diff --git a/AutoCoverTool/ref/so_vits_svc/inference/infer_tool.py b/AutoCoverTool/ref/so_vits_svc/inference/infer_tool.py index 2bfa86d..2f29b6f 100644 --- a/AutoCoverTool/ref/so_vits_svc/inference/infer_tool.py +++ b/AutoCoverTool/ref/so_vits_svc/inference/infer_tool.py @@ -1,327 +1,340 @@ import hashlib import json import logging import os import time from pathlib import Path import librosa import maad import numpy as np # import onnxruntime import parselmouth import soundfile import torch import torchaudio from hubert import hubert_model import utils from models import SynthesizerTrn - +import copy logging.getLogger('matplotlib').setLevel(logging.WARNING) +from mel_processing import spectrogram_torch, spec_to_mel_torch + + +def get_spec(audio): + audio_norm = audio + print(audio_norm) + spec = spectrogram_torch(audio_norm, 1280, 32000, 320, 1280, center=False) + return spec def read_temp(file_name): if not os.path.exists(file_name): with open(file_name, "w") as f: f.write(json.dumps({"info": "temp_dict"})) return {} else: try: with open(file_name, "r") as f: data = f.read() data_dict = json.loads(data) if os.path.getsize(file_name) > 50 * 1024 * 1024: f_name = file_name.replace("\\", "/").split("/")[-1] print(f"clean {f_name}") for wav_hash in list(data_dict.keys()): if int(time.time()) - int(data_dict[wav_hash]["time"]) > 14 * 24 * 3600: del data_dict[wav_hash] except Exception as e: print(e) print(f"{file_name} error,auto rebuild file") data_dict = {"info": "temp_dict"} return data_dict def write_temp(file_name, data): with open(file_name, "w") as f: f.write(json.dumps(data)) def timeit(func): def run(*args, **kwargs): t = time.time() res = func(*args, **kwargs) print('executing \'%s\' costed %.3fs' % (func.__name__, time.time() - t)) return res return run def format_wav(audio_path): if Path(audio_path).suffix == '.wav': return raw_audio, raw_sample_rate = librosa.load(audio_path, mono=True, sr=None) soundfile.write(Path(audio_path).with_suffix(".wav"), raw_audio, raw_sample_rate) def get_end_file(dir_path, end): file_lists = [] for root, dirs, files in os.walk(dir_path): files = [f for f in files if f[0] != '.'] dirs[:] = [d for d in dirs if d[0] != '.'] for f_file in files: if f_file.endswith(end): file_lists.append(os.path.join(root, f_file).replace("\\", "/")) return file_lists def get_md5(content): return hashlib.new("md5", content).hexdigest() def resize2d_f0(x, target_len): source = np.array(x) source[source < 0.001] = np.nan target = np.interp(np.arange(0, len(source) * target_len, len(source)) / target_len, np.arange(0, len(source)), source) res = np.nan_to_num(target) return res def get_f0(x, p_len, f0_up_key=0): time_step = 160 / 16000 * 1000 f0_min = 50 f0_max = 1100 f0_mel_min = 1127 * np.log(1 + f0_min / 700) f0_mel_max = 1127 * np.log(1 + f0_max / 700) f0 = parselmouth.Sound(x, 16000).to_pitch_ac( time_step=time_step / 1000, voicing_threshold=0.6, pitch_floor=f0_min, pitch_ceiling=f0_max).selected_array['frequency'] if len(f0) > p_len: f0 = f0[:p_len] pad_size = (p_len - len(f0) + 1) // 2 if (pad_size > 0 or p_len - len(f0) - pad_size > 0): f0 = np.pad(f0, [[pad_size, p_len - len(f0) - pad_size]], mode='constant') f0 *= pow(2, f0_up_key / 12) f0_mel = 1127 * np.log(1 + f0 / 700) f0_mel[f0_mel > 0] = (f0_mel[f0_mel > 0] - f0_mel_min) * 254 / (f0_mel_max - f0_mel_min) + 1 f0_mel[f0_mel <= 1] = 1 f0_mel[f0_mel > 255] = 255 f0_coarse = np.rint(f0_mel).astype(np.int) return f0_coarse, f0 def clean_pitch(input_pitch): num_nan = np.sum(input_pitch == 1) if num_nan / len(input_pitch) > 0.9: input_pitch[input_pitch != 1] = 1 return input_pitch def plt_pitch(input_pitch): input_pitch = input_pitch.astype(float) input_pitch[input_pitch == 1] = np.nan return input_pitch def f0_to_pitch(ff): f0_pitch = 69 + 12 * np.log2(ff / 440) return f0_pitch def fill_a_to_b(a, b): if len(a) < len(b): for _ in range(0, len(b) - len(a)): a.append(a[0]) def mkdir(paths: list): for path in paths: if not os.path.exists(path): os.mkdir(path) class Svc(object): def __init__(self, net_g_path, config_path, hubert_path="data/models/hubert-soft-0d54a1f4.pt", onnx=False): self.onnx = onnx self.net_g_path = net_g_path self.hubert_path = hubert_path self.dev = torch.device("cuda" if torch.cuda.is_available() else "cpu") self.net_g_ms = None self.hps_ms = utils.get_hparams_from_file(config_path) self.target_sample = self.hps_ms.data.sampling_rate self.hop_size = self.hps_ms.data.hop_length self.speakers = {} for spk, sid in self.hps_ms.spk.items(): self.speakers[sid] = spk self.spk2id = self.hps_ms.spk # 加载hubert self.hubert_soft = hubert_model.hubert_soft(hubert_path) if torch.cuda.is_available(): self.hubert_soft = self.hubert_soft.cuda() self.load_model() def load_model(self): # 获取模型配置 if self.onnx: raise NotImplementedError # self.net_g_ms = SynthesizerTrnForONNX( # 178, # self.hps_ms.data.filter_length // 2 + 1, # self.hps_ms.train.segment_size // self.hps_ms.data.hop_length, # n_speakers=self.hps_ms.data.n_speakers, # **self.hps_ms.model) # _ = utils.load_checkpoint(self.net_g_path, self.net_g_ms, None) else: self.net_g_ms = SynthesizerTrn( self.hps_ms.data.filter_length // 2 + 1, self.hps_ms.train.segment_size // self.hps_ms.data.hop_length, **self.hps_ms.model) _ = utils.load_checkpoint(self.net_g_path, self.net_g_ms, None) if "half" in self.net_g_path and torch.cuda.is_available(): _ = self.net_g_ms.half().eval().to(self.dev) else: _ = self.net_g_ms.eval().to(self.dev) def get_units(self, source, sr): source = source.unsqueeze(0).to(self.dev) with torch.inference_mode(): start = time.time() units = self.hubert_soft.units(source) use_time = time.time() - start print("hubert use time:{}".format(use_time)) return units def get_unit_pitch(self, in_path, tran): source, sr = torchaudio.load(in_path) + source_bak = copy.deepcopy(source) source = torchaudio.functional.resample(source, sr, 16000) if len(source.shape) == 2 and source.shape[1] >= 2: source = torch.mean(source, dim=0).unsqueeze(0) soft = self.get_units(source, sr).squeeze(0).cpu().numpy() f0_coarse, f0 = get_f0(source.cpu().numpy()[0], soft.shape[0] * 2, tran) - return soft, f0 + return soft, f0, source_bak def infer(self, speaker_id, tran, raw_path, dev=False): if type(speaker_id) == str: speaker_id = self.spk2id[speaker_id] sid = torch.LongTensor([int(speaker_id)]).to(self.dev).unsqueeze(0) - soft, pitch = self.get_unit_pitch(raw_path, tran) + soft, pitch, source = self.get_unit_pitch(raw_path, tran) f0 = torch.FloatTensor(clean_pitch(pitch)).unsqueeze(0).to(self.dev) if "half" in self.net_g_path and torch.cuda.is_available(): stn_tst = torch.HalfTensor(soft) else: stn_tst = torch.FloatTensor(soft) + + # 提取幅度谱 + # spec = get_spec(source).to(self.dev) with torch.no_grad(): x_tst = stn_tst.unsqueeze(0).to(self.dev) start = time.time() x_tst = torch.repeat_interleave(x_tst, repeats=2, dim=1).transpose(1, 2) audio = self.net_g_ms.infer(x_tst, f0=f0, g=sid)[0, 0].data.float() + # audio = self.net_g_ms.infer_v1(x_tst, spec[:, :, :f0.size(-1)], f0=f0, g=sid)[0, 0].data.float() use_time = time.time() - start print("vits use time:{}".format(use_time)) return audio, audio.shape[-1] # class SvcONNXInferModel(object): # def __init__(self, hubert_onnx, vits_onnx, config_path): # self.config_path = config_path # self.vits_onnx = vits_onnx # self.hubert_onnx = hubert_onnx # self.hubert_onnx_session = onnxruntime.InferenceSession(hubert_onnx, providers=['CUDAExecutionProvider', ]) # self.inspect_onnx(self.hubert_onnx_session) # self.vits_onnx_session = onnxruntime.InferenceSession(vits_onnx, providers=['CUDAExecutionProvider', ]) # self.inspect_onnx(self.vits_onnx_session) # self.hps_ms = utils.get_hparams_from_file(self.config_path) # self.target_sample = self.hps_ms.data.sampling_rate # self.feature_input = FeatureInput(self.hps_ms.data.sampling_rate, self.hps_ms.data.hop_length) # # @staticmethod # def inspect_onnx(session): # for i in session.get_inputs(): # print("name:{}\tshape:{}\tdtype:{}".format(i.name, i.shape, i.type)) # for i in session.get_outputs(): # print("name:{}\tshape:{}\tdtype:{}".format(i.name, i.shape, i.type)) # # def infer(self, speaker_id, tran, raw_path): # sid = np.array([int(speaker_id)], dtype=np.int64) # soft, pitch = self.get_unit_pitch(raw_path, tran) # pitch = np.expand_dims(pitch, axis=0).astype(np.int64) # stn_tst = soft # x_tst = np.expand_dims(stn_tst, axis=0) # x_tst_lengths = np.array([stn_tst.shape[0]], dtype=np.int64) # # 使用ONNX Runtime进行推理 # start = time.time() # audio = self.vits_onnx_session.run(output_names=["audio"], # input_feed={ # "hidden_unit": x_tst, # "lengths": x_tst_lengths, # "pitch": pitch, # "sid": sid, # })[0][0, 0] # use_time = time.time() - start # print("vits_onnx_session.run time:{}".format(use_time)) # audio = torch.from_numpy(audio) # return audio, audio.shape[-1] # # def get_units(self, source, sr): # source = torchaudio.functional.resample(source, sr, 16000) # if len(source.shape) == 2 and source.shape[1] >= 2: # source = torch.mean(source, dim=0).unsqueeze(0) # source = source.unsqueeze(0) # # 使用ONNX Runtime进行推理 # start = time.time() # units = self.hubert_onnx_session.run(output_names=["embed"], # input_feed={"source": source.numpy()})[0] # use_time = time.time() - start # print("hubert_onnx_session.run time:{}".format(use_time)) # return units # # def transcribe(self, source, sr, length, transform): # feature_pit = self.feature_input.compute_f0(source, sr) # feature_pit = feature_pit * 2 ** (transform / 12) # feature_pit = resize2d_f0(feature_pit, length) # coarse_pit = self.feature_input.coarse_f0(feature_pit) # return coarse_pit # # def get_unit_pitch(self, in_path, tran): # source, sr = torchaudio.load(in_path) # soft = self.get_units(source, sr).squeeze(0) # input_pitch = self.transcribe(source.numpy()[0], sr, soft.shape[0], tran) # return soft, input_pitch class RealTimeVC: def __init__(self): self.last_chunk = None self.last_o = None self.chunk_len = 16000 # 区块长度 self.pre_len = 3840 # 交叉淡化长度，640的倍数 """输入输出都是1维numpy 音频波形数组""" def process(self, svc_model, speaker_id, f_pitch_change, input_wav_path): audio, sr = torchaudio.load(input_wav_path) audio = audio.cpu().numpy()[0] temp_wav = io.BytesIO() if self.last_chunk is None: input_wav_path.seek(0) audio, sr = svc_model.infer(speaker_id, f_pitch_change, input_wav_path) audio = audio.cpu().numpy() self.last_chunk = audio[-self.pre_len:] self.last_o = audio return audio[-self.chunk_len:] else: audio = np.concatenate([self.last_chunk, audio]) soundfile.write(temp_wav, audio, sr, format="wav") temp_wav.seek(0) audio, sr = svc_model.infer(speaker_id, f_pitch_change, temp_wav) audio = audio.cpu().numpy() ret = maad.util.crossfade(self.last_o, audio, self.pre_len) self.last_chunk = audio[-self.pre_len:] self.last_o = audio return ret[self.chunk_len:2 * self.chunk_len] diff --git a/AutoCoverTool/ref/so_vits_svc/inference_main.py b/AutoCoverTool/ref/so_vits_svc/inference_main.py index e1579ec..326ad07 100644 --- a/AutoCoverTool/ref/so_vits_svc/inference_main.py +++ b/AutoCoverTool/ref/so_vits_svc/inference_main.py @@ -1,83 +1,85 @@ import io import os import sys import logging import time from pathlib import Path +from copy import deepcopy +import torch import librosa import numpy as np import soundfile from inference import infer_tool from inference import slicer from inference.infer_tool import Svc logging.getLogger('numba').setLevel(logging.WARNING) chunks_dict = infer_tool.read_temp("ref/so_vits_svc/inference/chunks_temp.json") def inf(model_path, config_path, raw_audio_path, dst_path, dev): # model_path = "logs/32k/G_174000-Copy1.pth" # config_path = "configs/config.json" svc_model = Svc(model_path, config_path) out_dir = os.path.dirname(dst_path) print(dst_path) os.makedirs(out_dir, exist_ok=True) # 支持多个wav文件，放在raw文件夹下 tran = 0 spk_list = ['speaker0'] # 每次同时合成多语者音色 slice_db = -40 # 默认-40，嘈杂的音频可以-30，干声保留呼吸可以-50 wav_format = 'wav' # 音频输出格式 # infer_tool.fill_a_to_b(trans, clean_names) # for clean_name, tran in zip(clean_names, trans): # raw_audio_path = f"raw/{clean_name}" # if "." not in raw_audio_path: # raw_audio_path += ".wav" infer_tool.format_wav(raw_audio_path) wav_path = Path(raw_audio_path).with_suffix('.wav') chunks = slicer.cut(wav_path, db_thresh=slice_db) audio_data, audio_sr = slicer.chunks2audio(wav_path, chunks) for spk in spk_list: audio = [] for (slice_tag, data) in audio_data: print(f'#=====segment start, {round(len(data) / audio_sr, 3)}s======') length = int(np.ceil(len(data) / audio_sr * svc_model.target_sample)) raw_path = io.BytesIO() soundfile.write(raw_path, data, audio_sr, format="wav") raw_path.seek(0) if slice_tag: print('jump empty segment') _audio = np.zeros(length) else: out_audio, out_sr = svc_model.infer(spk, tran, raw_path, dev == "test") _audio = out_audio.cpu().numpy() audio.extend(list(_audio)) soundfile.write(dst_path, audio, svc_model.target_sample, format=wav_format) if __name__ == '__main__': g_model = sys.argv[1] # 模型地址 g_config = sys.argv[2] # 配置文件地址 g_audio_path = sys.argv[3] # 输入的音频文件地址,wav g_dst_path = sys.argv[4] # 输出的音频文件地址 if os.path.exists(g_dst_path): print("{} success ...".format(g_dst_path)) exit(0) g_dev = "prod" if len(sys.argv) > 5: g_dev = sys.argv[5] g_aa, g_sr = librosa.load(g_audio_path) d = librosa.get_duration(g_aa, g_sr) # if g_dev != "test": # if d > 250: # print("{} too long".format(g_audio_path)) # exit(0) st = time.time() inf(g_model, g_config, g_audio_path, g_dst_path, g_dev) print("{}, inference sp={}".format(g_audio_path, time.time() - st)) diff --git a/AutoCoverTool/ref/so_vits_svc/losses.py b/AutoCoverTool/ref/so_vits_svc/losses.py index 41f9be6..c314696 100644 --- a/AutoCoverTool/ref/so_vits_svc/losses.py +++ b/AutoCoverTool/ref/so_vits_svc/losses.py @@ -1,61 +1,62 @@ import torch from torch.nn import functional as F import commons def feature_loss(fmap_r, fmap_g): loss = 0 for dr, dg in zip(fmap_r, fmap_g): for rl, gl in zip(dr, dg): rl = rl.float().detach() gl = gl.float() loss += torch.mean(torch.abs(rl - gl)) return loss * 2 def discriminator_loss(disc_real_outputs, disc_generated_outputs): loss = 0 r_losses = [] g_losses = [] for dr, dg in zip(disc_real_outputs, disc_generated_outputs): dr = dr.float() dg = dg.float() r_loss = torch.mean((1-dr)**2) g_loss = torch.mean(dg**2) loss += (r_loss + g_loss) r_losses.append(r_loss.item()) g_losses.append(g_loss.item()) return loss, r_losses, g_losses def generator_loss(disc_outputs): loss = 0 gen_losses = [] for dg in disc_outputs: dg = dg.float() l = torch.mean((1-dg)**2) gen_losses.append(l) loss += l return loss, gen_losses def kl_loss(z_p, logs_q, m_p, logs_p, z_mask): """ z_p, logs_q: [b, h, t_t] m_p, logs_p: [b, h, t_t] """ z_p = z_p.float() logs_q = logs_q.float() m_p = m_p.float() logs_p = logs_p.float() z_mask = z_mask.float() + # kl = -0.5 * torch.sum(1 + log_var - mean ** 2 - torch.exp(log_var)) #print(logs_p) kl = logs_p - logs_q - 0.5 kl += 0.5 * ((z_p - m_p)**2) * torch.exp(-2. * logs_p) kl = torch.sum(kl * z_mask) l = kl / torch.sum(z_mask) return l diff --git a/AutoCoverTool/ref/so_vits_svc/models.py b/AutoCoverTool/ref/so_vits_svc/models.py index 3e3498b..6fdea53 100644 --- a/AutoCoverTool/ref/so_vits_svc/models.py +++ b/AutoCoverTool/ref/so_vits_svc/models.py @@ -1,357 +1,388 @@ import copy import math import torch from torch import nn from torch.nn import functional as F import attentions import commons import modules from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d from torch.nn.utils import weight_norm, remove_weight_norm, spectral_norm from commons import init_weights, get_padding from vdecoder.hifigan.models import Generator from utils import f0_to_coarse class ResidualCouplingBlock(nn.Module): def __init__(self, channels, hidden_channels, kernel_size, dilation_rate, n_layers, n_flows=4, gin_channels=0): super().__init__() self.channels = channels self.hidden_channels = hidden_channels self.kernel_size = kernel_size self.dilation_rate = dilation_rate self.n_layers = n_layers self.n_flows = n_flows self.gin_channels = gin_channels self.flows = nn.ModuleList() for i in range(n_flows): self.flows.append(modules.ResidualCouplingLayer(channels, hidden_channels, kernel_size, dilation_rate, n_layers, gin_channels=gin_channels, mean_only=True)) self.flows.append(modules.Flip()) def forward(self, x, x_mask, g=None, reverse=False): if not reverse: for flow in self.flows: x, _ = flow(x, x_mask, g=g, reverse=reverse) else: for flow in reversed(self.flows): x = flow(x, x_mask, g=g, reverse=reverse) return x class Encoder(nn.Module): def __init__(self, in_channels, out_channels, hidden_channels, kernel_size, dilation_rate, n_layers, gin_channels=0): super().__init__() self.in_channels = in_channels self.out_channels = out_channels self.hidden_channels = hidden_channels self.kernel_size = kernel_size self.dilation_rate = dilation_rate self.n_layers = n_layers self.gin_channels = gin_channels self.pre = nn.Conv1d(in_channels, hidden_channels, 1) self.enc = modules.WN(hidden_channels, kernel_size, dilation_rate, n_layers, gin_channels=gin_channels) self.proj = nn.Conv1d(hidden_channels, out_channels * 2, 1) def forward(self, x, x_lengths, g=None): # print(x.shape,x_lengths.shape) + # commons.sequence_mask 对于batch层级有价值，x_lengths是每个batch中每一个元素的帧数 + # 比如输入（[3,5,2]， 5）那么得到 3 * 5的True/False矩阵，其中第一层矩阵为3个true,2个false,第二层全true，第三层前两个true,其余false + # 作用一个批次中允许不同长度的数据一起训练，此时较短的乘以false，剔除影响 x_mask = torch.unsqueeze(commons.sequence_mask(x_lengths, x.size(2)), 1).to(x.dtype) x = self.pre(x) * x_mask x = self.enc(x, x_mask, g=g) stats = self.proj(x) * x_mask m, logs = torch.split(stats, self.out_channels, dim=1) z = (m + torch.randn_like(m) * torch.exp(logs)) * x_mask return z, m, logs, x_mask class TextEncoder(nn.Module): def __init__(self, in_channels, out_channels, hidden_channels, kernel_size, dilation_rate, n_layers, gin_channels=0, filter_channels=None, n_heads=None, p_dropout=None): super().__init__() self.in_channels = in_channels self.out_channels = out_channels self.hidden_channels = hidden_channels self.kernel_size = kernel_size self.dilation_rate = dilation_rate self.n_layers = n_layers self.gin_channels = gin_channels self.pre = nn.Conv1d(in_channels, hidden_channels, 1) self.proj = nn.Conv1d(hidden_channels, out_channels * 2, 1) self.f0_emb = nn.Embedding(256, hidden_channels) self.enc_ = attentions.Encoder( hidden_channels, filter_channels, n_heads, n_layers, kernel_size, p_dropout) def forward(self, x, x_lengths, f0=None): # x->(b,256,frame_num), x_lengths -> (b) # commons.sequence_mask 对于batch层级有价值，x_lengths是每个batch中每一个元素的帧数 # 比如输入（[3,5,2]， 5）那么得到 3 * 5的True/False矩阵，其中第一层矩阵为3个true,2个false,第二层全true，第三层前两个true,其余false # 作用一个批次中允许不同长度的数据一起训练，此时较短的乘以false，剔除影响 x_mask = torch.unsqueeze(commons.sequence_mask(x_lengths, x.size(2)), 1).to(x.dtype) x = self.pre(x) * x_mask x = x + self.f0_emb(f0).transpose(1,2) x = self.enc_(x * x_mask, x_mask) stats = self.proj(x) * x_mask + # m是VAE过程中得到的mu,而log对应的是log(sigma) m, logs = torch.split(stats, self.out_channels, dim=1) + # z是随机采样过程 z = (m + torch.randn_like(m) * torch.exp(logs)) * x_mask return z, m, logs, x_mask class DiscriminatorP(torch.nn.Module): def __init__(self, period, kernel_size=5, stride=3, use_spectral_norm=False): super(DiscriminatorP, self).__init__() self.period = period self.use_spectral_norm = use_spectral_norm norm_f = weight_norm if use_spectral_norm == False else spectral_norm self.convs = nn.ModuleList([ norm_f(Conv2d(1, 32, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))), norm_f(Conv2d(32, 128, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))), norm_f(Conv2d(128, 512, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))), norm_f(Conv2d(512, 1024, (kernel_size, 1), (stride, 1), padding=(get_padding(kernel_size, 1), 0))), norm_f(Conv2d(1024, 1024, (kernel_size, 1), 1, padding=(get_padding(kernel_size, 1), 0))), ]) self.conv_post = norm_f(Conv2d(1024, 1, (3, 1), 1, padding=(1, 0))) def forward(self, x): fmap = [] # 1d to 2d b, c, t = x.shape if t % self.period != 0: # pad first n_pad = self.period - (t % self.period) x = F.pad(x, (0, n_pad), "reflect") t = t + n_pad x = x.view(b, c, t // self.period, self.period) for l in self.convs: x = l(x) x = F.leaky_relu(x, modules.LRELU_SLOPE) fmap.append(x) x = self.conv_post(x) fmap.append(x) x = torch.flatten(x, 1, -1) return x, fmap class DiscriminatorS(torch.nn.Module): def __init__(self, use_spectral_norm=False): super(DiscriminatorS, self).__init__() norm_f = weight_norm if use_spectral_norm == False else spectral_norm self.convs = nn.ModuleList([ norm_f(Conv1d(1, 16, 15, 1, padding=7)), norm_f(Conv1d(16, 64, 41, 4, groups=4, padding=20)), norm_f(Conv1d(64, 256, 41, 4, groups=16, padding=20)), norm_f(Conv1d(256, 1024, 41, 4, groups=64, padding=20)), norm_f(Conv1d(1024, 1024, 41, 4, groups=256, padding=20)), norm_f(Conv1d(1024, 1024, 5, 1, padding=2)), ]) self.conv_post = norm_f(Conv1d(1024, 1, 3, 1, padding=1)) def forward(self, x): fmap = [] for l in self.convs: x = l(x) x = F.leaky_relu(x, modules.LRELU_SLOPE) fmap.append(x) x = self.conv_post(x) fmap.append(x) x = torch.flatten(x, 1, -1) return x, fmap class MultiPeriodDiscriminator(torch.nn.Module): def __init__(self, use_spectral_norm=False): super(MultiPeriodDiscriminator, self).__init__() periods = [2,3,5,7,11] discs = [DiscriminatorS(use_spectral_norm=use_spectral_norm)] discs = discs + [DiscriminatorP(i, use_spectral_norm=use_spectral_norm) for i in periods] self.discriminators = nn.ModuleList(discs) def forward(self, y, y_hat): y_d_rs = [] y_d_gs = [] fmap_rs = [] fmap_gs = [] for i, d in enumerate(self.discriminators): y_d_r, fmap_r = d(y) y_d_g, fmap_g = d(y_hat) y_d_rs.append(y_d_r) y_d_gs.append(y_d_g) fmap_rs.append(fmap_r) fmap_gs.append(fmap_g) return y_d_rs, y_d_gs, fmap_rs, fmap_gs class SpeakerEncoder(torch.nn.Module): def __init__(self, mel_n_channels=80, model_num_layers=3, model_hidden_size=256, model_embedding_size=256): super(SpeakerEncoder, self).__init__() self.lstm = nn.LSTM(mel_n_channels, model_hidden_size, model_num_layers, batch_first=True) self.linear = nn.Linear(model_hidden_size, model_embedding_size) self.relu = nn.ReLU() def forward(self, mels): self.lstm.flatten_parameters() _, (hidden, _) = self.lstm(mels) embeds_raw = self.relu(self.linear(hidden[-1])) return embeds_raw / torch.norm(embeds_raw, dim=1, keepdim=True) def compute_partial_slices(self, total_frames, partial_frames, partial_hop): mel_slices = [] for i in range(0, total_frames-partial_frames, partial_hop): mel_range = torch.arange(i, i+partial_frames) mel_slices.append(mel_range) return mel_slices def embed_utterance(self, mel, partial_frames=128, partial_hop=64): mel_len = mel.size(1) last_mel = mel[:,-partial_frames:] if mel_len > partial_frames: mel_slices = self.compute_partial_slices(mel_len, partial_frames, partial_hop) mels = list(mel[:,s] for s in mel_slices) mels.append(last_mel) mels = torch.stack(tuple(mels), 0).squeeze(1) with torch.no_grad(): partial_embeds = self(mels) embed = torch.mean(partial_embeds, axis=0).unsqueeze(0) #embed = embed / torch.linalg.norm(embed, 2) else: with torch.no_grad(): embed = self(last_mel) return embed class SynthesizerTrn(nn.Module): """ Synthesizer for Training """ def __init__(self, spec_channels, segment_size, inter_channels, hidden_channels, filter_channels, n_heads, n_layers, kernel_size, p_dropout, resblock, resblock_kernel_sizes, resblock_dilation_sizes, upsample_rates, upsample_initial_channel, upsample_kernel_sizes, gin_channels, ssl_dim, n_speakers, **kwargs): super().__init__() self.spec_channels = spec_channels self.inter_channels = inter_channels self.hidden_channels = hidden_channels self.filter_channels = filter_channels self.n_heads = n_heads self.n_layers = n_layers self.kernel_size = kernel_size self.p_dropout = p_dropout self.resblock = resblock self.resblock_kernel_sizes = resblock_kernel_sizes self.resblock_dilation_sizes = resblock_dilation_sizes self.upsample_rates = upsample_rates self.upsample_initial_channel = upsample_initial_channel self.upsample_kernel_sizes = upsample_kernel_sizes self.segment_size = segment_size self.gin_channels = gin_channels self.ssl_dim = ssl_dim self.emb_g = nn.Embedding(n_speakers, gin_channels) self.enc_p_ = TextEncoder(ssl_dim, inter_channels, hidden_channels, 5, 1, 16,0, filter_channels, n_heads, p_dropout) hps = { "sampling_rate": 32000, "inter_channels": 192, "resblock": "1", "resblock_kernel_sizes": [3, 7, 11], "resblock_dilation_sizes": [[1, 3, 5], [1, 3, 5], [1, 3, 5]], "upsample_rates": [10, 8, 2, 2], "upsample_initial_channel": 512, "upsample_kernel_sizes": [16, 16, 4, 4], "gin_channels": 256, } self.dec = Generator(h=hps) self.enc_q = Encoder(spec_channels, inter_channels, hidden_channels, 5, 1, 16, gin_channels=gin_channels) self.flow = ResidualCouplingBlock(inter_channels, hidden_channels, 5, 1, 4, gin_channels=gin_channels) def forward(self, c, f0, spec, g=None, mel=None, c_lengths=None, spec_lengths=None): # hubert特征(b,256,frame_num), f0 (frame_num), 幅度谱特征, 说话人id,mel谱特征 if c_lengths == None: c_lengths = (torch.ones(c.size(0)) * c.size(-1)).to(c.device) # (b, frame_num) if spec_lengths == None: spec_lengths = (torch.ones(spec.size(0)) * spec.size(-1)).to(spec.device) # 说话人信息embding g = self.emb_g(g).transpose(1,2) + # 采样得到的z,vae需要的均值，logs_p是vae需要的log(sigma) + # 输入hubert特征(b,256,frame_num), f0 (frame_num)，对应的是文本出隐变量的那段模型 z_ptemp, m_p, logs_p, _ = self.enc_p_(c, c_lengths, f0=f0_to_coarse(f0)) + + # 输入幅度谱和说话人信息 + # 输出采样得到的z，m_q是均值，logs_q是log(sigma) z, m_q, logs_q, spec_mask = self.enc_q(spec, spec_lengths, g=g) + # 标准化流，增加分布复杂程度 z_p = self.flow(z, spec_mask, g=g) + # 由于整个batch中含有的音频帧数不一致，要求每一个元素都随机裁剪出segment_size长度的特征 + # 返回z的batch列表，pitch_slice列表和ids_slice的列表 z_slice, pitch_slice, ids_slice = commons.rand_slice_segments_with_pitch(z, f0, spec_lengths, self.segment_size) # o = self.dec(z_slice, g=g) + # 解码部分，输入未经过标准化的z,以及说话人信息和pitch，得到wav波形 o = self.dec(z_slice, g=g, f0=pitch_slice) + # 原始波形,批次中每个采样到的帧的位置,批次中幅度谱的有效帧位置， + # 幅度谱编码得到正态分布后随机采样得到的z, z经过标准化流之后得到z_p, hubert特征层得到的正态分布的均值， + # hubert特征层得到的正态分布的标准差(logs_p)，幅度谱和人声信息得到的均值(m_q),幅度谱和人声信息得到的标准差(logs_q) return o, ids_slice, spec_mask, (z, z_p, m_p, logs_p, m_q, logs_q) def infer(self, c, f0, g=None, mel=None, c_lengths=None): if c_lengths == None: c_lengths = (torch.ones(c.size(0)) * c.size(-1)).to(c.device) g = self.emb_g(g).transpose(1,2) z_p, m_p, logs_p, c_mask = self.enc_p_(c, c_lengths, f0=f0_to_coarse(f0)) z = self.flow(z_p, c_mask, g=g, reverse=True) o = self.dec(z * c_mask, g=g, f0=f0) return o + + def infer_v1(self, c, spec, f0, g): + print(c.shape, spec.shape, f0.shape, g.shape) + c_lengths = (torch.ones(c.size(0)) * c.size(-1)).to(c.device) # (b, frame_num) + spec_lengths = (torch.ones(spec.size(0)) * spec.size(-1)).to(spec.device) + g = self.emb_g(g).transpose(1, 2) + + # z_p, m_p, logs_p, c_mask = self.enc_p_(c, c_lengths, f0=f0_to_coarse(f0)) + # z = self.flow(z_p, c_mask, g=g, reverse=True) + # o = self.dec(z * c_mask, g=g, f0=f0) + # print(c_mask.shape, c_mask) + z, m_q, logs_q, spec_mask = self.enc_q(spec, spec_lengths, g=g) + o = self.dec(z, g=g, f0=f0) + return o