f197177ae1
- FastAPI 백엔드: /api/voices, /api/tts, /api/health - 한국어 메인 + 영어(5종 악센트), 중국어/일본어 UI 제거 - 언어별 목소리 모델 선택, 속도(0.5~2.0x)/피치(±12반음) 조절 - 로그인/사용제한/과금 유도 없음 (무제한 테스트) - NVIDIA GPU 자동 사용 (torch cu128, Blackwell sm_120) - 유저 친화 한국어 UI (frontend) - Docker/compose (GPU 예약), 8788 포트 - MeloTTS(순수 파이썬) 벤더링 + 검증된 버전 고정(constraints)
37 lines
1.8 KiB
Python
37 lines
1.8 KiB
Python
import click
|
|
import warnings
|
|
import os
|
|
|
|
|
|
@click.command
|
|
@click.argument('text')
|
|
@click.argument('output_path')
|
|
@click.option("--file", '-f', is_flag=True, show_default=True, default=False, help="Text is a file")
|
|
@click.option('--language', '-l', default='EN', help='Language, defaults to English', type=click.Choice(['EN', 'ES', 'FR', 'ZH', 'JP', 'KR'], case_sensitive=False))
|
|
@click.option('--speaker', '-spk', default='EN-Default', help='Speaker ID, only for English, leave empty for default, ignored if not English. If English, defaults to "EN-Default"', type=click.Choice(['EN-Default', 'EN-US', 'EN-BR', 'EN_INDIA', 'EN-AU']))
|
|
@click.option('--speed', '-s', default=1.0, help='Speed, defaults to 1.0', type=float)
|
|
@click.option('--device', '-d', default='auto', help='Device, defaults to auto')
|
|
def main(text, file, output_path, language, speaker, speed, device):
|
|
if file:
|
|
if not os.path.exists(text):
|
|
raise FileNotFoundError(f'Trying to load text from file due to --file/-f flag, but file not found. Remove the --file/-f flag to pass a string.')
|
|
else:
|
|
with open(text) as f:
|
|
text = f.read().strip()
|
|
if text == '':
|
|
raise ValueError('You entered empty text or the file you passed was empty.')
|
|
language = language.upper()
|
|
if language == '': language = 'EN'
|
|
if speaker == '': speaker = None
|
|
if (not language == 'EN') and speaker:
|
|
warnings.warn('You specified a speaker but the language is English.')
|
|
from melo.api import TTS
|
|
model = TTS(language=language, device=device)
|
|
speaker_ids = model.hps.data.spk2id
|
|
if language == 'EN':
|
|
if not speaker: speaker = 'EN-Default'
|
|
spkr = speaker_ids[speaker]
|
|
else:
|
|
spkr = speaker_ids[list(speaker_ids.keys())[0]]
|
|
model.tts_to_file(text, spkr, output_path, speed=speed)
|