bosonai

/

higgs-audio-v3-tts-4b

like 349

Follow

Boson AI 507

I am getting the Tupt sound at regular intervals (between 2-3 secs) of time when using the stream endpoint with sglang-omni

import base64
import json
import requests
import sounddevice as sd

BASE_URL = "http://10.10.52.25:8080"

inp = """<|emotion:awe|><|prosody:expressive_high|><|prosody:speed_slow|>
There are moments in life that change everything.
<|prosody:pause|>
Not because they are loud,
or dramatic,
or impossible to ignore.
<|prosody:long_pause|>
But because, in a single instant,
they remind us who we really are.
<|emotion:determination|>
And from that moment forward,
we choose to move ahead,
one step at a time."""

response = requests.post(
    f"{BASE_URL}/v1/audio/speech",
    json={
        "input": inp,
        "stream": True,
    },
    stream=True,
)

response.raise_for_status()

stream = sd.RawOutputStream(
    samplerate=24000,
    channels=1,
    dtype="int16",
)

stream.start()

try:
    for line in response.iter_lines():

        if not line:
            continue

        if isinstance(line, bytes):
            line = line.decode("utf-8")

        if not line.startswith("data: "):
            continue

        if line == "data: [DONE]":
            break

        event = json.loads(line[6:])

        audio = event.get("audio", {})

        if audio.get("data"):
            stream.write(
                base64.b64decode(audio["data"])
            )

finally:
    stream.stop()
    stream.close()
    response.close()