516 return self._compiled_call_impl(*args, **kwargs) # type: ignore[misc]
1517 else:
-> 1518 return self._call_impl(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1527, in Module._call_impl(self, *args, **kwargs)
1522 # If we don't have any hooks, we want to skip the rest of the logic in
1523 # this function, and just call forward.
1524 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
1525 or _global_backward_pre_hooks or _global_backward_hooks
1526 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1527 return forward_call(*args, **kwargs)
1529 try:
1530 result = None

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:641, in GLMTransformer.forward(self, hidden_states, attention_mask, rotary_pos_emb, kv_caches, use_cache, output_hidden_states)
631 layer_ret = torch.utils.checkpoint.checkpoint(
632 layer,
633 hidden_states,
(...)
638 use_reentrant=False
639 )
640 else:
--> 641 layer_ret = layer(
642 hidden_states,
643 attention_mask,
644 rotary_pos_emb,
645 kv_cache=kv_caches[index],
646 use_cache=use_cache
647 )
648 hidden_states, kv_cache = layer_ret
649 if use_cache:

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1518, in Module._wrapped_call_impl(self, *args, **kwargs)
1516 return self._compiled_call_impl(*args, **kwargs) # type: ignore[misc]
1517 else:
-> 1518 return self._call_impl(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1527, in Module._call_impl(self, *args, **kwargs)
1522 # If we don't have any hooks, we want to skip the rest of the logic in
1523 # this function, and just call forward.
1524 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
1525 or _global_backward_pre_hooks or _global_backward_hooks
1526 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1527 return forward_call(*args, **kwargs)
1529 try:
1530 result = None

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:544, in GLMBlock.forward(self, hidden_states, attention_mask, rotary_pos_emb, kv_cache, use_cache)
542 layernorm_output = self.input_layernorm(hidden_states)
543 # Self attention.
--> 544 attention_output, kv_cache = self.self_attention(
545 layernorm_output,
546 attention_mask,
547 rotary_pos_emb,
548 kv_cache=kv_cache,
549 use_cache=use_cache
550 )
552 # Residual connection.
553 if self.apply_residual_connection_post_layernorm:

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1518, in Module._wrapped_call_impl(self, *args, **kwargs)
1516 return self._compiled_call_impl(*args, **kwargs) # type: ignore[misc]
1517 else:
-> 1518 return self._call_impl(*args, **kwargs)

File /opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py:1527, in Module._call_impl(self, *args, **kwargs)
1522 # If we don't have any hooks, we want to skip the rest of the logic in
1523 # this function, and just call forward.
1524 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
1525 or _global_backward_pre_hooks or _global_backward_hooks
1526 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1527 return forward_call(*args, **kwargs)
1529 try:
1530 result = None

File ~/.cache/huggingface/modules/transformers_modules/THUDM/chatglm3-6b/06c7c873c843814171c51330b69c2e2a68e05178/modeling_chatglm.py:413, in SelfAttention.forward(self, hidden_states, attention_mask, rotary_pos_emb, kv_cache, use_cache)
411 # adjust key and value for inference
412 if kv_cache is not None:
--> 413 cache_k, cache_v = kv_cache
414 key_layer = torch.cat((cache_k, key_layer), dim=0)
415 value_layer = torch.cat((cache_v, value_layer), dim=0)

ValueError: too many values to unpack (expected 2)

[ViktorJiangC]

pip install transformers==4.41.2，试一下requirements.txt中最低的transformers版本