mosibi/man
0
1import os2import sys3import json4import argparse5import subprocess6import spaces7 8now_dir = os.getcwd()9sys.path.append(now_dir)10 11from rvc.configs.config import Config12 13from rvc.lib.tools.prerequisites_download import prequisites_download_pipeline14 15from rvc.infer.infer import infer_pipeline16 17from rvc.lib.tools.model_download import model_download_pipeline18 19config = Config()20current_script_directory = os.path.dirname(os.path.realpath(__file__))21logs_path = os.path.join(current_script_directory, "logs")22 23# Get TTS Voices24with open(os.path.join("rvc", "lib", "tools", "tts_voices.json"), "r") as f:25 voices_data = json.load(f)26 27locales = list({voice["Locale"] for voice in voices_data})28 29 30# Infer31@spaces.GPU32def run_infer_script(33 f0up_key,34 filter_radius,35 index_rate,36 rms_mix_rate,37 protect,38 hop_length,39 f0method,40 input_path,41 output_path,42 pth_path,43 index_path,44 split_audio,45 f0autotune,46 clean_audio,47 clean_strength,48 export_format,49 embedder_model,50 embedder_model_custom,51 upscale_audio,52):53 f0autotune = "True" if str(f0autotune) == "True" else "False"54 clean_audio = "True" if str(clean_audio) == "True" else "False"55 upscale_audio = "True" if str(upscale_audio) == "True" else "False"56 infer_pipeline(57 f0up_key,58 filter_radius,59 index_rate,60 rms_mix_rate,61 protect,62 hop_length,63 f0method,64 input_path,65 output_path,66 pth_path,67 index_path,68 split_audio,69 f0autotune,70 clean_audio,71 clean_strength,72 export_format,73 embedder_model,74 embedder_model_custom,75 upscale_audio,76 )77 return f"File {input_path} inferred successfully.", output_path.replace(78 ".wav", f".{export_format.lower()}"79 )80 81 82# Batch infer83@spaces.GPU84def run_batch_infer_script(85 f0up_key,86 filter_radius,87 index_rate,88 rms_mix_rate,89 protect,90 hop_length,91 f0method,92 input_folder,93 output_folder,94 pth_path,95 index_path,96 split_audio,97 f0autotune,98 clean_audio,99 clean_strength,100 export_format,101 embedder_model,102 embedder_model_custom,103 upscale_audio,104):105 f0autotune = "True" if str(f0autotune) == "True" else "False"106 clean_audio = "True" if str(clean_audio) == "True" else "False"107 upscale_audio = "True" if str(upscale_audio) == "True" else "False"108 audio_files = [109 f for f in os.listdir(input_folder) if f.endswith((".mp3", ".wav", ".flac"))110 ]111 print(f"Detected {len(audio_files)} audio files for inference.")112 113 for audio_file in audio_files:114 if "_output" in audio_file:115 pass116 else:117 input_path = os.path.join(input_folder, audio_file)118 output_file_name = os.path.splitext(os.path.basename(audio_file))[0]119 output_path = os.path.join(120 output_folder,121 f"{output_file_name}_output{os.path.splitext(audio_file)[1]}",122 )123 print(f"Inferring {input_path}...")124 125 infer_pipeline(126 f0up_key,127 filter_radius,128 index_rate,129 rms_mix_rate,130 protect,131 hop_length,132 f0method,133 input_path,134 output_path,135 pth_path,136 index_path,137 split_audio,138 f0autotune,139 clean_audio,140 clean_strength,141 export_format,142 embedder_model,143 embedder_model_custom,144 upscale_audio,145 )146 147 return f"Files from {input_folder} inferred successfully."148 149 150# TTS151@spaces.GPU152def run_tts_script(153 tts_text,154 tts_voice,155 tts_rate,156 f0up_key,157 filter_radius,158 index_rate,159 rms_mix_rate,160 protect,161 hop_length,162 f0method,163 output_tts_path,164 output_rvc_path,165 pth_path,166 index_path,167 split_audio,168 f0autotune,169 clean_audio,170 clean_strength,171 export_format,172 embedder_model,173 embedder_model_custom,174 upscale_audio,175):176 f0autotune = "True" if str(f0autotune) == "True" else "False"177 clean_audio = "True" if str(clean_audio) == "True" else "False"178 upscale_audio = "True" if str(upscale_audio) == "True" else "False"179 tts_script_path = os.path.join("rvc", "lib", "tools", "tts.py")180 181 if os.path.exists(output_tts_path):182 os.remove(output_tts_path)183 184 command_tts = [185 "python",186 tts_script_path,187 tts_text,188 tts_voice,189 str(tts_rate),190 output_tts_path,191 ]192 subprocess.run(command_tts)193 194 infer_pipeline(195 f0up_key,196 filter_radius,197 index_rate,198 rms_mix_rate,199 protect,200 hop_length,201 f0method,202 output_tts_path,203 output_rvc_path,204 pth_path,205 index_path,206 split_audio,207 f0autotune,208 clean_audio,209 clean_strength,210 export_format,211 embedder_model,212 embedder_model_custom,213 upscale_audio,214 )215 216 return f"Text {tts_text} synthesized successfully.", output_rvc_path.replace(217 ".wav", f".{export_format.lower()}"218 )219 220 221# Download222def run_download_script(model_link):223 model_download_pipeline(model_link)224 return f"Model downloaded successfully."225 226 227# Prerequisites228def run_prerequisites_script(pretraineds_v1, pretraineds_v2, models, exe):229 prequisites_download_pipeline(pretraineds_v1, pretraineds_v2, models, exe)230 return "Prerequisites installed successfully."231 232# Parse arguments233def parse_arguments():234 parser = argparse.ArgumentParser(235 description="Run the main.py script with specific parameters."236 )237 subparsers = parser.add_subparsers(238 title="subcommands", dest="mode", help="Choose a mode"239 )240 241 # Parser for 'infer' mode242 infer_parser = subparsers.add_parser("infer", help="Run inference")243 infer_parser.add_argument(244 "--f0up_key",245 type=str,246 help="Value for f0up_key",247 choices=[str(i) for i in range(-24, 25)],248 default="0",249 )250 infer_parser.add_argument(251 "--filter_radius",252 type=str,253 help="Value for filter_radius",254 choices=[str(i) for i in range(11)],255 default="3",256 )257 infer_parser.add_argument(258 "--index_rate",259 type=str,260 help="Value for index_rate",261 choices=[str(i / 10) for i in range(11)],262 default="0.3",263 )264 infer_parser.add_argument(265 "--rms_mix_rate",266 type=str,267 help="Value for rms_mix_rate",268 choices=[str(i / 10) for i in range(11)],269 default="1",270 )271 infer_parser.add_argument(272 "--protect",273 type=str,274 help="Value for protect",275 choices=[str(i / 10) for i in range(6)],276 default="0.33",277 )278 infer_parser.add_argument(279 "--hop_length",280 type=str,281 help="Value for hop_length",282 choices=[str(i) for i in range(1, 513)],283 default="128",284 )285 infer_parser.add_argument(286 "--f0method",287 type=str,288 help="Value for f0method",289 choices=[290 "pm",291 "harvest",292 "dio",293 "crepe",294 "crepe-tiny",295 "rmvpe",296 "fcpe",297 "hybrid[crepe+rmvpe]",298 "hybrid[crepe+fcpe]",299 "hybrid[rmvpe+fcpe]",300 "hybrid[crepe+rmvpe+fcpe]",301 ],302 default="rmvpe",303 )304 infer_parser.add_argument("--input_path", type=str, help="Input path")305 infer_parser.add_argument("--output_path", type=str, help="Output path")306 infer_parser.add_argument("--pth_path", type=str, help="Path to the .pth file")307 infer_parser.add_argument(308 "--index_path",309 type=str,310 help="Path to the .index file",311 )312 infer_parser.add_argument(313 "--split_audio",314 type=str,315 help="Enable split audio",316 choices=["True", "False"],317 default="False",318 )319 infer_parser.add_argument(320 "--f0autotune",321 type=str,322 help="Enable autotune",323 choices=["True", "False"],324 default="False",325 )326 infer_parser.add_argument(327 "--clean_audio",328 type=str,329 help="Enable clean audio",330 choices=["True", "False"],331 default="False",332 )333 infer_parser.add_argument(334 "--clean_strength",335 type=str,336 help="Value for clean_strength",337 choices=[str(i / 10) for i in range(11)],338 default="0.7",339 )340 infer_parser.add_argument(341 "--export_format",342 type=str,343 help="Export format",344 choices=["WAV", "MP3", "FLAC", "OGG", "M4A"],345 default="WAV",346 )347 infer_parser.add_argument(348 "--embedder_model",349 type=str,350 help="Embedder model",351 choices=["contentvec", "hubert", "custom"],352 default="hubert",353 )354 infer_parser.add_argument(355 "--embedder_model_custom",356 type=str,357 help="Custom Embedder model",358 default=None,359 )360 infer_parser.add_argument(361 "--upscale_audio",362 type=str,363 help="Enable audio upscaling",364 choices=["True", "False"],365 default="False",366 )367 368 # Parser for 'batch_infer' mode369 batch_infer_parser = subparsers.add_parser(370 "batch_infer", help="Run batch inference"371 )372 batch_infer_parser.add_argument(373 "--f0up_key",374 type=str,375 help="Value for f0up_key",376 choices=[str(i) for i in range(-24, 25)],377 default="0",378 )379 batch_infer_parser.add_argument(380 "--filter_radius",381 type=str,382 help="Value for filter_radius",383 choices=[str(i) for i in range(11)],384 default="3",385 )386 batch_infer_parser.add_argument(387 "--index_rate",388 type=str,389 help="Value for index_rate",390 choices=[str(i / 10) for i in range(11)],391 default="0.3",392 )393 batch_infer_parser.add_argument(394 "--rms_mix_rate",395 type=str,396 help="Value for rms_mix_rate",397 choices=[str(i / 10) for i in range(11)],398 default="1",399 )400 batch_infer_parser.add_argument(401 "--protect",402 type=str,403 help="Value for protect",404 choices=[str(i / 10) for i in range(6)],405 default="0.33",406 )407 batch_infer_parser.add_argument(408 "--hop_length",409 type=str,410 help="Value for hop_length",411 choices=[str(i) for i in range(1, 513)],412 default="128",413 )414 batch_infer_parser.add_argument(415 "--f0method",416 type=str,417 help="Value for f0method",418 choices=[419 "pm",420 "harvest",421 "dio",422 "crepe",423 "crepe-tiny",424 "rmvpe",425 "fcpe",426 "hybrid[crepe+rmvpe]",427 "hybrid[crepe+fcpe]",428 "hybrid[rmvpe+fcpe]",429 "hybrid[crepe+rmvpe+fcpe]",430 ],431 default="rmvpe",432 )433 batch_infer_parser.add_argument("--input_folder", type=str, help="Input folder")434 batch_infer_parser.add_argument("--output_folder", type=str, help="Output folder")435 batch_infer_parser.add_argument(436 "--pth_path", type=str, help="Path to the .pth file"437 )438 batch_infer_parser.add_argument(439 "--index_path",440 type=str,441 help="Path to the .index file",442 )443 batch_infer_parser.add_argument(444 "--split_audio",445 type=str,446 help="Enable split audio",447 choices=["True", "False"],448 default="False",449 )450 batch_infer_parser.add_argument(451 "--f0autotune",452 type=str,453 help="Enable autotune",454 choices=["True", "False"],455 default="False",456 )457 batch_infer_parser.add_argument(458 "--clean_audio",459 type=str,460 help="Enable clean audio",461 choices=["True", "False"],462 default="False",463 )464 batch_infer_parser.add_argument(465 "--clean_strength",466 type=str,467 help="Value for clean_strength",468 choices=[str(i / 10) for i in range(11)],469 default="0.7",470 )471 batch_infer_parser.add_argument(472 "--export_format",473 type=str,474 help="Export format",475 choices=["WAV", "MP3", "FLAC", "OGG", "M4A"],476 default="WAV",477 )478 batch_infer_parser.add_argument(479 "--embedder_model",480 type=str,481 help="Embedder model",482 choices=["contentvec", "hubert", "custom"],483 default="hubert",484 )485 batch_infer_parser.add_argument(486 "--embedder_model_custom",487 type=str,488 help="Custom Embedder model",489 default=None,490 )491 batch_infer_parser.add_argument(492 "--upscale_audio",493 type=str,494 help="Enable audio upscaling",495 choices=["True", "False"],496 default="False",497 )498 499 # Parser for 'tts' mode500 tts_parser = subparsers.add_parser("tts", help="Run TTS")501 tts_parser.add_argument(502 "--tts_text",503 type=str,504 help="Text to be synthesized",505 )506 tts_parser.add_argument(507 "--tts_voice",508 type=str,509 help="Voice to be used",510 choices=locales,511 )512 tts_parser.add_argument(513 "--tts_rate",514 type=str,515 help="Increase or decrease TTS speed",516 choices=[str(i) for i in range(-100, 100)],517 default="0",518 )519 tts_parser.add_argument(520 "--f0up_key",521 type=str,522 help="Value for f0up_key",523 choices=[str(i) for i in range(-24, 25)],524 default="0",525 )526 tts_parser.add_argument(527 "--filter_radius",528 type=str,529 help="Value for filter_radius",530 choices=[str(i) for i in range(11)],531 default="3",532 )533 tts_parser.add_argument(534 "--index_rate",535 type=str,536 help="Value for index_rate",537 choices=[str(i / 10) for i in range(11)],538 default="0.3",539 )540 tts_parser.add_argument(541 "--rms_mix_rate",542 type=str,543 help="Value for rms_mix_rate",544 choices=[str(i / 10) for i in range(11)],545 default="1",546 )547 tts_parser.add_argument(548 "--protect",549 type=str,550 help="Value for protect",551 choices=[str(i / 10) for i in range(6)],552 default="0.33",553 )554 tts_parser.add_argument(555 "--hop_length",556 type=str,557 help="Value for hop_length",558 choices=[str(i) for i in range(1, 513)],559 default="128",560 )561 tts_parser.add_argument(562 "--f0method",563 type=str,564 help="Value for f0method",565 choices=[566 "pm",567 "harvest",568 "dio",569 "crepe",570 "crepe-tiny",571 "rmvpe",572 "fcpe",573 "hybrid[crepe+rmvpe]",574 "hybrid[crepe+fcpe]",575 "hybrid[rmvpe+fcpe]",576 "hybrid[crepe+rmvpe+fcpe]",577 ],578 default="rmvpe",579 )580 tts_parser.add_argument("--output_tts_path", type=str, help="Output tts path")581 tts_parser.add_argument("--output_rvc_path", type=str, help="Output rvc path")582 tts_parser.add_argument("--pth_path", type=str, help="Path to the .pth file")583 tts_parser.add_argument(584 "--index_path",585 type=str,586 help="Path to the .index file",587 )588 tts_parser.add_argument(589 "--split_audio",590 type=str,591 help="Enable split audio",592 choices=["True", "False"],593 default="False",594 )595 tts_parser.add_argument(596 "--f0autotune",597 type=str,598 help="Enable autotune",599 choices=["True", "False"],600 default="False",601 )602 tts_parser.add_argument(603 "--clean_audio",604 type=str,605 help="Enable clean audio",606 choices=["True", "False"],607 default="False",608 )609 tts_parser.add_argument(610 "--clean_strength",611 type=str,612 help="Value for clean_strength",613 choices=[str(i / 10) for i in range(11)],614 default="0.7",615 )616 tts_parser.add_argument(617 "--export_format",618 type=str,619 help="Export format",620 choices=["WAV", "MP3", "FLAC", "OGG", "M4A"],621 default="WAV",622 )623 tts_parser.add_argument(624 "--embedder_model",625 type=str,626 help="Embedder model",627 choices=["contentvec", "hubert", "custom"],628 default="hubert",629 )630 tts_parser.add_argument(631 "--embedder_model_custom",632 type=str,633 help="Custom Embedder model",634 default=None,635 )636 tts_parser.add_argument(637 "--upscale_audio",638 type=str,639 help="Enable audio upscaling",640 choices=["True", "False"],641 default="False",642 )643 644 # Parser for 'download' mode645 download_parser = subparsers.add_parser("download", help="Download models")646 download_parser.add_argument(647 "--model_link",648 type=str,649 help="Link of the model",650 )651 652 # Parser for 'prerequisites' mode653 prerequisites_parser = subparsers.add_parser(654 "prerequisites", help="Install prerequisites"655 )656 prerequisites_parser.add_argument(657 "--pretraineds_v1",658 type=str,659 choices=["True", "False"],660 default="True",661 help="Download pretrained models for v1",662 )663 prerequisites_parser.add_argument(664 "--pretraineds_v2",665 type=str,666 choices=["True", "False"],667 default="True",668 help="Download pretrained models for v2",669 )670 prerequisites_parser.add_argument(671 "--models",672 type=str,673 choices=["True", "False"],674 default="True",675 help="Donwload models",676 )677 prerequisites_parser.add_argument(678 "--exe",679 type=str,680 choices=["True", "False"],681 default="True",682 help="Download executables",683 )684 685 return parser.parse_args()686 687 688def main():689 if len(sys.argv) == 1:690 print("Please run the script with '-h' for more information.")691 sys.exit(1)692 693 args = parse_arguments()694 695 try:696 if args.mode == "infer":697 run_infer_script(698 str(args.f0up_key),699 str(args.filter_radius),700 str(args.index_rate),701 str(args.rms_mix_rate),702 str(args.protect),703 str(args.hop_length),704 str(args.f0method),705 str(args.input_path),706 str(args.output_path),707 str(args.pth_path),708 str(args.index_path),709 str(args.split_audio),710 str(args.f0autotune),711 str(args.clean_audio),712 str(args.clean_strength),713 str(args.export_format),714 str(args.embedder_model),715 str(args.embedder_model_custom),716 str(args.upscale_audio),717 )718 elif args.mode == "batch_infer":719 run_batch_infer_script(720 str(args.f0up_key),721 str(args.filter_radius),722 str(args.index_rate),723 str(args.rms_mix_rate),724 str(args.protect),725 str(args.hop_length),726 str(args.f0method),727 str(args.input_folder),728 str(args.output_folder),729 str(args.pth_path),730 str(args.index_path),731 str(args.split_audio),732 str(args.f0autotune),733 str(args.clean_audio),734 str(args.clean_strength),735 str(args.export_format),736 str(args.embedder_model),737 str(args.embedder_model_custom),738 str(args.upscale_audio),739 )740 elif args.mode == "tts":741 run_tts_script(742 str(args.tts_text),743 str(args.tts_voice),744 str(args.tts_rate),745 str(args.f0up_key),746 str(args.filter_radius),747 str(args.index_rate),748 str(args.rms_mix_rate),749 str(args.protect),750 str(args.hop_length),751 str(args.f0method),752 str(args.output_tts_path),753 str(args.output_rvc_path),754 str(args.pth_path),755 str(args.index_path),756 str(args.split_audio),757 str(args.f0autotune),758 str(args.clean_audio),759 str(args.clean_strength),760 str(args.export_format),761 str(args.embedder_model),762 str(args.embedder_model_custom),763 str(args.upscale_audio),764 )765 elif args.mode == "download":766 run_download_script(767 str(args.model_link),768 )769 elif args.mode == "prerequisites":770 run_prerequisites_script(771 str(args.pretraineds_v1),772 str(args.pretraineds_v2),773 str(args.models),774 str(args.exe),775 )776 except Exception as error:777 print(f"Error: {error}")778 779 780if __name__ == "__main__":781 main()782