mirror of
https://github.com/BoardWare-Genius/jarvis-models.git
synced 2025-12-14 09:03:25 +00:00
93 lines
3.2 KiB
Python
93 lines
3.2 KiB
Python
import io
|
|
import time
|
|
|
|
import requests
|
|
from fastapi import Request, Response, status
|
|
from fastapi.responses import JSONResponse
|
|
from injector import inject
|
|
from injector import singleton
|
|
|
|
from ..log.logging_time import logging_time
|
|
|
|
from ..configuration import CosyVoiceConf
|
|
from .blackbox import Blackbox
|
|
|
|
import soundfile
|
|
import pyloudnorm as pyln
|
|
import sys
|
|
sys.path.append('/home/gpu/Workspace/CosyVoice')
|
|
from cosyvoice.cli.cosyvoice import CosyVoice
|
|
from cosyvoice.utils.file_utils import load_wav
|
|
import torchaudio
|
|
|
|
import os
|
|
import logging
|
|
logger = logging.getLogger(__name__)
|
|
|
|
@singleton
|
|
class CosyVoiceTTS(Blackbox):
|
|
mode: str
|
|
url: str
|
|
speed: int
|
|
device: str
|
|
language: str
|
|
speaker: str
|
|
|
|
@logging_time(logger=logger)
|
|
def model_init(self, cosyvoice_config: CosyVoiceConf) -> None:
|
|
self.speed = cosyvoice_config.speed
|
|
self.device = cosyvoice_config.device
|
|
self.language = cosyvoice_config.language
|
|
self.speaker = cosyvoice_config.speaker
|
|
self.device = cosyvoice_config.device
|
|
self.url = ''
|
|
self.mode = cosyvoice_config.mode
|
|
self.cosyvoicetts = None
|
|
self.speaker_ids = None
|
|
os.environ['CUDA_VISIBLE_DEVICES'] = str(cosyvoice_config.device)
|
|
if self.mode == 'local':
|
|
self.cosyvoicetts = CosyVoice('/home/gpu/Workspace/Models/CosyVoice/pretrained_models/CosyVoice-300M')
|
|
|
|
else:
|
|
self.url = cosyvoice_config.url
|
|
logging.info('#### Initializing CosyVoiceTTS Service in cuda:' + str(cosyvoice_config.device) + ' mode...')
|
|
|
|
@inject
|
|
def __init__(self, cosyvoice_config: CosyVoiceConf) -> None:
|
|
self.model_init(cosyvoice_config)
|
|
|
|
def __call__(self, *args, **kwargs):
|
|
return self.processing(*args, **kwargs)
|
|
|
|
def valid(self, *args, **kwargs) -> bool:
|
|
text = args[0]
|
|
return isinstance(text, str)
|
|
|
|
@logging_time(logger=logger)
|
|
def processing(self, *args, **kwargs) -> io.BytesIO | bytes:
|
|
text = args[0]
|
|
current_time = time.time()
|
|
if self.mode == 'local':
|
|
audio = self.cosyvoicetts.inference_sft(text, self.language)
|
|
f = io.BytesIO()
|
|
soundfile.write(f, audio['tts_speech'].cpu().numpy().squeeze(0), 22050, format='wav')
|
|
f.seek(0)
|
|
print("#### CosyVoiceTTS Service consume - local : ", (time.time() - current_time))
|
|
return f.read()
|
|
else:
|
|
message = {
|
|
"text": text
|
|
}
|
|
response = requests.post(self.url, json=message)
|
|
print("#### CosyVoiceTTS Service consume - docker : ", (time.time()-current_time))
|
|
return response.content
|
|
|
|
async def fast_api_handler(self, request: Request) -> Response:
|
|
try:
|
|
data = await request.json()
|
|
except:
|
|
return JSONResponse(content={"error": "json parse error"}, status_code=status.HTTP_400_BAD_REQUEST)
|
|
text = data.get("text")
|
|
if text is None:
|
|
return JSONResponse(content={"error": "text is required"}, status_code=status.HTTP_400_BAD_REQUEST)
|
|
return Response(content=self.processing(text), media_type="audio/wav", headers={"Content-Disposition": "attachment; filename=audio.wav"}) |