OpenAI: Whisper Large v3 Turbo
OpenAI speed-optimized Whisper Large v3 derivative: a pruned and finetuned large-v3 model with decoder layers reduced from 32 to 4 for faster inference at minor quality degradation.
プロバイダー
同じモデルを複数のプロバイダーがホストしている場合があります。バックエンドを固定したい場合は1つのプロバイダーを選び、そうでなければルーティングに選択を任せられます。
対応言語
en
zh
de
es
ru
ko
fr
ja
pt
tr
pl
ca
nl
ar
sv
it
id
hi
fi
vi
he
uk
el
ms
cs
ro
da
hu
ta
no
th
ur
hr
bg
lt
la
mi
ml
cy
sk
te
fa
lv
bn
sr
az
sl
kn
et
mk
br
eu
is
hy
ne
mn
bs
kk
sq
sw
gl
mr
pa
si
km
sn
yo
so
af
oc
ka
be
tg
sd
gu
am
yi
lo
uz
fo
ht
ps
tk
nn
mt
sa
lb
my
bo
tl
mg
as
tt
haw
ln
ha
ba
jw
su
yue
機能
Transcribes spoken audio into text.
speech_to_text
Transcribes a fully-submitted audio file in a single synchronous request.
synchronous
Supports transcription across the Whisper multilingual language set.
multilingual_transcription
Can translate supported source-language speech into English text.
translation_to_english
Can infer the source language automatically when the runtime/provider supports the Whisper language detection path.
automatic_language_detection
Supports timestamped transcription output at segment level, with word-level timestamps available in common Whisper inference pipelines.
timestamps
Uses a reduced decoder depth for faster inference compared with the full Whisper Large v3 checkpoint.
speed_optimized
コード
ElevenLabs SDKでのストリーミングSTT
マイクまたはファイルの音声を、ElevenLabs互換のAllModelsルート経由でストリーミングします。
import { ElevenLabsClient } from "elevenlabs";
const client = new ElevenLabsClient({
apiKey: process.env.ALLMODELS_API_KEY,
baseUrl: "https://api.allmodels.io/el"
});
const connection = await client.speechToText.realtime.connect({
modelId: "openai/whisper-large-v3-turbo",
languageCode: "en"
}, {
queryParams: {
provider: {
"only": [
"groq"
]
}
}
});
// Send audio frames to connection, then handle transcript events.