[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 -13
View File
@@ -3,9 +3,7 @@
import math
from collections.abc import Iterable, Mapping, Sequence
from typing import Literal
import numpy as np
import torch
import torch.nn.functional as F
from torch import nn
@@ -13,6 +11,8 @@ from transformers import PretrainedConfig
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 MultiModalDataDict, PromptType, TextPrompt
from vllm.logger import init_logger
@@ -1900,7 +1900,7 @@ class CohereASRDummyInputsBuilder(BaseDummyInputsBuilder[CohereASRProcessingInfo
self,
seq_len: int,
mm_counts: Mapping[str, int],
mm_options=None,
mm_options: Mapping[str, BaseDummyOptions],
mm_processor_kwargs=None,
) -> MultiModalDataDict:
feature_extractor = self.info.get_feature_extractor()
@@ -2021,16 +2021,12 @@ class CohereAsrForConditionalGeneration(
return super().validate_language(language)
@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,
) -> PromptType:
def get_generation_prompt(cls, stt_params: SpeechToTextParams) -> PromptType:
audio = stt_params.audio
stt_config = stt_params.stt_config
language = stt_params.language
request_prompt = stt_params.request_prompt
if language is None:
raise ValueError(
"Language must be specified when creating the CohereASR prompt"