vllm-project--vllm/vllm/v1/pool/metadata.py

# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from dataclasses import dataclass

import numpy as np
import torch

from vllm.pooling_params import PoolingParams
from vllm.tasks import PoolingTask
from vllm.utils.torch_utils import PIN_MEMORY


@dataclass
class PoolingCursor:
    first_token_indices_gpu: torch.Tensor
    last_token_indices_gpu: torch.Tensor
    prompt_lens_cpu: torch.Tensor
    seq_lens_cpu: torch.Tensor
    num_scheduled_tokens_cpu: torch.Tensor

    def __getitem__(self, indices: slice) -> "PoolingCursor":
        return PoolingCursor(
            first_token_indices_gpu=self.first_token_indices_gpu[indices],
            last_token_indices_gpu=self.last_token_indices_gpu[indices],
            prompt_lens_cpu=self.prompt_lens_cpu[indices],
            seq_lens_cpu=self.seq_lens_cpu[indices],
            num_scheduled_tokens_cpu=self.num_scheduled_tokens_cpu[indices],
        )

    def is_partial_prefill(self) -> bool:
        return not torch.all(self.prompt_lens_cpu == self.num_scheduled_tokens_cpu)

    def is_finished(self) -> torch.Tensor:
        return self.prompt_lens_cpu == self.seq_lens_cpu


class PoolingStates:
    def __init__(self) -> None:
        # for chunked prefill with ALL pooling
        self.hidden_states_cache: list[torch.Tensor] = []

    def clean(self) -> None:
        self.hidden_states_cache.clear()


@dataclass
class PoolingMetadata:
    """Tensors for pooling."""

    prompt_lens: torch.Tensor  # CPU Tensor
    prompt_token_ids: torch.Tensor | None  # Model-device tensor
    prompt_token_ids_cpu: torch.Tensor | None  # CPU tensor
    pooling_params: list[PoolingParams]
    pooling_states: list[PoolingStates]
    pooling_cursor: PoolingCursor | None = None

    def __post_init__(self) -> None:
        pooling_params = self.pooling_params

        tasks: list[PoolingTask] = [
            task
            for pooling_param in pooling_params
            if (task := pooling_param.task) is not None
        ]
        if len(pooling_params) != len(tasks):
            raise ValueError(
                "Every pooling param must have a task set, but got "
                f"{len(tasks)} tasks for {len(pooling_params)} pooling params"
            )

        self.tasks = tasks

    def __getitem__(self, indices: slice) -> "PoolingMetadata":
        return PoolingMetadata(
            prompt_lens=self.prompt_lens[indices],
            prompt_token_ids=None
            if self.prompt_token_ids is None
            else self.prompt_token_ids[indices],
            prompt_token_ids_cpu=None
            if self.prompt_token_ids_cpu is None
            else self.prompt_token_ids_cpu[indices],
            pooling_params=self.pooling_params[indices],
            pooling_states=self.pooling_states[indices],
            pooling_cursor=None
            if self.pooling_cursor is None
            else self.pooling_cursor[indices],
        )

    def _get_prompt_token_ids(
        self,
        prompt_token_ids: torch.Tensor | None,
    ) -> list[torch.Tensor]:
        if prompt_token_ids is None:
            raise ValueError(
                "prompt_token_ids is required but was not set. "
                "Please set `requires_token_ids=True` in `get_pooling_updates`"
            )
        return [prompt_token_ids[i, :num] for i, num in enumerate(self.prompt_lens)]

    def get_prompt_token_ids(self) -> list[torch.Tensor]:
        return self._get_prompt_token_ids(self.prompt_token_ids)

    def get_prompt_token_ids_cpu(self) -> list[torch.Tensor]:
        return self._get_prompt_token_ids(self.prompt_token_ids_cpu)

    def get_pooling_cursor(self) -> PoolingCursor:
        pooling_cursor = self.pooling_cursor
        if pooling_cursor is None:
            raise RuntimeError(
                "pooling_cursor has not been initialized. "
                "Call `build_pooling_cursor` before accessing it"
            )

        return pooling_cursor

    def build_pooling_cursor(
        self,
        num_scheduled_tokens_np: np.ndarray,
        seq_lens_cpu: torch.Tensor,
        device: torch.device,
        query_start_loc_gpu: torch.Tensor | None = None,
    ) -> None:
        n_seq = len(num_scheduled_tokens_np)
        prompt_lens = self.prompt_lens

        if len(prompt_lens) != n_seq:
            raise ValueError(
                f"prompt_lens length ({len(prompt_lens)}) does not match "
                f"the number of sequences ({n_seq})"
            )

        num_scheduled_tokens_cpu = torch.from_numpy(num_scheduled_tokens_np)
        if query_start_loc_gpu is None:
            cumsum = torch.zeros(
                n_seq + 1, dtype=torch.int64, pin_memory=PIN_MEMORY, device="cpu"
            )
            torch.cumsum(num_scheduled_tokens_cpu, dim=0, out=cumsum[1:])
            cumsum = cumsum.to(device, non_blocking=True)
        else:
            if query_start_loc_gpu.shape[0] != n_seq + 1:
                raise ValueError(
                    "query_start_loc_gpu length does not match "
                    f"the number of sequences: {query_start_loc_gpu.shape[0]} "
                    f"!= {n_seq + 1}."
                )
            if query_start_loc_gpu.device != device:
                raise ValueError(
                    "query_start_loc_gpu must be on the same device as the "
                    f"hidden states: {query_start_loc_gpu.device} != {device}."
                )
            cumsum = query_start_loc_gpu
        self.pooling_cursor = PoolingCursor(
            first_token_indices_gpu=cumsum[:n_seq],
            last_token_indices_gpu=cumsum[1:] - 1,
            prompt_lens_cpu=prompt_lens,
            seq_lens_cpu=seq_lens_cpu,
            num_scheduled_tokens_cpu=num_scheduled_tokens_cpu,
        )