更新
This commit is contained in:
Kevin Wong
92
models/CosyVoice/runtime/python/fastapi/client.py
Normal file
92
models/CosyVoice/runtime/python/fastapi/client.py
Normal file
@@ -0,0 +1,92 @@
|
||||
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
# You may obtain a copy of the License at
|
||||
#
|
||||
# http://www.apache.org/licenses/LICENSE-2.0
|
||||
#
|
||||
# Unless required by applicable law or agreed to in writing, software
|
||||
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
import argparse
|
||||
import logging
|
||||
import requests
|
||||
import torch
|
||||
import torchaudio
|
||||
import numpy as np
|
||||
|
||||
|
||||
def main():
|
||||
url = "http://{}:{}/inference_{}".format(args.host, args.port, args.mode)
|
||||
if args.mode == 'sft':
|
||||
payload = {
|
||||
'tts_text': args.tts_text,
|
||||
'spk_id': args.spk_id
|
||||
}
|
||||
response = requests.request("GET", url, data=payload, stream=True)
|
||||
elif args.mode == 'zero_shot':
|
||||
payload = {
|
||||
'tts_text': args.tts_text,
|
||||
'prompt_text': args.prompt_text
|
||||
}
|
||||
files = [('prompt_wav', ('prompt_wav', open(args.prompt_wav, 'rb'), 'application/octet-stream'))]
|
||||
response = requests.request("GET", url, data=payload, files=files, stream=True)
|
||||
elif args.mode == 'cross_lingual':
|
||||
payload = {
|
||||
'tts_text': args.tts_text,
|
||||
}
|
||||
files = [('prompt_wav', ('prompt_wav', open(args.prompt_wav, 'rb'), 'application/octet-stream'))]
|
||||
response = requests.request("GET", url, data=payload, files=files, stream=True)
|
||||
else:
|
||||
payload = {
|
||||
'tts_text': args.tts_text,
|
||||
'spk_id': args.spk_id,
|
||||
'instruct_text': args.instruct_text
|
||||
}
|
||||
response = requests.request("GET", url, data=payload, stream=True)
|
||||
tts_audio = b''
|
||||
for r in response.iter_content(chunk_size=16000):
|
||||
tts_audio += r
|
||||
tts_speech = torch.from_numpy(np.array(np.frombuffer(tts_audio, dtype=np.int16))).unsqueeze(dim=0)
|
||||
logging.info('save response to {}'.format(args.tts_wav))
|
||||
torchaudio.save(args.tts_wav, tts_speech, target_sr)
|
||||
logging.info('get response')
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument('--host',
|
||||
type=str,
|
||||
default='0.0.0.0')
|
||||
parser.add_argument('--port',
|
||||
type=int,
|
||||
default='50000')
|
||||
parser.add_argument('--mode',
|
||||
default='sft',
|
||||
choices=['sft', 'zero_shot', 'cross_lingual', 'instruct'],
|
||||
help='request mode')
|
||||
parser.add_argument('--tts_text',
|
||||
type=str,
|
||||
default='你好，我是通义千问语音合成大模型，请问有什么可以帮您的吗？')
|
||||
parser.add_argument('--spk_id',
|
||||
type=str,
|
||||
default='中文女')
|
||||
parser.add_argument('--prompt_text',
|
||||
type=str,
|
||||
default='希望你以后能够做的比我还好呦。')
|
||||
parser.add_argument('--prompt_wav',
|
||||
type=str,
|
||||
default='../../../asset/zero_shot_prompt.wav')
|
||||
parser.add_argument('--instruct_text',
|
||||
type=str,
|
||||
default='Theo \'Crimson\', is a fiery, passionate rebel leader. \
|
||||
Fights with fervor for justice, but struggles with impulsiveness.')
|
||||
parser.add_argument('--tts_wav',
|
||||
type=str,
|
||||
default='demo.wav')
|
||||
args = parser.parse_args()
|
||||
prompt_sr, target_sr = 16000, 22050
|
||||
main()
|
||||
95
models/CosyVoice/runtime/python/fastapi/server.py
Normal file
95
models/CosyVoice/runtime/python/fastapi/server.py
Normal file
@@ -0,0 +1,95 @@
|
||||
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
|
||||
#
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
# you may not use this file except in compliance with the License.
|
||||
# You may obtain a copy of the License at
|
||||
#
|
||||
# http://www.apache.org/licenses/LICENSE-2.0
|
||||
#
|
||||
# Unless required by applicable law or agreed to in writing, software
|
||||
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
# See the License for the specific language governing permissions and
|
||||
# limitations under the License.
|
||||
import os
|
||||
import sys
|
||||
import argparse
|
||||
import logging
|
||||
logging.getLogger('matplotlib').setLevel(logging.WARNING)
|
||||
from fastapi import FastAPI, UploadFile, Form, File
|
||||
from fastapi.responses import StreamingResponse
|
||||
from fastapi.middleware.cors import CORSMiddleware
|
||||
import uvicorn
|
||||
import numpy as np
|
||||
ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
|
||||
sys.path.append('{}/../../..'.format(ROOT_DIR))
|
||||
sys.path.append('{}/../../../third_party/Matcha-TTS'.format(ROOT_DIR))
|
||||
from cosyvoice.cli.cosyvoice import AutoModel
|
||||
from cosyvoice.utils.file_utils import load_wav
|
||||
|
||||
app = FastAPI()
|
||||
# set cross region allowance
|
||||
app.add_middleware(
|
||||
CORSMiddleware,
|
||||
allow_origins=["*"],
|
||||
allow_credentials=True,
|
||||
allow_methods=["*"],
|
||||
allow_headers=["*"])
|
||||
|
||||
|
||||
def generate_data(model_output):
|
||||
for i in model_output:
|
||||
tts_audio = (i['tts_speech'].numpy() * (2 ** 15)).astype(np.int16).tobytes()
|
||||
yield tts_audio
|
||||
|
||||
|
||||
@app.get("/inference_sft")
|
||||
@app.post("/inference_sft")
|
||||
async def inference_sft(tts_text: str = Form(), spk_id: str = Form()):
|
||||
model_output = cosyvoice.inference_sft(tts_text, spk_id)
|
||||
return StreamingResponse(generate_data(model_output))
|
||||
|
||||
|
||||
@app.get("/inference_zero_shot")
|
||||
@app.post("/inference_zero_shot")
|
||||
async def inference_zero_shot(tts_text: str = Form(), prompt_text: str = Form(), prompt_wav: UploadFile = File()):
|
||||
prompt_speech_16k = load_wav(prompt_wav.file, 16000)
|
||||
model_output = cosyvoice.inference_zero_shot(tts_text, prompt_text, prompt_speech_16k)
|
||||
return StreamingResponse(generate_data(model_output))
|
||||
|
||||
|
||||
@app.get("/inference_cross_lingual")
|
||||
@app.post("/inference_cross_lingual")
|
||||
async def inference_cross_lingual(tts_text: str = Form(), prompt_wav: UploadFile = File()):
|
||||
prompt_speech_16k = load_wav(prompt_wav.file, 16000)
|
||||
model_output = cosyvoice.inference_cross_lingual(tts_text, prompt_speech_16k)
|
||||
return StreamingResponse(generate_data(model_output))
|
||||
|
||||
|
||||
@app.get("/inference_instruct")
|
||||
@app.post("/inference_instruct")
|
||||
async def inference_instruct(tts_text: str = Form(), spk_id: str = Form(), instruct_text: str = Form()):
|
||||
model_output = cosyvoice.inference_instruct(tts_text, spk_id, instruct_text)
|
||||
return StreamingResponse(generate_data(model_output))
|
||||
|
||||
|
||||
@app.get("/inference_instruct2")
|
||||
@app.post("/inference_instruct2")
|
||||
async def inference_instruct2(tts_text: str = Form(), instruct_text: str = Form(), prompt_wav: UploadFile = File()):
|
||||
prompt_speech_16k = load_wav(prompt_wav.file, 16000)
|
||||
model_output = cosyvoice.inference_instruct2(tts_text, instruct_text, prompt_speech_16k)
|
||||
return StreamingResponse(generate_data(model_output))
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument('--port',
|
||||
type=int,
|
||||
default=50000)
|
||||
parser.add_argument('--model_dir',
|
||||
type=str,
|
||||
default='iic/CosyVoice2-0.5B',
|
||||
help='local path or modelscope repo id')
|
||||
args = parser.parse_args()
|
||||
cosyvoice = AutoModel(model_dir=args.model_dir)
|
||||
uvicorn.run(app, host="0.0.0.0", port=args.port)
|
||||
Reference in New Issue
Block a user