import argparse import os import sys print("Command-line arguments:", sys.argv) now_dir = os.getcwd() sys.path.append(now_dir) import sys import tqdm as tq from dotenv import load_dotenv from scipy.io import wavfile from configs.config import Config from infer.modules.vc.modules import VC def arg_parse() -> tuple: parser = argparse.ArgumentParser() parser.add_argument("--f0up_key", type=int, default=0) parser.add_argument("--input_path", type=str, help="input path") parser.add_argument("--index_path", type=str, help="index path") parser.add_argument("--f0method", type=str, default="harvest", help="harvest or pm") parser.add_argument("--opt_path", type=str, help="opt path") parser.add_argument("--model_name", type=str, help="store in assets/weight_root") parser.add_argument("--index_rate", type=float, default=0.66, help="index rate") parser.add_argument("--device", type=str, help="device") parser.add_argument("--is_half", type=bool, help="use half -> True") parser.add_argument("--filter_radius", type=int, default=3, help="filter radius") parser.add_argument("--resample_sr", type=int, default=0, help="resample sr") parser.add_argument("--rms_mix_rate", type=float, default=1, help="rms mix rate") parser.add_argument("--protect", type=float, default=0.33, help="protect") args = parser.parse_args() sys.argv = sys.argv[:1] return args def main(): load_dotenv() args = arg_parse() config = Config() config.device = args.device if args.device else config.device config.is_half = args.is_half if args.is_half else config.is_half vc = VC(config) vc.get_vc(args.model_name) audios = os.listdir(args.input_path) for file in tq.tqdm(audios): if file.endswith(".wav"): file_path = os.path.join(args.input_path, file) _, wav_opt = vc.vc_single( 0, file_path, args.f0up_key, None, args.f0method, args.index_path, None, args.index_rate, args.filter_radius, args.resample_sr, args.rms_mix_rate, args.protect, ) out_path = os.path.join(args.opt_path, file) wavfile.write(out_path, wav_opt[0], wav_opt[1]) if __name__ == "__main__": main()