diff --git a/extract_f0_print.py b/extract_f0_print.py index 0f6f5a8..4f6c806 100644 --- a/extract_f0_print.py +++ b/extract_f0_print.py @@ -79,9 +79,7 @@ class FeatureInput(object): from lib.rmvpe import RMVPE print("loading rmvpe model") - self.model_rmvpe = RMVPE( - "rmvpe.pt", is_half=False, device="cpu" - ) + self.model_rmvpe = RMVPE("rmvpe.pt", is_half=False, device="cpu") f0 = self.model_rmvpe.infer_from_audio(x, thred=0.03) return f0 diff --git a/extract_f0_rmvpe.py b/extract_f0_rmvpe.py index a0c9bba..55dd97b 100644 --- a/extract_f0_rmvpe.py +++ b/extract_f0_rmvpe.py @@ -23,7 +23,6 @@ def printt(strr): f.flush() - class FeatureInput(object): def __init__(self, samplerate=16000, hop_size=160): self.fs = samplerate @@ -38,14 +37,12 @@ class FeatureInput(object): def compute_f0(self, path, f0_method): x = load_audio(path, self.fs) p_len = x.shape[0] // self.hop - if(f0_method=="rmvpe"): + if f0_method == "rmvpe": if hasattr(self, "model_rmvpe") == False: from lib.rmvpe import RMVPE print("loading rmvpe model") - self.model_rmvpe = RMVPE( - "rmvpe.pt", is_half=True, device="cuda" - ) + self.model_rmvpe = RMVPE("rmvpe.pt", is_half=True, device="cuda") f0 = self.model_rmvpe.infer_from_audio(x, thred=0.03) return f0 @@ -117,7 +114,7 @@ if __name__ == "__main__": opt_path2 = "%s/%s" % (opt_root2, name) paths.append([inp_path, opt_path1, opt_path2]) try: - featureInput.go(paths[i_part::n_part],"rmvpe") + featureInput.go(paths[i_part::n_part], "rmvpe") except: printt("f0_all_fail-%s" % (traceback.format_exc())) # ps = [] diff --git a/infer-web.py b/infer-web.py index 13b2964..b61ac87 100644 --- a/infer-web.py +++ b/infer-web.py @@ -43,9 +43,7 @@ logging.getLogger("numba").setLevel(logging.WARNING) now_dir = os.getcwd() tmp = os.path.join(now_dir, "TEMP") shutil.rmtree(tmp, ignore_errors=True) -shutil.rmtree( - "%s/runtime/Lib/site-packages/infer_pack" % (now_dir), ignore_errors=True -) +shutil.rmtree("%s/runtime/Lib/site-packages/infer_pack" % (now_dir), ignore_errors=True) shutil.rmtree("%s/runtime/Lib/site-packages/uvr5_pack" % (now_dir), ignore_errors=True) os.makedirs(tmp, exist_ok=True) os.makedirs(os.path.join(now_dir, "logs"), exist_ok=True) @@ -570,13 +568,13 @@ def preprocess_dataset(trainset_dir, exp_dir, sr, n_p): # but2.click(extract_f0,[gpus6,np7,f0method8,if_f0_3,trainset_dir4],[info2]) -def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19,gpus_rmvpe): +def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19, gpus_rmvpe): gpus = gpus.split("-") os.makedirs("%s/logs/%s" % (now_dir, exp_dir), exist_ok=True) f = open("%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "w") f.close() if if_f0: - if(f0method!="rmvpe_gpu"): + if f0method != "rmvpe_gpu": cmd = config.python_cmd + ' extract_f0_print.py "%s/logs/%s" %s %s' % ( now_dir, exp_dir, @@ -584,7 +582,9 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19,gpus_rmvpe f0method, ) print(cmd) - p = Popen(cmd, shell=True, cwd=now_dir) # , stdin=PIPE, stdout=PIPE,stderr=PIPE + p = Popen( + cmd, shell=True, cwd=now_dir + ) # , stdin=PIPE, stdout=PIPE,stderr=PIPE ###煞笔gr, popen read都非得全跑完了再一次性读取, 不用gr就正常读一句输出一句;只能额外弄出一个文本流定时读 done = [False] threading.Thread( @@ -602,7 +602,9 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19,gpus_rmvpe sleep(1) if done[0]: break - with open("%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r") as f: + with open( + "%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r" + ) as f: log = f.read() print(log) yield log @@ -612,16 +614,9 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19,gpus_rmvpe ps = [] for idx, n_g in enumerate(gpus_rmvpe): cmd = ( - config.python_cmd - + ' extract_f0_rmvpe.py %s %s %s "%s/logs/%s" %s ' - % ( - leng, - idx, - n_g, - now_dir, - exp_dir, - config.is_half - ) + config.python_cmd + + ' extract_f0_rmvpe.py %s %s %s "%s/logs/%s" %s ' + % (leng, idx, n_g, now_dir, exp_dir, config.is_half) ) print(cmd) p = Popen( @@ -638,12 +633,16 @@ def extract_f0_feature(gpus, n_p, f0method, if_f0, exp_dir, version19,gpus_rmvpe ), ).start() while 1: - with open("%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r") as f: + with open( + "%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r" + ) as f: yield (f.read()) sleep(1) if done[0]: break - with open("%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r") as f: + with open( + "%s/logs/%s/extract_f0_feature.log" % (now_dir, exp_dir), "r" + ) as f: log = f.read() print(log) yield log @@ -1037,7 +1036,8 @@ def train1key( gpus16, if_cache_gpu17, if_save_every_weights18, - version19,gpus_rmvpe + version19, + gpus_rmvpe, ): infos = [] @@ -1074,7 +1074,7 @@ def train1key( open(extract_f0_feature_log_path, "w") if if_f0_3: yield get_info_str("step2a:正在提取音高") - if(f0method8!="rmvpe_gpu"): + if f0method8 != "rmvpe_gpu": cmd = config.python_cmd + ' extract_f0_print.py "%s" %s %s' % ( model_log_dir, np7, @@ -1088,16 +1088,12 @@ def train1key( leng = len(gpus_rmvpe) ps = [] for idx, n_g in enumerate(gpus_rmvpe): - cmd = ( - config.python_cmd - + ' extract_f0_rmvpe.py %s %s %s "%s" %s ' - % ( - leng, - idx, - n_g, - model_log_dir, - config.is_half - ) + cmd = config.python_cmd + ' extract_f0_rmvpe.py %s %s %s "%s" %s ' % ( + leng, + idx, + n_g, + model_log_dir, + config.is_half, ) yield get_info_str(cmd) p = Popen( @@ -1318,11 +1314,15 @@ def change_info_(ckpt_path): traceback.print_exc() return {"__type__": "update"}, {"__type__": "update"}, {"__type__": "update"} + def change_f0_method(f0method8): - if(f0method8=="rmvpe_gpu"):visible=True - else:visible=False + if f0method8 == "rmvpe_gpu": + visible = True + else: + visible = False return {"visible": visible, "__type__": "update"} + def export_onnx(ModelPath, ExportedPath): global cpt cpt = torch.load(ModelPath, map_location="cpu") @@ -1755,10 +1755,12 @@ with gr.Blocks(title="RVC WebUI") as app: interactive=True, ) gpus_rmvpe = gr.Textbox( - label=i18n("rmvpe卡号配置：以-分隔输入使用的不同进程卡号,例如0-0-1使用在卡0上跑2个进程并在卡1上跑1个进程"), - value="%s-%s"%(gpus,gpus), + label=i18n( + "rmvpe卡号配置：以-分隔输入使用的不同进程卡号,例如0-0-1使用在卡0上跑2个进程并在卡1上跑1个进程" + ), + value="%s-%s" % (gpus, gpus), interactive=True, - visible=True + visible=True, ) but2 = gr.Button(i18n("特征提取"), variant="primary") info2 = gr.Textbox(label=i18n("输出信息"), value="", max_lines=8) @@ -1769,7 +1771,15 @@ with gr.Blocks(title="RVC WebUI") as app: ) but2.click( extract_f0_feature, - [gpus6, np7, f0method8, if_f0_3, exp_dir1, version19,gpus_rmvpe], + [ + gpus6, + np7, + f0method8, + if_f0_3, + exp_dir1, + version19, + gpus_rmvpe, + ], [info2], ) with gr.Group(): @@ -1894,7 +1904,8 @@ with gr.Blocks(title="RVC WebUI") as app: gpus16, if_cache_gpu17, if_save_every_weights18, - version19,gpus_rmvpe + version19, + gpus_rmvpe, ], info3, ) diff --git a/infer_cli.py b/infer_cli.py index 75e905f..5eadea6 100644 --- a/infer_cli.py +++ b/infer_cli.py @@ -1,4 +1,5 @@ -import os,sys,pdb,torch +import os, sys, pdb, torch + now_dir = os.getcwd() sys.path.append(now_dir) import argparse @@ -9,35 +10,36 @@ import numpy as np from multiprocessing import cpu_count #### -#USAGE +# USAGE # -#In your Terminal or CMD or whatever -#python infer_cli.py [TRANSPOSE_VALUE] "[INPUT_PATH]" "[OUTPUT_PATH]" "[MODEL_PATH]" "[INDEX_FILE_PATH]" "[INFERENCE_DEVICE]" "[METHOD]" +# In your Terminal or CMD or whatever +# python infer_cli.py [TRANSPOSE_VALUE] "[INPUT_PATH]" "[OUTPUT_PATH]" "[MODEL_PATH]" "[INDEX_FILE_PATH]" "[INFERENCE_DEVICE]" "[METHOD]" using_cli = False device = "cuda:0" is_half = False -if(len(sys.argv) > 0): - f0_up_key=int(sys.argv[1]) #transpose value - input_path=sys.argv[2] - output_path=sys.argv[3] - model_path=sys.argv[4] - file_index=sys.argv[5] #.index file - device=sys.argv[6] - f0_method=sys.argv[7] #pm or harvest or crepe - +if len(sys.argv) > 0: + f0_up_key = int(sys.argv[1]) # transpose value + input_path = sys.argv[2] + output_path = sys.argv[3] + model_path = sys.argv[4] + file_index = sys.argv[5] # .index file + device = sys.argv[6] + f0_method = sys.argv[7] # pm or harvest or crepe + using_cli = True - #file_index2=sys.argv[8] - #index_rate=float(sys.argv[10]) #search feature ratio - #filter_radius=float(sys.argv[11]) #median filter - #resample_sr=float(sys.argv[12]) #resample audio in post processing - #rms_mix_rate=float(sys.argv[13]) #search feature + # file_index2=sys.argv[8] + # index_rate=float(sys.argv[10]) #search feature ratio + # filter_radius=float(sys.argv[11]) #median filter + # resample_sr=float(sys.argv[12]) #resample audio in post processing + # rms_mix_rate=float(sys.argv[13]) #search feature print(sys.argv) + class Config: - def __init__(self,device,is_half): + def __init__(self, device, is_half): self.device = device self.is_half = is_half self.n_cpu = 0 @@ -113,8 +115,9 @@ class Config: return x_pad, x_query, x_center, x_max -config=Config(device,is_half) -now_dir=os.getcwd() + +config = Config(device, is_half) +now_dir = os.getcwd() sys.path.append(now_dir) from vc_infer_pipeline import VC from infer_pack.models import SynthesizerTrnMs256NSFsid, SynthesizerTrnMs256NSFsid_nono @@ -122,7 +125,9 @@ from my_utils import load_audio from fairseq import checkpoint_utils from scipy.io import wavfile -hubert_model=None +hubert_model = None + + def load_hubert(): global hubert_model models, _, _ = checkpoint_utils.load_model_ensemble_and_task( @@ -137,41 +142,42 @@ def load_hubert(): hubert_model = hubert_model.float() hubert_model.eval() + def vc_single( sid=0, input_audio_path=None, - f0_up_key=0, + f0_up_key=0, f0_file=None, - f0_method="pm", - file_index="", #.index file + f0_method="pm", + file_index="", # .index file file_index2="", # file_big_npy, - index_rate=1.0, - filter_radius=3, - resample_sr=0, - rms_mix_rate=1.0, + index_rate=1.0, + filter_radius=3, + resample_sr=0, + rms_mix_rate=1.0, model_path="", output_path="", - protect=0.33 + protect=0.33, ): global tgt_sr, net_g, vc, hubert_model, version get_vc(model_path) if input_audio_path is None: return "You need to upload an audio file", None - + f0_up_key = int(f0_up_key) audio = load_audio(input_audio_path, 16000) audio_max = np.abs(audio).max() / 0.95 - + if audio_max > 1: audio /= audio_max times = [0, 0, 0] - + if hubert_model == None: load_hubert() - + if_f0 = cpt.get("f0", 1) - + file_index = ( ( file_index.strip(" ") @@ -184,7 +190,7 @@ def vc_single( if file_index != "" else file_index2 ) - + audio_opt = vc.pipeline( hubert_model, net_g, @@ -204,32 +210,49 @@ def vc_single( rms_mix_rate, version, f0_file=f0_file, - protect=protect + protect=protect, ) wavfile.write(output_path, tgt_sr, audio_opt) - return('processed') + return "processed" def get_vc(model_path): - global n_spk,tgt_sr,net_g,vc,cpt,device,is_half, version - print("loading pth %s"%model_path) + global n_spk, tgt_sr, net_g, vc, cpt, device, is_half, version + print("loading pth %s" % model_path) cpt = torch.load(model_path, map_location="cpu") tgt_sr = cpt["config"][-1] - cpt["config"][-3]=cpt["weight"]["emb_g.weight"].shape[0]#n_spk - if_f0=cpt.get("f0",1) + cpt["config"][-3] = cpt["weight"]["emb_g.weight"].shape[0] # n_spk + if_f0 = cpt.get("f0", 1) version = cpt.get("version", "v1") - if(if_f0==1): + if if_f0 == 1: net_g = SynthesizerTrnMs256NSFsid(*cpt["config"], is_half=is_half) else: net_g = SynthesizerTrnMs256NSFsid_nono(*cpt["config"]) del net_g.enc_q print(net_g.load_state_dict(cpt["weight"], strict=False)) net_g.eval().to(device) - if (is_half):net_g = net_g.half() - else:net_g = net_g.float() + if is_half: + net_g = net_g.half() + else: + net_g = net_g.float() vc = VC(tgt_sr, config) - n_spk=cpt["config"][-3] + n_spk = cpt["config"][-3] # return {"visible": True,"maximum": n_spk, "__type__": "update"} -if(using_cli): - vc_single(sid=0,input_audio_path=input_path,f0_up_key=f0_up_key,f0_file=None,f0_method=f0_method,file_index=file_index,file_index2="",index_rate=1,filter_radius=3,resample_sr=0,rms_mix_rate=0,model_path=model_path,output_path=output_path) + +if using_cli: + vc_single( + sid=0, + input_audio_path=input_path, + f0_up_key=f0_up_key, + f0_file=None, + f0_method=f0_method, + file_index=file_index, + file_index2="", + index_rate=1, + filter_radius=3, + resample_sr=0, + rms_mix_rate=0, + model_path=model_path, + output_path=output_path, + ) diff --git a/train_nsf_sim_cache_sid_load_pretrain.py b/train_nsf_sim_cache_sid_load_pretrain.py index dafc44e..9fe5735 100644 --- a/train_nsf_sim_cache_sid_load_pretrain.py +++ b/train_nsf_sim_cache_sid_load_pretrain.py @@ -1,4 +1,5 @@ -import os,sys +import os, sys + now_dir = os.getcwd() sys.path.append(os.path.join(now_dir)) sys.path.append(os.path.join(now_dir, "train"))