[Audio] Bundle get_generation_prompt() params into SpeechToTextParams (#36268)

Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
This commit is contained in:
Ekagra Ranjan
2026-04-22 12:24:18 +08:00
committed by GitHub
co-authored by gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Cyrus Leung
parent 6d09769700
commit 4eafc72928
18 changed files with 214 additions and 161 deletions
+9 -8
View File
@@ -5,7 +5,7 @@ import enum
import math
from collections.abc import Iterable, Mapping, Sequence
from contextlib import nullcontext
from typing import Annotated, Literal
from typing import Annotated
import numpy as np
import torch
@@ -20,6 +20,7 @@ from transformers.models.whisper.modeling_whisper import sinusoids
from vllm.compilation.decorators import support_torch_compile
from vllm.config import CacheConfig, ModelConfig, SpeechToTextConfig, VllmConfig
from vllm.config.multimodal import BaseDummyOptions
from vllm.config.speech_to_text import SpeechToTextParams
from vllm.distributed import get_tensor_model_parallel_world_size
from vllm.inputs import (
ExplicitEncoderDecoderPrompt,
@@ -830,14 +831,14 @@ class WhisperForConditionalGeneration(
@classmethod
def get_generation_prompt(
cls,
audio: np.ndarray,
model_config: ModelConfig, # not needed here
stt_config: SpeechToTextConfig,
language: str | None,
task_type: Literal["transcribe", "translate"],
request_prompt: str,
to_language: str | None,
stt_params: SpeechToTextParams,
) -> PromptType:
audio = stt_params.audio
stt_config = stt_params.stt_config
language = stt_params.language
task_type = stt_params.task_type
request_prompt = stt_params.request_prompt
if language is None:
raise ValueError(
"Language must be specified when creating the Whisper prompt"