LlamaTokenizerFast returns token_type_ids but the forward pass of the LlamaModel does not receive token_type_ids

cfierro · June 7, 2023, 3:13pm

huggingface/transformers/blob/092c14c37d9baa6a12be5481138564a720a49a8a/src/transformers/models/llama/modeling_llama.py#L484-L495


      
          def forward(
              self,
              input_ids: torch.LongTensor = None,
              attention_mask: Optional[torch.Tensor] = None,
              position_ids: Optional[torch.LongTensor] = None,
              past_key_values: Optional[List[torch.FloatTensor]] = None,
              inputs_embeds: Optional[torch.FloatTensor] = None,
              use_cache: Optional[bool] = None,
              output_attentions: Optional[bool] = None,
              output_hidden_states: Optional[bool] = None,
              return_dict: Optional[bool] = None,
          ) -> Union[Tuple, BaseModelOutputWithPast]:

Python 3.10.11
transformers==4.28.1

ArthurZ · June 9, 2023, 7:37am

Hey thanks for reporting! This was adressed in [Lllama] Update tokenization code to ensure parsing of the special tokens [core] by ArthurZucker · Pull Request #24042 · huggingface/transformers · GitHub !

Topic		Replies	Views
Are the slow and fast tokenizer results the same output for the same input? 🤗Tokenizers	0	384	August 30, 2023
Getting the same embedding from llama 2 class token for any input 🤗Transformers	1	765	December 4, 2023
Tokenization: different results when tokenizing in one pass vs sample-by-sample Intermediate	3	1157	October 23, 2023
LongFormer tokenizer has the same token_type_ids for sequence pairs 🤗Tokenizers	0	667	December 20, 2021
TypeError: forward() got an unexpected keyword argument 'token_type_ids' Beginners	3	2870	June 10, 2022

LlamaTokenizerFast returns token_type_ids but the forward pass of the LlamaModel does not receive token_type_ids

Related Topics