[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
+7 -11
View File
@@ -3,7 +3,6 @@
from collections.abc import Iterable, Mapping, Sequence
from typing import Annotated, Any, Literal
import numpy as np
import torch
from torch import nn
from transformers import AutoModel, BatchFeature
@@ -19,6 +18,7 @@ from transformers.models.siglip import SiglipImageProcessorFast
from vllm.config import ModelConfig, SpeechToTextConfig, VllmConfig
from vllm.config.multimodal import BaseDummyOptions
from vllm.config.speech_to_text import SpeechToTextParams
from vllm.inputs import MultiModalDataDict, PromptType, TextPrompt
from vllm.logger import init_logger
from vllm.model_executor.layers.layernorm import RMSNorm
@@ -769,21 +769,17 @@ class Gemma3nForConditionalGeneration(
raise ValueError(f"Unsupported modality: {modality}")
@classmethod
def get_generation_prompt(
cls,
audio: np.ndarray,
stt_config: SpeechToTextConfig,
model_config: ModelConfig,
language: str | None,
task_type: Literal["transcribe", "translate"],
request_prompt: str,
to_language: str | None,
) -> PromptType:
def get_generation_prompt(cls, stt_params: SpeechToTextParams) -> PromptType:
"""
Gemma3n supports "free-form" transcription.
We fix its prompt here to standardize transcriptions/translations
requests.
"""
audio = stt_params.audio
stt_config = stt_params.stt_config
language = stt_params.language
task_type = stt_params.task_type
to_language = stt_params.to_language
# Transcribe this audio [into <>] | for transcription
# Translate this audio [from <> into <>] | for translation
prompt = "<start_of_turn>user\n"