[Deprecation] Deprecate `seed_everything` and `scatter_mm_placeholders` in v0.15 (#33362)

Signed-off-by: yewentao256 <zhyanwentao@126.com>

vllm/model_executor/models/voxtral.py

@@ -103,14 +103,6 @@ class VoxtralProcessorAdapter:
     def begin_audio_token_id(self) -> int:
         return self._audio_processor.special_ids.begin_audio
 
-    # @cached_property
-    # def begin_transcript_token_id(self) -> int:
-    #     return self._audio_processor.special_ids.begin_transcript
-
-    # @cached_property
-    # def end_transcript_token_id(self) -> int:
-    #     return self._audio_processor.special_ids.end_transcript
-
     @cached_property
     def sampling_rate(self) -> int:
         return self._audio_processor.audio_config.sampling_rate

vllm/platforms/interface.py

@@ -4,14 +4,11 @@ import contextlib
 import enum
 import os
 import platform
-import random
 import sys
 from datetime import timedelta
 from typing import TYPE_CHECKING, Any, NamedTuple
 
-import numpy as np
 import torch
-from typing_extensions import deprecated
 
 from vllm.logger import init_logger
 from vllm.v1.attention.backends.registry import AttentionBackendEnum
@@ -365,23 +362,6 @@ class Platform:
         """
         return torch.inference_mode(mode=True)
 
-    @classmethod
-    @deprecated(
-        "`seed_everything` is deprecated. It will be removed in v0.15.0 or later. "
-        "Please use `vllm.utils.torch_utils.set_random_seed` instead."
-    )
-    def seed_everything(cls, seed: int | None = None) -> None:
-        """
-        Set the seed of each random module.
-        `torch.manual_seed` will set seed on all devices.
-
-        Loosely based on: https://github.com/Lightning-AI/pytorch-lightning/blob/2.4.0/src/lightning/fabric/utilities/seed.py#L20
-        """
-        if seed is not None:
-            random.seed(seed)
-            np.random.seed(seed)
-            torch.manual_seed(seed)
-
     @classmethod
     def set_device(cls, device: torch.device) -> None:
         """

vllm/v1/worker/utils.py

@@ -5,7 +5,6 @@ from collections import defaultdict
 from dataclasses import dataclass, field
 
 import torch
-from typing_extensions import deprecated
 
 from vllm.config import CacheConfig, VllmConfig
 from vllm.logger import init_logger
@@ -201,52 +200,6 @@ def sanity_check_mm_encoder_outputs(
     )
 
 
-@deprecated("`scatter_mm_placeholders` is deprecated and will be removed in v0.15.0.")
-def scatter_mm_placeholders(
-    embeds: torch.Tensor,
-    is_embed: torch.Tensor | None,
-) -> torch.Tensor:
-    """
-    Scatter the multimodal embeddings into a contiguous tensor that represents
-    the placeholder tokens.
-
-    [`vllm.multimodal.processing.PromptUpdateDetails.is_embed`][].
-
-    Args:
-        embeds: The multimodal embeddings.
-            Shape: `(num_embeds, embed_dim)`
-        is_embed: A boolean mask indicating which positions in the placeholder
-            tokens need to be filled with multimodal embeddings.
-            Shape: `(num_placeholders, num_embeds)`
-    """
-    if is_embed is None:
-        return embeds
-
-    placeholders = embeds.new_full(
-        (is_embed.shape[0], embeds.shape[-1]),
-        fill_value=torch.nan,
-    )
-    placeholders[is_embed] = embeds
-    return placeholders
-
-
-@deprecated("`gather_mm_placeholders` is deprecated and will be removed in v0.15.0.")
-def gather_mm_placeholders(
-    placeholders: torch.Tensor,
-    is_embed: torch.Tensor | None,
-) -> torch.Tensor:
-    """
-    Reconstructs the embeddings from the placeholder tokens.
-
-    This is the operation of [`scatter_mm_placeholders`]
-    [vllm.v1.worker.utils.scatter_mm_placeholders].
-    """
-    if is_embed is None:
-        return placeholders
-
-    return placeholders[is_embed]
-
-
 def request_memory(init_snapshot: MemorySnapshot, cache_config: CacheConfig) -> int:
     """
     Calculate the amount of memory required by vLLM, then validate