forward_tensor接口

def forward_tensor(
        self,
        input_ids: torch.Tensor,
        position_ids: torch.Tensor,
        is_prefill: bool,
        kv_cache: List[Tuple[torch.Tensor, torch.Tensor]],
        block_tables: torch.Tensor,
        slots: torch.Tensor,
        input_lengths: torch.Tensor,
        max_seq_len: int,
        lm_head_indices: Optional[torch.Tensor] = None,
        **kwargs
    ):
        """Call the `forward_tensor` of `model_wrapper`."""
        adapter_ids = kwargs.get("adapter_ids")
        batch_size = input_lengths.shape[0]
        if adapter_ids is not None and len(adapter_ids) > batch_size:
            message = "The length of `adapter_ids` should not be larger than batch size."
            logger.error(message, ErrorCode.TEXT_GENERATOR_INTERNAL_ERROR)
            raise ValueError(message)
        logits = self.model_wrapper.forward_tensor(
            input_ids=input_ids,
            position_ids=position_ids,
            is_prefill=is_prefill,
            kv_cache=kv_cache,
            block_tables=block_tables,
            slots=slots,
            input_lengths=input_lengths,
            max_seq_len=max_seq_len,
            lm_head_indices=lm_head_indices,
            **kwargs,
        )
        return logits