ValueError: not enough values to unpack (expected 2, got 1)

[Roopesh-Bharatwaj-K-R]

Hi,

I'm trying to run the Cross-Encoder given example.
i faced this error in Line no 41.

   torch.nn.functional.log_softmax(model(model_input)[0], dim=-1).cpu()

i tried modifying dimentions but no luck, tried to give input as 2 seperate parameters but also did not work.

Please kindly help me which parameter should i update to avoid unpack error.

Thank You and Much Appreciated

 ValueError                                Traceback (most recent call last)
 Cell In [54], line 41
 37 print(input_len)
 38 # print('model',model(model_input))
 39 
 40 # [seq_len] -> [seq_len, vocab]
  ---> 41 logprobs = torch.nn.functional.log_softmax(model(model_input)[0], dim=-1).cpu()
 42 # [seq_len, vocab] -> [continuation_len, vocab]
 43 logprobs = logprobs[input_len-continuation_len:]

 File /anaconda/envs/azureml_py38/lib/python3.8/site-packages/torch/nn/modules/module.py:1130, in Module._call_impl(self, *input, **kwargs)
   1126 # If we don't have any hooks, we want to skip the rest of the logic in
    1127 # this function, and just call forward.
    1128 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
    1129         or _global_forward_hooks or _global_forward_pre_hooks):
    -> 1130     return forward_call(*input, **kwargs)
    1131 # Do not call functions when jit is used
    1132 full_backward_hooks, non_full_backward_hooks = [], []

    File /anaconda/envs/azureml_py38/lib/python3.8/site-packages/transformers/models/gpt_neo/modeling_gpt_neo.py:974, in 
    GPTNeoForCausalLM.forward(self, input_ids, past_key_values, attention_mask, token_type_ids, position_ids, head_mask, 
    inputs_embeds, labels, use_cache, output_attentions, output_hidden_states, return_dict)
    966 r"""
    967 labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`):
    968     Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
    969     ``labels = input_ids`` Indices are selected in ``[-100, 0, ..., config.vocab_size]`` All labels set to
    970     ``-100`` are ignored (masked), the loss is only computed for labels in ``[0, ..., config.vocab_size]``
    971 """
    972 return_dict = return_dict if return_dict is not None else self.config.use_return_dict
    --> 974 transformer_outputs = self.transformer(
    975     input_ids,
    976     past_key_values=past_key_values,
    977     attention_mask=attention_mask,
    978     token_type_ids=token_type_ids,
    979     position_ids=position_ids,
    980     head_mask=head_mask,
    981     inputs_embeds=inputs_embeds,
    982     use_cache=use_cache,
    983     output_attentions=output_attentions,
    984     output_hidden_states=output_hidden_states,
    985     return_dict=return_dict,
    986 )
    987 hidden_states = transformer_outputs[0]
    989 lm_logits = self.lm_head(hidden_states)

   File /anaconda/envs/azureml_py38/lib/python3.8/site-packages/torch/nn/modules/module.py:1130, in Module._call_impl(self, 
  *input, **kwargs)
   1126 # If we don't have any hooks, we want to skip the rest of the logic in
    1127 # this function, and just call forward.
    1128 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
     1129         or _global_forward_hooks or _global_forward_pre_hooks):
    -> 1130     return forward_call(*input, **kwargs)
    1131 # Do not call functions when jit is used
    1132 full_backward_hooks, non_full_backward_hooks = [], []

     File /anaconda/envs/azureml_py38/lib/python3.8/site-packages/transformers/models/gpt_neo/modeling_gpt_neo.py:799, in 
     GPTNeoModel.forward(self, input_ids, past_key_values, attention_mask, token_type_ids, position_ids, head_mask, 
     inputs_embeds, use_cache, output_attentions, output_hidden_states, return_dict)
     796     global_attention_mask = None
     798 # Local causal attention mask
     --> 799 batch_size, seq_length = input_shape
     800 full_seq_length = seq_length + past_length
     801 local_attention_mask = GPTNeoAttentionMixin.create_local_attention_mask(
     802     batch_size, full_seq_length, self.config.window_size, device, attention_mask
     803 )

     ValueError: not enough values to unpack (expected 2, got 1)

[Muennighoff]

It works fine on my side, see this notebook: https://colab.research.google.com/drive/1mxH15422ZnguaItPBKR2PZ_qACM5QE0l?usp=sharing

Maybe you're on an older transformers version?