import os ## 01 step # os.system(f"python resample.py") # os.system(f"python preprocess_flist_config.py --speech_encoder vec768l12") ## 02 step speech_encoder = "vec768l12" #@param ["vec768l12", "vec256l9", "hubertsoft", "whisper-ppg", "whisper-ppg-large"] from pretrain.meta import download_dict download_dict = download_dict() url = download_dict[speech_encoder]["url"] output = download_dict[speech_encoder]["output"] if not os.path.exists(output): os.system(f"curl -L {url} -o {output}") ##03 step f0_predictor = "crepe" #@param ["crepe", "pm", "dio", "harvest"] use_diff = True #@param {type:"boolean"} diff_param = "" if use_diff: diff_param = "--use_diff" if not os.path.exists("./pretrain/nsf_hifigan/model"): os.system(f"curl -L https://github.com/openvpi/vocoders/releases/download/nsf-hifigan-v1/nsf_hifigan_20221211.zip -o nsf_hifigan_20221211.zip") os.system(f"unzip nsf_hifigan_20221211.zip") os.system(f"rm -rf pretrain/nsf_hifigan") os.system(f"mv -v nsf_hifigan pretrain") os.system(f"python preprocess_hubert_f0.py --f0_predictor={f0_predictor} {diff_param}") ##04 step os.system(f"python train.py -c configs/config.json -m 44k")