first commit without model

.gitignore

@@ -0,0 +1 @@
+.DS_Store

README.md

@@ -1,12 +1,89 @@
----
-title: Vits Chinese
-emoji: 🏃
-colorFrom: red
-colorTo: yellow
-sdk: gradio
-sdk_version: 3.9
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+# This code is slightly modified on the [VITS repo](https://github.com/jaywalnut310/vits)
+## chinese train
+`python3 train.py -c configs/woman_csmsc.json -m woman_csmsc`
+
+## inference 
+you can run infer.py
+
+## switch english or chinese
+just modify 
+`chinese_mode = True`
+in ./text/__init__.py
+
+## data example
+copy **24000hz** data-baker datasets to ./test/csmsc
+
+**For copyright reasons, you can only download it yourself**
+
+## models
+you can get example model in ./logs/woman_csmsc/G*.pth
+
+## Prosody model
+you can try chinese Prosody model in this repo.
+
+
+
+======================================================================================================
+
+# VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech
+
+### Jaehyeon Kim, Jungil Kong, and Juhee Son
+
+In our recent [paper](https://arxiv.org/abs/2106.06103), we propose VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech.
+
+Several recent end-to-end text-to-speech (TTS) models enabling single-stage training and parallel sampling have been proposed, but their sample quality does not match that of two-stage TTS systems. In this work, we present a parallel end-to-end TTS method that generates more natural sounding audio than current two-stage models. Our method adopts variational inference augmented with normalizing flows and an adversarial training process, which improves the expressive power of generative modeling. We also propose a stochastic duration predictor to synthesize speech with diverse rhythms from input text. With the uncertainty modeling over latent variables and the stochastic duration predictor, our method expresses the natural one-to-many relationship in which a text input can be spoken in multiple ways with different pitches and rhythms. A subjective human evaluation (mean opinion score, or MOS) on the LJ Speech, a single speaker dataset, shows that our method outperforms the best publicly available TTS systems and achieves a MOS comparable to ground truth.
+
+Visit our [demo](https://jaywalnut310.github.io/vits-demo/index.html) for audio samples.
+
+We also provide the [pretrained models](https://drive.google.com/drive/folders/1ksarh-cJf3F5eKJjLVWY0X1j1qsQqiS2?usp=sharing).
+
+** Update note: Thanks to [Rishikesh (ऋषिकेश)](https://github.com/jaywalnut310/vits/issues/1), our interactive TTS demo is now available on [Colab Notebook](https://colab.research.google.com/drive/1CO61pZizDj7en71NQG_aqqKdGaA_SaBf?usp=sharing).
+
+<table style="width:100%">
+ <tr>
+ <th>VITS at training</th>
+ <th>VITS at inference</th>
+ </tr>
+ <tr>
+ <td><img src="resources/fig_1a.png" alt="VITS at training" height="400"></td>
+ <td><img src="resources/fig_1b.png" alt="VITS at inference" height="400"></td>
+ </tr>
+</table>
+
+
+## Pre-requisites
+0. Python >= 3.6
+0. Clone this repository
+0. Install python requirements. Please refer [requirements.txt](requirements.txt)
+ 1. You may need to install espeak first: `apt-get install espeak`
+0. Download datasets
+ 1. Download and extract the LJ Speech dataset, then rename or create a link to the dataset folder: `ln -s /path/to/LJSpeech-1.1/wavs DUMMY1`
+ 1. For mult-speaker setting, download and extract the VCTK dataset, and downsample wav files to 22050 Hz. Then rename or create a link to the dataset folder: `ln -s /path/to/VCTK-Corpus/downsampled_wavs DUMMY2`
+0. Build Monotonic Alignment Search and run preprocessing if you use your own datasets.
+```sh
+# Cython-version Monotonoic Alignment Search
+cd monotonic_align
+python setup.py build_ext --inplace
+
+# Preprocessing (g2p) for your own datasets. Preprocessed phonemes for LJ Speech and VCTK have been already provided.
+# python preprocess.py --text_index 1 --filelists filelists/ljs_audio_text_train_filelist.txt filelists/ljs_audio_text_val_filelist.txt filelists/ljs_audio_text_test_filelist.txt 
+# python preprocess.py --text_index 2 --filelists filelists/vctk_audio_sid_text_train_filelist.txt filelists/vctk_audio_sid_text_val_filelist.txt filelists/vctk_audio_sid_text_test_filelist.txt
+```
+
+
+## Training Exmaple
+```sh
+# LJ Speech
+python train.py -c configs/ljs_base.json -m ljs_base
+
+# VCTK
+python train_ms.py -c configs/vctk_base.json -m vctk_base
+
+
+
+```
+
+
+## Inference Example
+See [inference.ipynb](inference.ipynb)

app.py

@@ -0,0 +1,55 @@
+import matplotlib.pyplot as plt
+import gradio as gr
+import os
+os.system('cd monotonic_align && python setup.py build_ext --inplace && cd ..')
+import torch
+
+import commons
+import utils
+from models import SynthesizerTrn
+from text.symbols import symbols
+from text import text_to_sequence
+
+def get_text(text, hps):
+ text_norm = text_to_sequence(text, hps.data.text_cleaners)
+ if hps.data.add_blank:
+ text_norm = commons.intersperse(text_norm, 0)
+ text_norm = torch.LongTensor(text_norm)
+ return text_norm
+
+
+hps = utils.get_hparams_from_file("configs/config.json")
+
+net_g = SynthesizerTrn(
+ len(symbols),
+ hps.data.filter_length // 2 + 1,
+ hps.train.segment_size // hps.data.hop_length,
+ **hps.model)
+_ = net_g.eval()
+
+# _ = utils.load_checkpoint("logs/woman_csmsc/G_100000.pth", net_g, None) 
+_ = utils.load_checkpoint("logs/G_98000.pth", net_g, None) 
+
+def vc_fn(input):
+ stn_tst = get_text(input, hps)
+ with torch.no_grad():
+ x_tst = stn_tst.unsqueeze(0)
+ x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
+
+ # x_tst = stn_tst.cpu().unsqueeze(0)
+ # x_tst_lengths = torch.LongTensor([stn_tst.size(0)]).cpu()
+
+ audio = net_g.infer(x_tst, x_tst_lengths, noise_scale=.667, noise_scale_w=0.8, length_scale=1)[0][0,0].data.cpu().float().numpy()
+ sampling_rate = 44100
+ return (sampling_rate, audio)
+ 
+app = gr.Blocks()
+with app:
+ with gr.Tabs():
+ with gr.TabItem("Basic"):
+ vc_input = gr.Textbox(label="Input Message")
+ vc_submit = gr.Button("Convert", variant="primary")
+ vc_output = gr.Audio(label="Output Audio")
+ vc_submit.click(vc_fn, [ vc_input], [vc_output])
+
+ app.launch()

attentions.py

@@ -0,0 +1,303 @@
+import copy
+import math
+import numpy as np
+import torch
+from torch import nn
+from torch.nn import functional as F
+
+import commons
+import modules
+from modules import LayerNorm
+ 
+
+class Encoder(nn.Module):
+ def __init__(self, hidden_channels, filter_channels, n_heads, n_layers, kernel_size=1, p_dropout=0., window_size=4, **kwargs):
+ super().__init__()
+ 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.window_size = window_size
+
+ self.drop = nn.Dropout(p_dropout)
+ self.attn_layers = nn.ModuleList()
+ self.norm_layers_1 = nn.ModuleList()
+ self.ffn_layers = nn.ModuleList()
+ self.norm_layers_2 = nn.ModuleList()
+ for i in range(self.n_layers):
+ self.attn_layers.append(MultiHeadAttention(hidden_channels, hidden_channels, n_heads, p_dropout=p_dropout, window_size=window_size))
+ self.norm_layers_1.append(LayerNorm(hidden_channels))
+ self.ffn_layers.append(FFN(hidden_channels, hidden_channels, filter_channels, kernel_size, p_dropout=p_dropout))
+ self.norm_layers_2.append(LayerNorm(hidden_channels))
+
+ def forward(self, x, x_mask):
+ attn_mask = x_mask.unsqueeze(2) * x_mask.unsqueeze(-1)
+ x = x * x_mask
+ for i in range(self.n_layers):
+ y = self.attn_layers[i](x, x, attn_mask)
+ y = self.drop(y)
+ x = self.norm_layers_1[i](x + y)
+
+ y = self.ffn_layers[i](x, x_mask)
+ y = self.drop(y)
+ x = self.norm_layers_2[i](x + y)
+ x = x * x_mask
+ return x
+
+
+class Decoder(nn.Module):
+ def __init__(self, hidden_channels, filter_channels, n_heads, n_layers, kernel_size=1, p_dropout=0., proximal_bias=False, proximal_init=True, **kwargs):
+ super().__init__()
+ 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.proximal_bias = proximal_bias
+ self.proximal_init = proximal_init
+
+ self.drop = nn.Dropout(p_dropout)
+ self.self_attn_layers = nn.ModuleList()
+ self.norm_layers_0 = nn.ModuleList()
+ self.encdec_attn_layers = nn.ModuleList()
+ self.norm_layers_1 = nn.ModuleList()
+ self.ffn_layers = nn.ModuleList()
+ self.norm_layers_2 = nn.ModuleList()
+ for i in range(self.n_layers):
+ self.self_attn_layers.append(MultiHeadAttention(hidden_channels, hidden_channels, n_heads, p_dropout=p_dropout, proximal_bias=proximal_bias, proximal_init=proximal_init))
+ self.norm_layers_0.append(LayerNorm(hidden_channels))
+ self.encdec_attn_layers.append(MultiHeadAttention(hidden_channels, hidden_channels, n_heads, p_dropout=p_dropout))
+ self.norm_layers_1.append(LayerNorm(hidden_channels))
+ self.ffn_layers.append(FFN(hidden_channels, hidden_channels, filter_channels, kernel_size, p_dropout=p_dropout, causal=True))
+ self.norm_layers_2.append(LayerNorm(hidden_channels))
+
+ def forward(self, x, x_mask, h, h_mask):
+ """
+ x: decoder input
+ h: encoder output
+ """
+ self_attn_mask = commons.subsequent_mask(x_mask.size(2)).to(device=x.device, dtype=x.dtype)
+ encdec_attn_mask = h_mask.unsqueeze(2) * x_mask.unsqueeze(-1)
+ x = x * x_mask
+ for i in range(self.n_layers):
+ y = self.self_attn_layers[i](x, x, self_attn_mask)
+ y = self.drop(y)
+ x = self.norm_layers_0[i](x + y)
+
+ y = self.encdec_attn_layers[i](x, h, encdec_attn_mask)
+ y = self.drop(y)
+ x = self.norm_layers_1[i](x + y)
+ 
+ y = self.ffn_layers[i](x, x_mask)
+ y = self.drop(y)
+ x = self.norm_layers_2[i](x + y)
+ x = x * x_mask
+ return x
+
+
+class MultiHeadAttention(nn.Module):
+ def __init__(self, channels, out_channels, n_heads, p_dropout=0., window_size=None, heads_share=True, block_length=None, proximal_bias=False, proximal_init=False):
+ super().__init__()
+ assert channels % n_heads == 0
+
+ self.channels = channels
+ self.out_channels = out_channels
+ self.n_heads = n_heads
+ self.p_dropout = p_dropout
+ self.window_size = window_size
+ self.heads_share = heads_share
+ self.block_length = block_length
+ self.proximal_bias = proximal_bias
+ self.proximal_init = proximal_init
+ self.attn = None
+
+ self.k_channels = channels // n_heads
+ self.conv_q = nn.Conv1d(channels, channels, 1)
+ self.conv_k = nn.Conv1d(channels, channels, 1)
+ self.conv_v = nn.Conv1d(channels, channels, 1)
+ self.conv_o = nn.Conv1d(channels, out_channels, 1)
+ self.drop = nn.Dropout(p_dropout)
+
+ if window_size is not None:
+ n_heads_rel = 1 if heads_share else n_heads
+ rel_stddev = self.k_channels**-0.5
+ self.emb_rel_k = nn.Parameter(torch.randn(n_heads_rel, window_size * 2 + 1, self.k_channels) * rel_stddev)
+ self.emb_rel_v = nn.Parameter(torch.randn(n_heads_rel, window_size * 2 + 1, self.k_channels) * rel_stddev)
+
+ nn.init.xavier_uniform_(self.conv_q.weight)
+ nn.init.xavier_uniform_(self.conv_k.weight)
+ nn.init.xavier_uniform_(self.conv_v.weight)
+ if proximal_init:
+ with torch.no_grad():
+ self.conv_k.weight.copy_(self.conv_q.weight)
+ self.conv_k.bias.copy_(self.conv_q.bias)
+ 
+ def forward(self, x, c, attn_mask=None):
+ q = self.conv_q(x)
+ k = self.conv_k(c)
+ v = self.conv_v(c)
+ 
+ x, self.attn = self.attention(q, k, v, mask=attn_mask)
+
+ x = self.conv_o(x)
+ return x
+
+ def attention(self, query, key, value, mask=None):
+ # reshape [b, d, t] -> [b, n_h, t, d_k]
+ b, d, t_s, t_t = (*key.size(), query.size(2))
+ query = query.view(b, self.n_heads, self.k_channels, t_t).transpose(2, 3)
+ key = key.view(b, self.n_heads, self.k_channels, t_s).transpose(2, 3)
+ value = value.view(b, self.n_heads, self.k_channels, t_s).transpose(2, 3)
+
+ scores = torch.matmul(query / math.sqrt(self.k_channels), key.transpose(-2, -1))
+ if self.window_size is not None:
+ assert t_s == t_t, "Relative attention is only available for self-attention."
+ key_relative_embeddings = self._get_relative_embeddings(self.emb_rel_k, t_s)
+ rel_logits = self._matmul_with_relative_keys(query /math.sqrt(self.k_channels), key_relative_embeddings)
+ scores_local = self._relative_position_to_absolute_position(rel_logits)
+ scores = scores + scores_local
+ if self.proximal_bias:
+ assert t_s == t_t, "Proximal bias is only available for self-attention."
+ scores = scores + self._attention_bias_proximal(t_s).to(device=scores.device, dtype=scores.dtype)
+ if mask is not None:
+ scores = scores.masked_fill(mask == 0, -1e4)
+ if self.block_length is not None:
+ assert t_s == t_t, "Local attention is only available for self-attention."
+ block_mask = torch.ones_like(scores).triu(-self.block_length).tril(self.block_length)
+ scores = scores.masked_fill(block_mask == 0, -1e4)
+ p_attn = F.softmax(scores, dim=-1) # [b, n_h, t_t, t_s]
+ p_attn = self.drop(p_attn)
+ output = torch.matmul(p_attn, value)
+ if self.window_size is not None:
+ relative_weights = self._absolute_position_to_relative_position(p_attn)
+ value_relative_embeddings = self._get_relative_embeddings(self.emb_rel_v, t_s)
+ output = output + self._matmul_with_relative_values(relative_weights, value_relative_embeddings)
+ output = output.transpose(2, 3).contiguous().view(b, d, t_t) # [b, n_h, t_t, d_k] -> [b, d, t_t]
+ return output, p_attn
+
+ def _matmul_with_relative_values(self, x, y):
+ """
+ x: [b, h, l, m]
+ y: [h or 1, m, d]
+ ret: [b, h, l, d]
+ """
+ ret = torch.matmul(x, y.unsqueeze(0))
+ return ret
+
+ def _matmul_with_relative_keys(self, x, y):
+ """
+ x: [b, h, l, d]
+ y: [h or 1, m, d]
+ ret: [b, h, l, m]
+ """
+ ret = torch.matmul(x, y.unsqueeze(0).transpose(-2, -1))
+ return ret
+
+ def _get_relative_embeddings(self, relative_embeddings, length):
+ max_relative_position = 2 * self.window_size + 1
+ # Pad first before slice to avoid using cond ops.
+ pad_length = max(length - (self.window_size + 1), 0)
+ slice_start_position = max((self.window_size + 1) - length, 0)
+ slice_end_position = slice_start_position + 2 * length - 1
+ if pad_length > 0:
+ padded_relative_embeddings = F.pad(
+ relative_embeddings,
+ commons.convert_pad_shape([[0, 0], [pad_length, pad_length], [0, 0]]))
+ else:
+ padded_relative_embeddings = relative_embeddings
+ used_relative_embeddings = padded_relative_embeddings[:,slice_start_position:slice_end_position]
+ return used_relative_embeddings
+
+ def _relative_position_to_absolute_position(self, x):
+ """
+ x: [b, h, l, 2*l-1]
+ ret: [b, h, l, l]
+ """
+ batch, heads, length, _ = x.size()
+ # Concat columns of pad to shift from relative to absolute indexing.
+ x = F.pad(x, commons.convert_pad_shape([[0,0],[0,0],[0,0],[0,1]]))
+
+ # Concat extra elements so to add up to shape (len+1, 2*len-1).
+ x_flat = x.view([batch, heads, length * 2 * length])
+ x_flat = F.pad(x_flat, commons.convert_pad_shape([[0,0],[0,0],[0,length-1]]))
+
+ # Reshape and slice out the padded elements.
+ x_final = x_flat.view([batch, heads, length+1, 2*length-1])[:, :, :length, length-1:]
+ return x_final
+
+ def _absolute_position_to_relative_position(self, x):
+ """
+ x: [b, h, l, l]
+ ret: [b, h, l, 2*l-1]
+ """
+ batch, heads, length, _ = x.size()
+ # padd along column
+ x = F.pad(x, commons.convert_pad_shape([[0, 0], [0, 0], [0, 0], [0, length-1]]))
+ x_flat = x.view([batch, heads, length**2 + length*(length -1)])
+ # add 0's in the beginning that will skew the elements after reshape
+ x_flat = F.pad(x_flat, commons.convert_pad_shape([[0, 0], [0, 0], [length, 0]]))
+ x_final = x_flat.view([batch, heads, length, 2*length])[:,:,:,1:]
+ return x_final
+
+ def _attention_bias_proximal(self, length):
+ """Bias for self-attention to encourage attention to close positions.
+ Args:
+ length: an integer scalar.
+ Returns:
+ a Tensor with shape [1, 1, length, length]
+ """
+ r = torch.arange(length, dtype=torch.float32)
+ diff = torch.unsqueeze(r, 0) - torch.unsqueeze(r, 1)
+ return torch.unsqueeze(torch.unsqueeze(-torch.log1p(torch.abs(diff)), 0), 0)
+
+
+class FFN(nn.Module):
+ def __init__(self, in_channels, out_channels, filter_channels, kernel_size, p_dropout=0., activation=None, causal=False):
+ super().__init__()
+ self.in_channels = in_channels
+ self.out_channels = out_channels
+ self.filter_channels = filter_channels
+ self.kernel_size = kernel_size
+ self.p_dropout = p_dropout
+ self.activation = activation
+ self.causal = causal
+
+ if causal:
+ self.padding = self._causal_padding
+ else:
+ self.padding = self._same_padding
+
+ self.conv_1 = nn.Conv1d(in_channels, filter_channels, kernel_size)
+ self.conv_2 = nn.Conv1d(filter_channels, out_channels, kernel_size)
+ self.drop = nn.Dropout(p_dropout)
+
+ def forward(self, x, x_mask):
+ x = self.conv_1(self.padding(x * x_mask))
+ if self.activation == "gelu":
+ x = x * torch.sigmoid(1.702 * x)
+ else:
+ x = torch.relu(x)
+ x = self.drop(x)
+ x = self.conv_2(self.padding(x * x_mask))
+ return x * x_mask
+ 
+ def _causal_padding(self, x):
+ if self.kernel_size == 1:
+ return x
+ pad_l = self.kernel_size - 1
+ pad_r = 0
+ padding = [[0, 0], [0, 0], [pad_l, pad_r]]
+ x = F.pad(x, commons.convert_pad_shape(padding))
+ return x
+
+ def _same_padding(self, x):
+ if self.kernel_size == 1:
+ return x
+ pad_l = (self.kernel_size - 1) // 2
+ pad_r = self.kernel_size // 2
+ padding = [[0, 0], [0, 0], [pad_l, pad_r]]
+ x = F.pad(x, commons.convert_pad_shape(padding))
+ return x

commons.py

@@ -0,0 +1,161 @@
+import math
+import numpy as np
+import torch
+from torch import nn
+from torch.nn import functional as F
+
+
+def init_weights(m, mean=0.0, std=0.01):
+ classname = m.__class__.__name__
+ if classname.find("Conv") != -1:
+ m.weight.data.normal_(mean, std)
+
+
+def get_padding(kernel_size, dilation=1):
+ return int((kernel_size*dilation - dilation)/2)
+
+
+def convert_pad_shape(pad_shape):
+ l = pad_shape[::-1]
+ pad_shape = [item for sublist in l for item in sublist]
+ return pad_shape
+
+
+def intersperse(lst, item):
+ result = [item] * (len(lst) * 2 + 1)
+ result[1::2] = lst
+ return result
+
+
+def kl_divergence(m_p, logs_p, m_q, logs_q):
+ """KL(P||Q)"""
+ kl = (logs_q - logs_p) - 0.5
+ kl += 0.5 * (torch.exp(2. * logs_p) + ((m_p - m_q)**2)) * torch.exp(-2. * logs_q)
+ return kl
+
+
+def rand_gumbel(shape):
+ """Sample from the Gumbel distribution, protect from overflows."""
+ uniform_samples = torch.rand(shape) * 0.99998 + 0.00001
+ return -torch.log(-torch.log(uniform_samples))
+
+
+def rand_gumbel_like(x):
+ g = rand_gumbel(x.size()).to(dtype=x.dtype, device=x.device)
+ return g
+
+
+def slice_segments(x, ids_str, segment_size=4):
+ ret = torch.zeros_like(x[:, :, :segment_size])
+ for i in range(x.size(0)):
+ idx_str = ids_str[i]
+ idx_end = idx_str + segment_size
+ ret[i] = x[i, :, idx_str:idx_end]
+ return ret
+
+
+def rand_slice_segments(x, x_lengths=None, segment_size=4):
+ b, d, t = x.size()
+ if x_lengths is None:
+ x_lengths = t
+ ids_str_max = x_lengths - segment_size + 1
+ ids_str = (torch.rand([b]).to(device=x.device) * ids_str_max).to(dtype=torch.long)
+ ret = slice_segments(x, ids_str, segment_size)
+ return ret, ids_str
+
+
+def get_timing_signal_1d(
+ length, channels, min_timescale=1.0, max_timescale=1.0e4):
+ position = torch.arange(length, dtype=torch.float)
+ num_timescales = channels // 2
+ log_timescale_increment = (
+ math.log(float(max_timescale) / float(min_timescale)) /
+ (num_timescales - 1))
+ inv_timescales = min_timescale * torch.exp(
+ torch.arange(num_timescales, dtype=torch.float) * -log_timescale_increment)
+ scaled_time = position.unsqueeze(0) * inv_timescales.unsqueeze(1)
+ signal = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)], 0)
+ signal = F.pad(signal, [0, 0, 0, channels % 2])
+ signal = signal.view(1, channels, length)
+ return signal
+
+
+def add_timing_signal_1d(x, min_timescale=1.0, max_timescale=1.0e4):
+ b, channels, length = x.size()
+ signal = get_timing_signal_1d(length, channels, min_timescale, max_timescale)
+ return x + signal.to(dtype=x.dtype, device=x.device)
+
+
+def cat_timing_signal_1d(x, min_timescale=1.0, max_timescale=1.0e4, axis=1):
+ b, channels, length = x.size()
+ signal = get_timing_signal_1d(length, channels, min_timescale, max_timescale)
+ return torch.cat([x, signal.to(dtype=x.dtype, device=x.device)], axis)
+
+
+def subsequent_mask(length):
+ mask = torch.tril(torch.ones(length, length)).unsqueeze(0).unsqueeze(0)
+ return mask
+
+
+@torch.jit.script
+def fused_add_tanh_sigmoid_multiply(input_a, input_b, n_channels):
+ n_channels_int = n_channels[0]
+ in_act = input_a + input_b
+ t_act = torch.tanh(in_act[:, :n_channels_int, :])
+ s_act = torch.sigmoid(in_act[:, n_channels_int:, :])
+ acts = t_act * s_act
+ return acts
+
+
+def convert_pad_shape(pad_shape):
+ l = pad_shape[::-1]
+ pad_shape = [item for sublist in l for item in sublist]
+ return pad_shape
+
+
+def shift_1d(x):
+ x = F.pad(x, convert_pad_shape([[0, 0], [0, 0], [1, 0]]))[:, :, :-1]
+ return x
+
+
+def sequence_mask(length, max_length=None):
+ if max_length is None:
+ max_length = length.max()
+ x = torch.arange(max_length, dtype=length.dtype, device=length.device)
+ return x.unsqueeze(0) < length.unsqueeze(1)
+
+
+def generate_path(duration, mask):
+ """
+ duration: [b, 1, t_x]
+ mask: [b, 1, t_y, t_x]
+ """
+ device = duration.device
+ 
+ b, _, t_y, t_x = mask.shape
+ cum_duration = torch.cumsum(duration, -1)
+ 
+ cum_duration_flat = cum_duration.view(b * t_x)
+ path = sequence_mask(cum_duration_flat, t_y).to(mask.dtype)
+ path = path.view(b, t_x, t_y)
+ path = path - F.pad(path, convert_pad_shape([[0, 0], [1, 0], [0, 0]]))[:, :-1]
+ path = path.unsqueeze(1).transpose(2,3) * mask
+ return path
+
+
+def clip_grad_value_(parameters, clip_value, norm_type=2):
+ if isinstance(parameters, torch.Tensor):
+ parameters = [parameters]
+ parameters = list(filter(lambda p: p.grad is not None, parameters))
+ norm_type = float(norm_type)
+ if clip_value is not None:
+ clip_value = float(clip_value)
+
+ total_norm = 0
+ for p in parameters:
+ param_norm = p.grad.data.norm(norm_type)
+ total_norm += param_norm.item() ** norm_type
+ if clip_value is not None:
+ p.grad.data.clamp_(min=-clip_value, max=clip_value)
+ total_norm = total_norm ** (1. / norm_type)
+ return total_norm

configs/config.json

@@ -0,0 +1,52 @@
+{
+ "train": {
+ "log_interval": 200,
+ "eval_interval": 2000,
+ "seed": 1234,
+ "epochs": 20000,
+ "learning_rate": 2e-4,
+ "betas": [0.8, 0.99],
+ "eps": 1e-9,
+ "batch_size": 24,
+ "fp16_run": true,
+ "lr_decay": 0.999875,
+ "segment_size": 8192,
+ "init_lr_ratio": 1,
+ "warmup_epochs": 0,
+ "c_mel": 45,
+ "c_kl": 1.0
+ },
+ "data": {
+ "training_files":"test/dsq.txt.cleaned",
+ "validation_files":"test/dsq_val.txt.cleaned",
+ "text_cleaners":["chinese_cleaners1"],
+ "max_wav_value": 32768.0,
+ "sampling_rate": 44100,
+ "filter_length": 1024,
+ "hop_length": 256,
+ "win_length": 1024,
+ "n_mel_channels": 80,
+ "mel_fmin": 0.0,
+ "mel_fmax": null,
+ "add_blank": true,
+ "n_speakers": 0,
+ "cleaned_text": true
+ },
+ "model": {
+ "inter_channels": 192,
+ "hidden_channels": 192,
+ "filter_channels": 768,
+ "n_heads": 2,
+ "n_layers": 6,
+ "kernel_size": 3,
+ "p_dropout": 0.1,
+ "resblock": "1",
+ "resblock_kernel_sizes": [3,7,11],
+ "resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
+ "upsample_rates": [8,8,2,2],
+ "upsample_initial_channel": 512,
+ "upsample_kernel_sizes": [16,16,4,4],
+ "n_layers_q": 3,
+ "use_spectral_norm": false
+ }
+}

models.py

@@ -0,0 +1,534 @@
+import copy
+import math
+import torch
+from torch import nn
+from torch.nn import functional as F
+
+import commons
+import modules
+import attentions
+import monotonic_align
+
+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
+
+
+class StochasticDurationPredictor(nn.Module):
+ def __init__(self, in_channels, filter_channels, kernel_size, p_dropout, n_flows=4, gin_channels=0):
+ super().__init__()
+ filter_channels = in_channels # it needs to be removed from future version.
+ self.in_channels = in_channels
+ self.filter_channels = filter_channels
+ self.kernel_size = kernel_size
+ self.p_dropout = p_dropout
+ self.n_flows = n_flows
+ self.gin_channels = gin_channels
+
+ self.log_flow = modules.Log()
+ self.flows = nn.ModuleList()
+ self.flows.append(modules.ElementwiseAffine(2))
+ for i in range(n_flows):
+ self.flows.append(modules.ConvFlow(2, filter_channels, kernel_size, n_layers=3))
+ self.flows.append(modules.Flip())
+
+ self.post_pre = nn.Conv1d(1, filter_channels, 1)
+ self.post_proj = nn.Conv1d(filter_channels, filter_channels, 1)
+ self.post_convs = modules.DDSConv(filter_channels, kernel_size, n_layers=3, p_dropout=p_dropout)
+ self.post_flows = nn.ModuleList()
+ self.post_flows.append(modules.ElementwiseAffine(2))
+ for i in range(4):
+ self.post_flows.append(modules.ConvFlow(2, filter_channels, kernel_size, n_layers=3))
+ self.post_flows.append(modules.Flip())
+
+ self.pre = nn.Conv1d(in_channels, filter_channels, 1)
+ self.proj = nn.Conv1d(filter_channels, filter_channels, 1)
+ self.convs = modules.DDSConv(filter_channels, kernel_size, n_layers=3, p_dropout=p_dropout)
+ if gin_channels != 0:
+ self.cond = nn.Conv1d(gin_channels, filter_channels, 1)
+
+ def forward(self, x, x_mask, w=None, g=None, reverse=False, noise_scale=1.0):
+ x = torch.detach(x)
+ x = self.pre(x)
+ if g is not None:
+ g = torch.detach(g)
+ x = x + self.cond(g)
+ x = self.convs(x, x_mask)
+ x = self.proj(x) * x_mask
+
+ if not reverse:
+ flows = self.flows
+ assert w is not None
+
+ logdet_tot_q = 0 
+ h_w = self.post_pre(w)
+ h_w = self.post_convs(h_w, x_mask)
+ h_w = self.post_proj(h_w) * x_mask
+ e_q = torch.randn(w.size(0), 2, w.size(2)).to(device=x.device, dtype=x.dtype) * x_mask
+ z_q = e_q
+ for flow in self.post_flows:
+ z_q, logdet_q = flow(z_q, x_mask, g=(x + h_w))
+ logdet_tot_q += logdet_q
+ z_u, z1 = torch.split(z_q, [1, 1], 1) 
+ u = torch.sigmoid(z_u) * x_mask
+ z0 = (w - u) * x_mask
+ logdet_tot_q += torch.sum((F.logsigmoid(z_u) + F.logsigmoid(-z_u)) * x_mask, [1,2])
+ logq = torch.sum(-0.5 * (math.log(2*math.pi) + (e_q**2)) * x_mask, [1,2]) - logdet_tot_q
+
+ logdet_tot = 0
+ z0, logdet = self.log_flow(z0, x_mask)
+ logdet_tot += logdet
+ z = torch.cat([z0, z1], 1)
+ for flow in flows:
+ z, logdet = flow(z, x_mask, g=x, reverse=reverse)
+ logdet_tot = logdet_tot + logdet
+ nll = torch.sum(0.5 * (math.log(2*math.pi) + (z**2)) * x_mask, [1,2]) - logdet_tot
+ return nll + logq # [b]
+ else:
+ flows = list(reversed(self.flows))
+ flows = flows[:-2] + [flows[-1]] # remove a useless vflow
+ z = torch.randn(x.size(0), 2, x.size(2)).to(device=x.device, dtype=x.dtype) * noise_scale
+ for flow in flows:
+ z = flow(z, x_mask, g=x, reverse=reverse)
+ z0, z1 = torch.split(z, [1, 1], 1)
+ logw = z0
+ return logw
+
+
+class DurationPredictor(nn.Module):
+ def __init__(self, in_channels, filter_channels, kernel_size, p_dropout, gin_channels=0):
+ super().__init__()
+
+ self.in_channels = in_channels
+ self.filter_channels = filter_channels
+ self.kernel_size = kernel_size
+ self.p_dropout = p_dropout
+ self.gin_channels = gin_channels
+
+ self.drop = nn.Dropout(p_dropout)
+ self.conv_1 = nn.Conv1d(in_channels, filter_channels, kernel_size, padding=kernel_size//2)
+ self.norm_1 = modules.LayerNorm(filter_channels)
+ self.conv_2 = nn.Conv1d(filter_channels, filter_channels, kernel_size, padding=kernel_size//2)
+ self.norm_2 = modules.LayerNorm(filter_channels)
+ self.proj = nn.Conv1d(filter_channels, 1, 1)
+
+ if gin_channels != 0:
+ self.cond = nn.Conv1d(gin_channels, in_channels, 1)
+
+ def forward(self, x, x_mask, g=None):
+ x = torch.detach(x)
+ if g is not None:
+ g = torch.detach(g)
+ x = x + self.cond(g)
+ x = self.conv_1(x * x_mask)
+ x = torch.relu(x)
+ x = self.norm_1(x)
+ x = self.drop(x)
+ x = self.conv_2(x * x_mask)
+ x = torch.relu(x)
+ x = self.norm_2(x)
+ x = self.drop(x)
+ x = self.proj(x * x_mask)
+ return x * x_mask
+
+
+class TextEncoder(nn.Module):
+ def __init__(self,
+ n_vocab,
+ out_channels,
+ hidden_channels,
+ filter_channels,
+ n_heads,
+ n_layers,
+ kernel_size,
+ p_dropout):
+ super().__init__()
+ self.n_vocab = n_vocab
+ self.out_channels = out_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.emb = nn.Embedding(n_vocab, hidden_channels)
+ nn.init.normal_(self.emb.weight, 0.0, hidden_channels**-0.5)
+
+ self.encoder = attentions.Encoder(
+ hidden_channels,
+ filter_channels,
+ n_heads,
+ n_layers,
+ kernel_size,
+ p_dropout)
+ self.proj= nn.Conv1d(hidden_channels, out_channels * 2, 1)
+
+ def forward(self, x, x_lengths):
+ x = self.emb(x) * math.sqrt(self.hidden_channels) # [b, t, h]
+ x = torch.transpose(x, 1, -1) # [b, h, t]
+ x_mask = torch.unsqueeze(commons.sequence_mask(x_lengths, x.size(2)), 1).to(x.dtype)
+
+ x = self.encoder(x * x_mask, x_mask)
+ stats = self.proj(x) * x_mask
+
+ m, logs = torch.split(stats, self.out_channels, dim=1)
+ return x, m, logs, x_mask
+
+
+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 PosteriorEncoder(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):
+ 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 Generator(torch.nn.Module):
+ def __init__(self, initial_channel, resblock, resblock_kernel_sizes, resblock_dilation_sizes, upsample_rates, upsample_initial_channel, upsample_kernel_sizes, gin_channels=0):
+ super(Generator, self).__init__()
+ self.num_kernels = len(resblock_kernel_sizes)
+ self.num_upsamples = len(upsample_rates)
+ self.conv_pre = Conv1d(initial_channel, upsample_initial_channel, 7, 1, padding=3)
+ resblock = modules.ResBlock1 if resblock == '1' else modules.ResBlock2
+
+ self.ups = nn.ModuleList()
+ for i, (u, k) in enumerate(zip(upsample_rates, upsample_kernel_sizes)):
+ self.ups.append(weight_norm(
+ ConvTranspose1d(upsample_initial_channel//(2**i), upsample_initial_channel//(2**(i+1)),
+ k, u, padding=(k-u)//2)))
+
+ self.resblocks = nn.ModuleList()
+ for i in range(len(self.ups)):
+ ch = upsample_initial_channel//(2**(i+1))
+ for j, (k, d) in enumerate(zip(resblock_kernel_sizes, resblock_dilation_sizes)):
+ self.resblocks.append(resblock(ch, k, d))
+
+ self.conv_post = Conv1d(ch, 1, 7, 1, padding=3, bias=False)
+ self.ups.apply(init_weights)
+
+ if gin_channels != 0:
+ self.cond = nn.Conv1d(gin_channels, upsample_initial_channel, 1)
+
+ def forward(self, x, g=None):
+ x = self.conv_pre(x)
+ if g is not None:
+ x = x + self.cond(g)
+
+ for i in range(self.num_upsamples):
+ x = F.leaky_relu(x, modules.LRELU_SLOPE)
+ x = self.ups[i](x)
+ xs = None
+ for j in range(self.num_kernels):
+ if xs is None:
+ xs = self.resblocks[i*self.num_kernels+j](x)
+ else:
+ xs += self.resblocks[i*self.num_kernels+j](x)
+ x = xs / self.num_kernels
+ x = F.leaky_relu(x)
+ x = self.conv_post(x)
+ x = torch.tanh(x)
+
+ return x
+
+ def remove_weight_norm(self):
+ print('Removing weight norm...')
+ for l in self.ups:
+ remove_weight_norm(l)
+ for l in self.resblocks:
+ l.remove_weight_norm()
+
+
+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 SynthesizerTrn(nn.Module):
+ """
+ Synthesizer for Training
+ """
+
+ def __init__(self, 
+ n_vocab,
+ 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,
+ n_speakers=0,
+ gin_channels=0,
+ use_sdp=True,
+ **kwargs):
+
+ super().__init__()
+ self.n_vocab = n_vocab
+ 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.n_speakers = n_speakers
+ self.gin_channels = gin_channels
+
+ self.use_sdp = use_sdp
+
+ self.enc_p = TextEncoder(n_vocab,
+ inter_channels,
+ hidden_channels,
+ filter_channels,
+ n_heads,
+ n_layers,
+ kernel_size,
+ p_dropout)
+ self.dec = Generator(inter_channels, resblock, resblock_kernel_sizes, resblock_dilation_sizes, upsample_rates, upsample_initial_channel, upsample_kernel_sizes, gin_channels=gin_channels)
+ self.enc_q = PosteriorEncoder(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)
+
+ if use_sdp:
+ self.dp = StochasticDurationPredictor(hidden_channels, 192, 3, 0.5, 4, gin_channels=gin_channels)
+ else:
+ self.dp = DurationPredictor(hidden_channels, 256, 3, 0.5, gin_channels=gin_channels)
+
+ if n_speakers > 1:
+ self.emb_g = nn.Embedding(n_speakers, gin_channels)
+
+ def forward(self, x, x_lengths, y, y_lengths, sid=None):
+
+ x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths)
+ if self.n_speakers > 0:
+ g = self.emb_g(sid).unsqueeze(-1) # [b, h, 1]
+ else:
+ g = None
+
+ z, m_q, logs_q, y_mask = self.enc_q(y, y_lengths, g=g)
+ z_p = self.flow(z, y_mask, g=g)
+
+ with torch.no_grad():
+ # negative cross-entropy
+ s_p_sq_r = torch.exp(-2 * logs_p) # [b, d, t]
+ neg_cent1 = torch.sum(-0.5 * math.log(2 * math.pi) - logs_p, [1], keepdim=True) # [b, 1, t_s]
+ neg_cent2 = torch.matmul(-0.5 * (z_p ** 2).transpose(1, 2), s_p_sq_r) # [b, t_t, d] x [b, d, t_s] = [b, t_t, t_s]
+ neg_cent3 = torch.matmul(z_p.transpose(1, 2), (m_p * s_p_sq_r)) # [b, t_t, d] x [b, d, t_s] = [b, t_t, t_s]
+ neg_cent4 = torch.sum(-0.5 * (m_p ** 2) * s_p_sq_r, [1], keepdim=True) # [b, 1, t_s]
+ neg_cent = neg_cent1 + neg_cent2 + neg_cent3 + neg_cent4
+
+ attn_mask = torch.unsqueeze(x_mask, 2) * torch.unsqueeze(y_mask, -1)
+ attn = monotonic_align.maximum_path(neg_cent, attn_mask.squeeze(1)).unsqueeze(1).detach()
+
+ w = attn.sum(2)
+ if self.use_sdp:
+ l_length = self.dp(x, x_mask, w, g=g)
+ l_length = l_length / torch.sum(x_mask)
+ else:
+ logw_ = torch.log(w + 1e-6) * x_mask
+ logw = self.dp(x, x_mask, g=g)
+ l_length = torch.sum((logw - logw_)**2, [1,2]) / torch.sum(x_mask) # for averaging 
+
+ # expand prior
+ m_p = torch.matmul(attn.squeeze(1), m_p.transpose(1, 2)).transpose(1, 2)
+ logs_p = torch.matmul(attn.squeeze(1), logs_p.transpose(1, 2)).transpose(1, 2)
+
+ z_slice, ids_slice = commons.rand_slice_segments(z, y_lengths, self.segment_size)
+ o = self.dec(z_slice, g=g)
+ return o, l_length, attn, ids_slice, x_mask, y_mask, (z, z_p, m_p, logs_p, m_q, logs_q)
+
+ def infer(self, x, x_lengths, sid=None, noise_scale=1, length_scale=1, noise_scale_w=1., max_len=None):
+ x, m_p, logs_p, x_mask = self.enc_p(x, x_lengths)
+ if self.n_speakers > 0:
+ g = self.emb_g(sid).unsqueeze(-1) # [b, h, 1]
+ else:
+ g = None
+
+ if self.use_sdp:
+ logw = self.dp(x, x_mask, g=g, reverse=True, noise_scale=noise_scale_w)
+ else:
+ logw = self.dp(x, x_mask, g=g)
+ w = torch.exp(logw) * x_mask * length_scale
+ w_ceil = torch.ceil(w)
+ y_lengths = torch.clamp_min(torch.sum(w_ceil, [1, 2]), 1).long()
+ y_mask = torch.unsqueeze(commons.sequence_mask(y_lengths, None), 1).to(x_mask.dtype)
+ attn_mask = torch.unsqueeze(x_mask, 2) * torch.unsqueeze(y_mask, -1)
+ attn = commons.generate_path(w_ceil, attn_mask)
+
+ m_p = torch.matmul(attn.squeeze(1), m_p.transpose(1, 2)).transpose(1, 2) # [b, t', t], [b, t, d] -> [b, d, t']
+ logs_p = torch.matmul(attn.squeeze(1), logs_p.transpose(1, 2)).transpose(1, 2) # [b, t', t], [b, t, d] -> [b, d, t']
+
+ z_p = m_p + torch.randn_like(m_p) * torch.exp(logs_p) * noise_scale
+ z = self.flow(z_p, y_mask, g=g, reverse=True)
+ o = self.dec((z * y_mask)[:,:,:max_len], g=g)
+ return o, attn, y_mask, (z, z_p, m_p, logs_p)
+
+ def voice_conversion(self, y, y_lengths, sid_src, sid_tgt):
+ assert self.n_speakers > 0, "n_speakers have to be larger than 0."
+ g_src = self.emb_g(sid_src).unsqueeze(-1)
+ g_tgt = self.emb_g(sid_tgt).unsqueeze(-1)
+ z, m_q, logs_q, y_mask = self.enc_q(y, y_lengths, g=g_src)
+ z_p = self.flow(z, y_mask, g=g_src)
+ z_hat = self.flow(z_p, y_mask, g=g_tgt, reverse=True)
+ o_hat = self.dec(z_hat * y_mask, g=g_tgt)
+ return o_hat, y_mask, (z, z_p, z_hat)
+

modules.py

@@ -0,0 +1,390 @@
+import copy
+import math
+import numpy as np
+import scipy
+import torch
+from torch import nn
+from torch.nn import functional as F
+
+from torch.nn import Conv1d, ConvTranspose1d, AvgPool1d, Conv2d
+from torch.nn.utils import weight_norm, remove_weight_norm
+
+import commons
+from commons import init_weights, get_padding
+from transforms import piecewise_rational_quadratic_transform
+
+
+LRELU_SLOPE = 0.1
+
+
+class LayerNorm(nn.Module):
+ def __init__(self, channels, eps=1e-5):
+ super().__init__()
+ self.channels = channels
+ self.eps = eps
+
+ self.gamma = nn.Parameter(torch.ones(channels))
+ self.beta = nn.Parameter(torch.zeros(channels))
+
+ def forward(self, x):
+ x = x.transpose(1, -1)
+ x = F.layer_norm(x, (self.channels,), self.gamma, self.beta, self.eps)
+ return x.transpose(1, -1)
+
+ 
+class ConvReluNorm(nn.Module):
+ def __init__(self, in_channels, hidden_channels, out_channels, kernel_size, n_layers, p_dropout):
+ super().__init__()
+ self.in_channels = in_channels
+ self.hidden_channels = hidden_channels
+ self.out_channels = out_channels
+ self.kernel_size = kernel_size
+ self.n_layers = n_layers
+ self.p_dropout = p_dropout
+ assert n_layers > 1, "Number of layers should be larger than 0."
+
+ self.conv_layers = nn.ModuleList()
+ self.norm_layers = nn.ModuleList()
+ self.conv_layers.append(nn.Conv1d(in_channels, hidden_channels, kernel_size, padding=kernel_size//2))
+ self.norm_layers.append(LayerNorm(hidden_channels))
+ self.relu_drop = nn.Sequential(
+ nn.ReLU(),
+ nn.Dropout(p_dropout))
+ for _ in range(n_layers-1):
+ self.conv_layers.append(nn.Conv1d(hidden_channels, hidden_channels, kernel_size, padding=kernel_size//2))
+ self.norm_layers.append(LayerNorm(hidden_channels))
+ self.proj = nn.Conv1d(hidden_channels, out_channels, 1)
+ self.proj.weight.data.zero_()
+ self.proj.bias.data.zero_()
+
+ def forward(self, x, x_mask):
+ x_org = x
+ for i in range(self.n_layers):
+ x = self.conv_layers[i](x * x_mask)
+ x = self.norm_layers[i](x)
+ x = self.relu_drop(x)
+ x = x_org + self.proj(x)
+ return x * x_mask
+
+
+class DDSConv(nn.Module):
+ """
+ Dialted and Depth-Separable Convolution
+ """
+ def __init__(self, channels, kernel_size, n_layers, p_dropout=0.):
+ super().__init__()
+ self.channels = channels
+ self.kernel_size = kernel_size
+ self.n_layers = n_layers
+ self.p_dropout = p_dropout
+
+ self.drop = nn.Dropout(p_dropout)
+ self.convs_sep = nn.ModuleList()
+ self.convs_1x1 = nn.ModuleList()
+ self.norms_1 = nn.ModuleList()
+ self.norms_2 = nn.ModuleList()
+ for i in range(n_layers):
+ dilation = kernel_size ** i
+ padding = (kernel_size * dilation - dilation) // 2
+ self.convs_sep.append(nn.Conv1d(channels, channels, kernel_size, 
+ groups=channels, dilation=dilation, padding=padding
+ ))
+ self.convs_1x1.append(nn.Conv1d(channels, channels, 1))
+ self.norms_1.append(LayerNorm(channels))
+ self.norms_2.append(LayerNorm(channels))
+
+ def forward(self, x, x_mask, g=None):
+ if g is not None:
+ x = x + g
+ for i in range(self.n_layers):
+ y = self.convs_sep[i](x * x_mask)
+ y = self.norms_1[i](y)
+ y = F.gelu(y)
+ y = self.convs_1x1[i](y)
+ y = self.norms_2[i](y)
+ y = F.gelu(y)
+ y = self.drop(y)
+ x = x + y
+ return x * x_mask
+
+
+class WN(torch.nn.Module):
+ def __init__(self, hidden_channels, kernel_size, dilation_rate, n_layers, gin_channels=0, p_dropout=0):
+ super(WN, self).__init__()
+ assert(kernel_size % 2 == 1)
+ 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.p_dropout = p_dropout
+
+ self.in_layers = torch.nn.ModuleList()
+ self.res_skip_layers = torch.nn.ModuleList()
+ self.drop = nn.Dropout(p_dropout)
+
+ if gin_channels != 0:
+ cond_layer = torch.nn.Conv1d(gin_channels, 2*hidden_channels*n_layers, 1)
+ self.cond_layer = torch.nn.utils.weight_norm(cond_layer, name='weight')
+
+ for i in range(n_layers):
+ dilation = dilation_rate ** i
+ padding = int((kernel_size * dilation - dilation) / 2)
+ in_layer = torch.nn.Conv1d(hidden_channels, 2*hidden_channels, kernel_size,
+ dilation=dilation, padding=padding)
+ in_layer = torch.nn.utils.weight_norm(in_layer, name='weight')
+ self.in_layers.append(in_layer)
+
+ # last one is not necessary
+ if i < n_layers - 1:
+ res_skip_channels = 2 * hidden_channels
+ else:
+ res_skip_channels = hidden_channels
+
+ res_skip_layer = torch.nn.Conv1d(hidden_channels, res_skip_channels, 1)
+ res_skip_layer = torch.nn.utils.weight_norm(res_skip_layer, name='weight')
+ self.res_skip_layers.append(res_skip_layer)
+
+ def forward(self, x, x_mask, g=None, **kwargs):
+ output = torch.zeros_like(x)
+ n_channels_tensor = torch.IntTensor([self.hidden_channels])
+
+ if g is not None:
+ g = self.cond_layer(g)
+
+ for i in range(self.n_layers):
+ x_in = self.in_layers[i](x)
+ if g is not None:
+ cond_offset = i * 2 * self.hidden_channels
+ g_l = g[:,cond_offset:cond_offset+2*self.hidden_channels,:]
+ else:
+ g_l = torch.zeros_like(x_in)
+
+ acts = commons.fused_add_tanh_sigmoid_multiply(
+ x_in,
+ g_l,
+ n_channels_tensor)
+ acts = self.drop(acts)
+
+ res_skip_acts = self.res_skip_layers[i](acts)
+ if i < self.n_layers - 1:
+ res_acts = res_skip_acts[:,:self.hidden_channels,:]
+ x = (x + res_acts) * x_mask
+ output = output + res_skip_acts[:,self.hidden_channels:,:]
+ else:
+ output = output + res_skip_acts
+ return output * x_mask
+
+ def remove_weight_norm(self):
+ if self.gin_channels != 0:
+ torch.nn.utils.remove_weight_norm(self.cond_layer)
+ for l in self.in_layers:
+ torch.nn.utils.remove_weight_norm(l)
+ for l in self.res_skip_layers:
+ torch.nn.utils.remove_weight_norm(l)
+
+
+class ResBlock1(torch.nn.Module):
+ def __init__(self, channels, kernel_size=3, dilation=(1, 3, 5)):
+ super(ResBlock1, self).__init__()
+ self.convs1 = nn.ModuleList([
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0],
+ padding=get_padding(kernel_size, dilation[0]))),
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1],
+ padding=get_padding(kernel_size, dilation[1]))),
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[2],
+ padding=get_padding(kernel_size, dilation[2])))
+ ])
+ self.convs1.apply(init_weights)
+
+ self.convs2 = nn.ModuleList([
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
+ padding=get_padding(kernel_size, 1))),
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
+ padding=get_padding(kernel_size, 1))),
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1,
+ padding=get_padding(kernel_size, 1)))
+ ])
+ self.convs2.apply(init_weights)
+
+ def forward(self, x, x_mask=None):
+ for c1, c2 in zip(self.convs1, self.convs2):
+ xt = F.leaky_relu(x, LRELU_SLOPE)
+ if x_mask is not None:
+ xt = xt * x_mask
+ xt = c1(xt)
+ xt = F.leaky_relu(xt, LRELU_SLOPE)
+ if x_mask is not None:
+ xt = xt * x_mask
+ xt = c2(xt)
+ x = xt + x
+ if x_mask is not None:
+ x = x * x_mask
+ return x
+
+ def remove_weight_norm(self):
+ for l in self.convs1:
+ remove_weight_norm(l)
+ for l in self.convs2:
+ remove_weight_norm(l)
+
+
+class ResBlock2(torch.nn.Module):
+ def __init__(self, channels, kernel_size=3, dilation=(1, 3)):
+ super(ResBlock2, self).__init__()
+ self.convs = nn.ModuleList([
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0],
+ padding=get_padding(kernel_size, dilation[0]))),
+ weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1],
+ padding=get_padding(kernel_size, dilation[1])))
+ ])
+ self.convs.apply(init_weights)
+
+ def forward(self, x, x_mask=None):
+ for c in self.convs:
+ xt = F.leaky_relu(x, LRELU_SLOPE)
+ if x_mask is not None:
+ xt = xt * x_mask
+ xt = c(xt)
+ x = xt + x
+ if x_mask is not None:
+ x = x * x_mask
+ return x
+
+ def remove_weight_norm(self):
+ for l in self.convs:
+ remove_weight_norm(l)
+
+
+class Log(nn.Module):
+ def forward(self, x, x_mask, reverse=False, **kwargs):
+ if not reverse:
+ y = torch.log(torch.clamp_min(x, 1e-5)) * x_mask
+ logdet = torch.sum(-y, [1, 2])
+ return y, logdet
+ else:
+ x = torch.exp(x) * x_mask
+ return x
+ 
+
+class Flip(nn.Module):
+ def forward(self, x, *args, reverse=False, **kwargs):
+ x = torch.flip(x, [1])
+ if not reverse:
+ logdet = torch.zeros(x.size(0)).to(dtype=x.dtype, device=x.device)
+ return x, logdet
+ else:
+ return x
+
+
+class ElementwiseAffine(nn.Module):
+ def __init__(self, channels):
+ super().__init__()
+ self.channels = channels
+ self.m = nn.Parameter(torch.zeros(channels,1))
+ self.logs = nn.Parameter(torch.zeros(channels,1))
+
+ def forward(self, x, x_mask, reverse=False, **kwargs):
+ if not reverse:
+ y = self.m + torch.exp(self.logs) * x
+ y = y * x_mask
+ logdet = torch.sum(self.logs * x_mask, [1,2])
+ return y, logdet
+ else:
+ x = (x - self.m) * torch.exp(-self.logs) * x_mask
+ return x
+
+
+class ResidualCouplingLayer(nn.Module):
+ def __init__(self,
+ channels,
+ hidden_channels,
+ kernel_size,
+ dilation_rate,
+ n_layers,
+ p_dropout=0,
+ gin_channels=0,
+ mean_only=False):
+ assert channels % 2 == 0, "channels should be divisible by 2"
+ 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.half_channels = channels // 2
+ self.mean_only = mean_only
+
+ self.pre = nn.Conv1d(self.half_channels, hidden_channels, 1)
+ self.enc = WN(hidden_channels, kernel_size, dilation_rate, n_layers, p_dropout=p_dropout, gin_channels=gin_channels)
+ self.post = nn.Conv1d(hidden_channels, self.half_channels * (2 - mean_only), 1)
+ self.post.weight.data.zero_()
+ self.post.bias.data.zero_()
+
+ def forward(self, x, x_mask, g=None, reverse=False):
+ x0, x1 = torch.split(x, [self.half_channels]*2, 1)
+ h = self.pre(x0) * x_mask
+ h = self.enc(h, x_mask, g=g)
+ stats = self.post(h) * x_mask
+ if not self.mean_only:
+ m, logs = torch.split(stats, [self.half_channels]*2, 1)
+ else:
+ m = stats
+ logs = torch.zeros_like(m)
+
+ if not reverse:
+ x1 = m + x1 * torch.exp(logs) * x_mask
+ x = torch.cat([x0, x1], 1)
+ logdet = torch.sum(logs, [1,2])
+ return x, logdet
+ else:
+ x1 = (x1 - m) * torch.exp(-logs) * x_mask
+ x = torch.cat([x0, x1], 1)
+ return x
+
+
+class ConvFlow(nn.Module):
+ def __init__(self, in_channels, filter_channels, kernel_size, n_layers, num_bins=10, tail_bound=5.0):
+ super().__init__()
+ self.in_channels = in_channels
+ self.filter_channels = filter_channels
+ self.kernel_size = kernel_size
+ self.n_layers = n_layers
+ self.num_bins = num_bins
+ self.tail_bound = tail_bound
+ self.half_channels = in_channels // 2
+
+ self.pre = nn.Conv1d(self.half_channels, filter_channels, 1)
+ self.convs = DDSConv(filter_channels, kernel_size, n_layers, p_dropout=0.)
+ self.proj = nn.Conv1d(filter_channels, self.half_channels * (num_bins * 3 - 1), 1)
+ self.proj.weight.data.zero_()
+ self.proj.bias.data.zero_()
+
+ def forward(self, x, x_mask, g=None, reverse=False):
+ x0, x1 = torch.split(x, [self.half_channels]*2, 1)
+ h = self.pre(x0)
+ h = self.convs(h, x_mask, g=g)
+ h = self.proj(h) * x_mask
+
+ b, c, t = x0.shape
+ h = h.reshape(b, c, -1, t).permute(0, 1, 3, 2) # [b, cx?, t] -> [b, c, t, ?]
+
+ unnormalized_widths = h[..., :self.num_bins] / math.sqrt(self.filter_channels)
+ unnormalized_heights = h[..., self.num_bins:2*self.num_bins] / math.sqrt(self.filter_channels)
+ unnormalized_derivatives = h[..., 2 * self.num_bins:]
+
+ x1, logabsdet = piecewise_rational_quadratic_transform(x1,
+ unnormalized_widths,
+ unnormalized_heights,
+ unnormalized_derivatives,
+ inverse=reverse,
+ tails='linear',
+ tail_bound=self.tail_bound
+ )
+
+ x = torch.cat([x0, x1], 1) * x_mask
+ logdet = torch.sum(logabsdet * x_mask, [1,2])
+ if not reverse:
+ return x, logdet
+ else:
+ return x

monotonic_align/__init__.py

@@ -0,0 +1,19 @@
+import numpy as np
+import torch
+from .monotonic_align.core import maximum_path_c
+
+
+def maximum_path(neg_cent, mask):
+ """ Cython optimized version.
+ neg_cent: [b, t_t, t_s]
+ mask: [b, t_t, t_s]
+ """
+ device = neg_cent.device
+ dtype = neg_cent.dtype
+ neg_cent = neg_cent.data.cpu().numpy().astype(np.float32)
+ path = np.zeros(neg_cent.shape, dtype=np.int32)
+
+ t_t_max = mask.sum(1)[:, 0].data.cpu().numpy().astype(np.int32)
+ t_s_max = mask.sum(2)[:, 0].data.cpu().numpy().astype(np.int32)
+ maximum_path_c(path, neg_cent, t_t_max, t_s_max)
+ return torch.from_numpy(path).to(device=device, dtype=dtype)

monotonic_align/core.pyx

@@ -0,0 +1,42 @@
+cimport cython
+from cython.parallel import prange
+
+
+@cython.boundscheck(False)
+@cython.wraparound(False)
+cdef void maximum_path_each(int[:,::1] path, float[:,::1] value, int t_y, int t_x, float max_neg_val=-1e9) nogil:
+ cdef int x
+ cdef int y
+ cdef float v_prev
+ cdef float v_cur
+ cdef float tmp
+ cdef int index = t_x - 1
+
+ for y in range(t_y):
+ for x in range(max(0, t_x + y - t_y), min(t_x, y + 1)):
+ if x == y:
+ v_cur = max_neg_val
+ else:
+ v_cur = value[y-1, x]
+ if x == 0:
+ if y == 0:
+ v_prev = 0.
+ else:
+ v_prev = max_neg_val
+ else:
+ v_prev = value[y-1, x-1]
+ value[y, x] += max(v_prev, v_cur)
+
+ for y in range(t_y - 1, -1, -1):
+ path[y, index] = 1
+ if index != 0 and (index == y or value[y-1, index] < value[y-1, index-1]):
+ index = index - 1
+
+
+@cython.boundscheck(False)
+@cython.wraparound(False)
+cpdef void maximum_path_c(int[:,:,::1] paths, float[:,:,::1] values, int[::1] t_ys, int[::1] t_xs) nogil:
+ cdef int b = paths.shape[0]
+ cdef int i
+ for i in prange(b, nogil=True):
+ maximum_path_each(paths[i], values[i], t_ys[i], t_xs[i])

monotonic_align/setup.py

@@ -0,0 +1,9 @@
+from distutils.core import setup
+from Cython.Build import cythonize
+import numpy
+
+setup(
+ name = 'monotonic_align',
+ ext_modules = cythonize("core.pyx"),
+ include_dirs=[numpy.get_include()]
+)

text/__init__.py

@@ -0,0 +1,64 @@
+""" from https://github.com/keithito/tacotron """
+from text import cleaners
+from text.symbols import symbols,symbols_zh
+
+
+# Mappings from symbol to numeric ID and vice versa:
+# _symbol_to_id = {s: i for i, s in enumerate(symbols)}
+# _id_to_symbol = {i: s for i, s in enumerate(symbols)}
+
+chinese_mode = True
+if chinese_mode:
+ _symbol_to_id = {s: i for i, s in enumerate(symbols_zh)}
+ _id_to_symbol = {i: s for i, s in enumerate(symbols_zh)}
+else:
+ _symbol_to_id = {s: i for i, s in enumerate(symbols)}
+ _id_to_symbol = {i: s for i, s in enumerate(symbols)}
+
+def text_to_sequence(text, cleaner_names, ):
+ '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
+ Args:
+ text: string to convert to a sequence
+ cleaner_names: names of the cleaner functions to run the text through
+ Returns:
+ List of integers corresponding to the symbols in the text
+ '''
+ sequence = []
+
+ clean_text = _clean_text(text, cleaner_names)
+ for symbol in clean_text:
+ symbol_id = _symbol_to_id[symbol]
+ sequence += [symbol_id]
+ return sequence
+
+
+def cleaned_text_to_sequence(cleaned_text, chinese_mode=True):
+ '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
+ Args:
+ text: string to convert to a sequence
+ Returns:
+ List of integers corresponding to the symbols in the text
+ '''
+ # if chinese_mode:
+ # sequence = [_symbol_to_id_zh[symbol] for symbol in cleaned_text]
+ # else:
+ sequence = [_symbol_to_id[symbol] for symbol in cleaned_text]
+ return sequence
+
+
+def sequence_to_text(sequence):
+ '''Converts a sequence of IDs back to a string'''
+ result = ''
+ for symbol_id in sequence:
+ s = _id_to_symbol[symbol_id]
+ result += s
+ return result
+
+
+def _clean_text(text, cleaner_names):
+ for name in cleaner_names:
+ cleaner = getattr(cleaners, name)
+ if not cleaner:
+ raise Exception('Unknown cleaner: %s' % name)
+ text = cleaner(text)
+ return text

text/cleaners.py

@@ -0,0 +1,138 @@
+""" from https://github.com/keithito/tacotron """
+
+'''
+Cleaners are transformations that run over the input text at both training and eval time.
+
+Cleaners can be selected by passing a comma-delimited list of cleaner names as the "cleaners"
+hyperparameter. Some cleaners are English-specific. You'll typically want to use:
+ 1. "english_cleaners" for English text
+ 2. "transliteration_cleaners" for non-English text that can be transliterated to ASCII using
+ the Unidecode library (https://pypi.python.org/pypi/Unidecode)
+ 3. "basic_cleaners" if you do not want to transliterate (in this case, you should also update
+ the symbols in symbols.py to match your data).
+'''
+
+import re
+from unidecode import unidecode
+from phonemizer import phonemize
+from pypinyin import Style, pinyin
+from pypinyin.style._utils import get_finals, get_initials
+# Regular expression matching whitespace:
+_whitespace_re = re.compile(r'\s+')
+
+# List of (regular expression, replacement) pairs for abbreviations:
+_abbreviations = [(re.compile('\\b%s\\.' % x[0], re.IGNORECASE), x[1]) for x in [
+ ('mrs', 'misess'),
+ ('mr', 'mister'),
+ ('dr', 'doctor'),
+ ('st', 'saint'),
+ ('co', 'company'),
+ ('jr', 'junior'),
+ ('maj', 'major'),
+ ('gen', 'general'),
+ ('drs', 'doctors'),
+ ('rev', 'reverend'),
+ ('lt', 'lieutenant'),
+ ('hon', 'honorable'),
+ ('sgt', 'sergeant'),
+ ('capt', 'captain'),
+ ('esq', 'esquire'),
+ ('ltd', 'limited'),
+ ('col', 'colonel'),
+ ('ft', 'fort'),
+]]
+
+
+def expand_abbreviations(text):
+ for regex, replacement in _abbreviations:
+ text = re.sub(regex, replacement, text)
+ return text
+
+
+def expand_numbers(text):
+ return normalize_numbers(text)
+
+
+def lowercase(text):
+ return text.lower()
+
+
+def collapse_whitespace(text):
+ return re.sub(_whitespace_re, ' ', text)
+
+
+def convert_to_ascii(text):
+ return unidecode(text)
+
+
+def basic_cleaners(text):
+ '''Basic pipeline that lowercases and collapses whitespace without transliteration.'''
+ text = lowercase(text)
+ text = collapse_whitespace(text)
+ return text
+
+
+def transliteration_cleaners(text):
+ '''Pipeline for non-English text that transliterates to ASCII.'''
+ text = convert_to_ascii(text)
+ text = lowercase(text)
+ text = collapse_whitespace(text)
+ return text
+
+
+def english_cleaners(text):
+ '''Pipeline for English text, including abbreviation expansion.'''
+ text = convert_to_ascii(text)
+ text = lowercase(text)
+ text = expand_abbreviations(text)
+ phonemes = phonemize(text, language='en-us', backend='espeak', strip=True)
+ phonemes = collapse_whitespace(phonemes)
+ return phonemes
+
+
+def english_cleaners2(text):
+ '''Pipeline for English text, including abbreviation expansion. + punctuation + stress'''
+ text = convert_to_ascii(text)
+ text = lowercase(text)
+ text = expand_abbreviations(text)
+ phonemes = phonemize(text, language='en-us', backend='espeak', strip=True, preserve_punctuation=True, with_stress=True)
+ phonemes = collapse_whitespace(phonemes)
+ return phonemes
+
+
+
+
+def chinese_cleaners1(text):
+ from pypinyin import Style, pinyin
+
+ phones = [phone[0] for phone in pinyin(text, style=Style.TONE3)]
+ return ' '.join(phones)
+ 
+ 
+def chinese_cleaners2(text):
+ phones = [
+ p
+ for phone in pinyin(text, style=Style.TONE3)
+ for p in [
+ get_initials(phone[0], strict=True),
+ get_finals(phone[0][:-1], strict=True) + phone[0][-1]
+ if phone[0][-1].isdigit()
+ else get_finals(phone[0], strict=True)
+ if phone[0][-1].isalnum()
+ else phone[0],
+ ]
+ # Remove the case of individual tones as a phoneme
+ if len(p) != 0 and not p.isdigit()
+ ]
+ return phones
+ # return phonemes
+ 
+if __name__ == '__main__':
+ res = chinese_cleaners2('这是语音测试！')
+ print(res)
+ res = chinese_cleaners1('"第一，南京不是发展的不行，是大家对他期望很高，')
+ print(res)
+ 
+ 
+ res = english_cleaners2('this is a club test for one train.GDP')
+ print(res)

text/symbols.py

@@ -0,0 +1,24 @@
+""" from https://github.com/keithito/tacotron """
+
+'''
+Defines the set of symbols used in text input to the model.
+'''
+_pad = '_'
+_punctuation = ';:,.!?¡¿—…"«»“” '
+
+_punctuation_zh = '；：，。！？-“”《》、 '
+_letters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz'
+
+_numbers = '123450'
+_others = ''
+
+_letters_ipa = "ɑɐɒæɓʙβɔɕçɗɖðʤəɘɚɛɜɝɞɟʄɡɠɢʛɦɧħɥʜɨɪʝɭɬɫɮʟɱɯɰŋɳɲɴøɵɸθœɶʘɹɺɾɻʀʁɽʂʃʈʧʉʊʋⱱʌɣɤʍχʎʏʑʐʒʔʡʕʢǀǁǂǃˈˌːˑʼʴʰʱʲʷˠˤ˞↓↑→↗↘'̩'ᵻ"
+
+
+# Export all symbols:
+symbols = [_pad] + list(_punctuation) + list(_letters) + list(_letters_ipa)
+
+symbols_zh = [_pad] + list(_punctuation_zh) + list(_letters) + list(_numbers)
+
+# Special symbol ids
+SPACE_ID = symbols.index(" ")

transforms.py

@@ -0,0 +1,193 @@
+import torch
+from torch.nn import functional as F
+
+import numpy as np
+
+
+DEFAULT_MIN_BIN_WIDTH = 1e-3
+DEFAULT_MIN_BIN_HEIGHT = 1e-3
+DEFAULT_MIN_DERIVATIVE = 1e-3
+
+
+def piecewise_rational_quadratic_transform(inputs, 
+ unnormalized_widths,
+ unnormalized_heights,
+ unnormalized_derivatives,
+ inverse=False,
+ tails=None, 
+ tail_bound=1.,
+ min_bin_width=DEFAULT_MIN_BIN_WIDTH,
+ min_bin_height=DEFAULT_MIN_BIN_HEIGHT,
+ min_derivative=DEFAULT_MIN_DERIVATIVE):
+
+ if tails is None:
+ spline_fn = rational_quadratic_spline
+ spline_kwargs = {}
+ else:
+ spline_fn = unconstrained_rational_quadratic_spline
+ spline_kwargs = {
+ 'tails': tails,
+ 'tail_bound': tail_bound
+ }
+
+ outputs, logabsdet = spline_fn(
+ inputs=inputs,
+ unnormalized_widths=unnormalized_widths,
+ unnormalized_heights=unnormalized_heights,
+ unnormalized_derivatives=unnormalized_derivatives,
+ inverse=inverse,
+ min_bin_width=min_bin_width,
+ min_bin_height=min_bin_height,
+ min_derivative=min_derivative,
+ **spline_kwargs
+ )
+ return outputs, logabsdet
+
+
+def searchsorted(bin_locations, inputs, eps=1e-6):
+ bin_locations[..., -1] += eps
+ return torch.sum(
+ inputs[..., None] >= bin_locations,
+ dim=-1
+ ) - 1
+
+
+def unconstrained_rational_quadratic_spline(inputs,
+ unnormalized_widths,
+ unnormalized_heights,
+ unnormalized_derivatives,
+ inverse=False,
+ tails='linear',
+ tail_bound=1.,
+ min_bin_width=DEFAULT_MIN_BIN_WIDTH,
+ min_bin_height=DEFAULT_MIN_BIN_HEIGHT,
+ min_derivative=DEFAULT_MIN_DERIVATIVE):
+ inside_interval_mask = (inputs >= -tail_bound) & (inputs <= tail_bound)
+ outside_interval_mask = ~inside_interval_mask
+
+ outputs = torch.zeros_like(inputs)
+ logabsdet = torch.zeros_like(inputs)
+
+ if tails == 'linear':
+ unnormalized_derivatives = F.pad(unnormalized_derivatives, pad=(1, 1))
+ constant = np.log(np.exp(1 - min_derivative) - 1)
+ unnormalized_derivatives[..., 0] = constant
+ unnormalized_derivatives[..., -1] = constant
+
+ outputs[outside_interval_mask] = inputs[outside_interval_mask]
+ logabsdet[outside_interval_mask] = 0
+ else:
+ raise RuntimeError('{} tails are not implemented.'.format(tails))
+
+ outputs[inside_interval_mask], logabsdet[inside_interval_mask] = rational_quadratic_spline(
+ inputs=inputs[inside_interval_mask],
+ unnormalized_widths=unnormalized_widths[inside_interval_mask, :],
+ unnormalized_heights=unnormalized_heights[inside_interval_mask, :],
+ unnormalized_derivatives=unnormalized_derivatives[inside_interval_mask, :],
+ inverse=inverse,
+ left=-tail_bound, right=tail_bound, bottom=-tail_bound, top=tail_bound,
+ min_bin_width=min_bin_width,
+ min_bin_height=min_bin_height,
+ min_derivative=min_derivative
+ )
+
+ return outputs, logabsdet
+
+def rational_quadratic_spline(inputs,
+ unnormalized_widths,
+ unnormalized_heights,
+ unnormalized_derivatives,
+ inverse=False,
+ left=0., right=1., bottom=0., top=1.,
+ min_bin_width=DEFAULT_MIN_BIN_WIDTH,
+ min_bin_height=DEFAULT_MIN_BIN_HEIGHT,
+ min_derivative=DEFAULT_MIN_DERIVATIVE):
+ if torch.min(inputs) < left or torch.max(inputs) > right:
+ raise ValueError('Input to a transform is not within its domain')
+
+ num_bins = unnormalized_widths.shape[-1]
+
+ if min_bin_width * num_bins > 1.0:
+ raise ValueError('Minimal bin width too large for the number of bins')
+ if min_bin_height * num_bins > 1.0:
+ raise ValueError('Minimal bin height too large for the number of bins')
+
+ widths = F.softmax(unnormalized_widths, dim=-1)
+ widths = min_bin_width + (1 - min_bin_width * num_bins) * widths
+ cumwidths = torch.cumsum(widths, dim=-1)
+ cumwidths = F.pad(cumwidths, pad=(1, 0), mode='constant', value=0.0)
+ cumwidths = (right - left) * cumwidths + left
+ cumwidths[..., 0] = left
+ cumwidths[..., -1] = right
+ widths = cumwidths[..., 1:] - cumwidths[..., :-1]
+
+ derivatives = min_derivative + F.softplus(unnormalized_derivatives)
+
+ heights = F.softmax(unnormalized_heights, dim=-1)
+ heights = min_bin_height + (1 - min_bin_height * num_bins) * heights
+ cumheights = torch.cumsum(heights, dim=-1)
+ cumheights = F.pad(cumheights, pad=(1, 0), mode='constant', value=0.0)
+ cumheights = (top - bottom) * cumheights + bottom
+ cumheights[..., 0] = bottom
+ cumheights[..., -1] = top
+ heights = cumheights[..., 1:] - cumheights[..., :-1]
+
+ if inverse:
+ bin_idx = searchsorted(cumheights, inputs)[..., None]
+ else:
+ bin_idx = searchsorted(cumwidths, inputs)[..., None]
+
+ input_cumwidths = cumwidths.gather(-1, bin_idx)[..., 0]
+ input_bin_widths = widths.gather(-1, bin_idx)[..., 0]
+
+ input_cumheights = cumheights.gather(-1, bin_idx)[..., 0]
+ delta = heights / widths
+ input_delta = delta.gather(-1, bin_idx)[..., 0]
+
+ input_derivatives = derivatives.gather(-1, bin_idx)[..., 0]
+ input_derivatives_plus_one = derivatives[..., 1:].gather(-1, bin_idx)[..., 0]
+
+ input_heights = heights.gather(-1, bin_idx)[..., 0]
+
+ if inverse:
+ a = (((inputs - input_cumheights) * (input_derivatives
+ + input_derivatives_plus_one
+ - 2 * input_delta)
+ + input_heights * (input_delta - input_derivatives)))
+ b = (input_heights * input_derivatives
+ - (inputs - input_cumheights) * (input_derivatives
+ + input_derivatives_plus_one
+ - 2 * input_delta))
+ c = - input_delta * (inputs - input_cumheights)
+
+ discriminant = b.pow(2) - 4 * a * c
+ assert (discriminant >= 0).all()
+
+ root = (2 * c) / (-b - torch.sqrt(discriminant))
+ outputs = root * input_bin_widths + input_cumwidths
+
+ theta_one_minus_theta = root * (1 - root)
+ denominator = input_delta + ((input_derivatives + input_derivatives_plus_one - 2 * input_delta)
+ * theta_one_minus_theta)
+ derivative_numerator = input_delta.pow(2) * (input_derivatives_plus_one * root.pow(2)
+ + 2 * input_delta * theta_one_minus_theta
+ + input_derivatives * (1 - root).pow(2))
+ logabsdet = torch.log(derivative_numerator) - 2 * torch.log(denominator)
+
+ return outputs, -logabsdet
+ else:
+ theta = (inputs - input_cumwidths) / input_bin_widths
+ theta_one_minus_theta = theta * (1 - theta)
+
+ numerator = input_heights * (input_delta * theta.pow(2)
+ + input_derivatives * theta_one_minus_theta)
+ denominator = input_delta + ((input_derivatives + input_derivatives_plus_one - 2 * input_delta)
+ * theta_one_minus_theta)
+ outputs = input_cumheights + numerator / denominator
+
+ derivative_numerator = input_delta.pow(2) * (input_derivatives_plus_one * theta.pow(2)
+ + 2 * input_delta * theta_one_minus_theta
+ + input_derivatives * (1 - theta).pow(2))
+ logabsdet = torch.log(derivative_numerator) - 2 * torch.log(denominator)
+
+ return outputs, logabsdet

utils.py

@@ -0,0 +1,259 @@
+import os
+import glob
+import sys
+import argparse
+import logging
+import json
+import subprocess
+import numpy as np
+from scipy.io.wavfile import read
+import torch
+
+MATPLOTLIB_FLAG = False
+
+logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
+logger = logging
+
+
+def load_checkpoint(checkpoint_path, model, optimizer=None):
+ assert os.path.isfile(checkpoint_path)
+ checkpoint_dict = torch.load(checkpoint_path, map_location='cpu')
+ iteration = checkpoint_dict['iteration']
+ learning_rate = checkpoint_dict['learning_rate']
+ if optimizer is not None:
+ optimizer.load_state_dict(checkpoint_dict['optimizer'])
+ saved_state_dict = checkpoint_dict['model']
+ if hasattr(model, 'module'):
+ state_dict = model.module.state_dict()
+ else:
+ state_dict = model.state_dict()
+ new_state_dict= {}
+ for k, v in state_dict.items():
+ try:
+ new_state_dict[k] = saved_state_dict[k]
+ except:
+ logger.info("%s is not in the checkpoint" % k)
+ new_state_dict[k] = v
+ if hasattr(model, 'module'):
+ model.module.load_state_dict(new_state_dict)
+ else:
+ model.load_state_dict(new_state_dict)
+ logger.info("Loaded checkpoint '{}' (iteration {})" .format(
+ checkpoint_path, iteration))
+ return model, optimizer, learning_rate, iteration
+
+
+def save_checkpoint(model, optimizer, learning_rate, iteration, checkpoint_path):
+ logger.info("Saving model and optimizer state at iteration {} to {}".format(
+ iteration, checkpoint_path))
+ if hasattr(model, 'module'):
+ state_dict = model.module.state_dict()
+ else:
+ state_dict = model.state_dict()
+ torch.save({'model': state_dict,
+ 'iteration': iteration,
+ 'optimizer': optimizer.state_dict(),
+ 'learning_rate': learning_rate}, checkpoint_path)
+
+
+def summarize(writer, global_step, scalars={}, histograms={}, images={}, audios={}, audio_sampling_rate=22050):
+ for k, v in scalars.items():
+ writer.add_scalar(k, v, global_step)
+ for k, v in histograms.items():
+ writer.add_histogram(k, v, global_step)
+ for k, v in images.items():
+ writer.add_image(k, v, global_step, dataformats='HWC')
+ for k, v in audios.items():
+ writer.add_audio(k, v, global_step, audio_sampling_rate)
+
+
+def latest_checkpoint_path(dir_path, regex="G_*.pth"):
+ f_list = glob.glob(os.path.join(dir_path, regex))
+ f_list.sort(key=lambda f: int("".join(filter(str.isdigit, f))))
+ x = f_list[-1]
+ print(x)
+ return x
+
+
+def plot_spectrogram_to_numpy(spectrogram):
+ global MATPLOTLIB_FLAG
+ if not MATPLOTLIB_FLAG:
+ import matplotlib
+ matplotlib.use("Agg")
+ MATPLOTLIB_FLAG = True
+ mpl_logger = logging.getLogger('matplotlib')
+ mpl_logger.setLevel(logging.WARNING)
+ import matplotlib.pylab as plt
+ import numpy as np
+ 
+ fig, ax = plt.subplots(figsize=(10,2))
+ im = ax.imshow(spectrogram, aspect="auto", origin="lower",
+ interpolation='none')
+ plt.colorbar(im, ax=ax)
+ plt.xlabel("Frames")
+ plt.ylabel("Channels")
+ plt.tight_layout()
+
+ fig.canvas.draw()
+ data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep='')
+ data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
+ plt.close()
+ return data
+
+
+def plot_alignment_to_numpy(alignment, info=None):
+ global MATPLOTLIB_FLAG
+ if not MATPLOTLIB_FLAG:
+ import matplotlib
+ matplotlib.use("Agg")
+ MATPLOTLIB_FLAG = True
+ mpl_logger = logging.getLogger('matplotlib')
+ mpl_logger.setLevel(logging.WARNING)
+ import matplotlib.pylab as plt
+ import numpy as np
+
+ fig, ax = plt.subplots(figsize=(6, 4))
+ im = ax.imshow(alignment.transpose(), aspect='auto', origin='lower',
+ interpolation='none')
+ fig.colorbar(im, ax=ax)
+ xlabel = 'Decoder timestep'
+ if info is not None:
+ xlabel += '\n\n' + info
+ plt.xlabel(xlabel)
+ plt.ylabel('Encoder timestep')
+ plt.tight_layout()
+
+ fig.canvas.draw()
+ data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep='')
+ data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
+ plt.close()
+ return data
+
+
+def load_wav_to_torch(full_path):
+ sampling_rate, data = read(full_path)
+ return torch.FloatTensor(data.astype(np.float32)), sampling_rate
+
+
+def load_filepaths_and_text(filename, split="|"):
+ with open(filename, encoding='utf-8') as f:
+ filepaths_and_text = [line.strip().split(split) for line in f]
+ return filepaths_and_text
+
+
+def get_hparams(init=True):
+ parser = argparse.ArgumentParser()
+ parser.add_argument('-c', '--config', type=str, default="./configs/woman_csmsc.json",
+ help='JSON file for configuration')
+ parser.add_argument('-m', '--model', type=str, default="./woman_csmsc",
+ # required=True,
+ help='Model name')
+ 
+ args = parser.parse_args()
+ model_dir = os.path.join("./logs", args.model)
+
+ if not os.path.exists(model_dir):
+ os.makedirs(model_dir)
+
+ config_path = args.config
+ config_save_path = os.path.join(model_dir, "config.json")
+ if init:
+ with open(config_path, "r") as f:
+ data = f.read()
+ with open(config_save_path, "w") as f:
+ f.write(data)
+ else:
+ with open(config_save_path, "r") as f:
+ data = f.read()
+ config = json.loads(data)
+ 
+ hparams = HParams(**config)
+ hparams.model_dir = model_dir
+ return hparams
+
+
+def get_hparams_from_dir(model_dir):
+ config_save_path = os.path.join(model_dir, "config.json")
+ with open(config_save_path, "r") as f:
+ data = f.read()
+ config = json.loads(data)
+
+ hparams =HParams(**config)
+ hparams.model_dir = model_dir
+ return hparams
+
+
+def get_hparams_from_file(config_path):
+ with open(config_path, "r") as f:
+ data = f.read()
+ config = json.loads(data)
+
+ hparams =HParams(**config)
+ return hparams
+
+
+def check_git_hash(model_dir):
+ source_dir = os.path.dirname(os.path.realpath(__file__))
+ if not os.path.exists(os.path.join(source_dir, ".git")):
+ logger.warn("{} is not a git repository, therefore hash value comparison will be ignored.".format(
+ source_dir
+ ))
+ return
+
+ cur_hash = subprocess.getoutput("git rev-parse HEAD")
+
+ path = os.path.join(model_dir, "githash")
+ if os.path.exists(path):
+ saved_hash = open(path).read()
+ if saved_hash != cur_hash:
+ logger.warn("git hash values are different. {}(saved) != {}(current)".format(
+ saved_hash[:8], cur_hash[:8]))
+ else:
+ open(path, "w").write(cur_hash)
+
+
+def get_logger(model_dir, filename="train.log"):
+ global logger
+ logger = logging.getLogger(os.path.basename(model_dir))
+ logger.setLevel(logging.DEBUG)
+ 
+ formatter = logging.Formatter("%(asctime)s\t%(name)s\t%(levelname)s\t%(message)s")
+ if not os.path.exists(model_dir):
+ os.makedirs(model_dir)
+ h = logging.FileHandler(os.path.join(model_dir, filename))
+ h.setLevel(logging.DEBUG)
+ h.setFormatter(formatter)
+ logger.addHandler(h)
+ return logger
+
+
+class HParams():
+ def __init__(self, **kwargs):
+ for k, v in kwargs.items():
+ if type(v) == dict:
+ v = HParams(**v)
+ self[k] = v
+ 
+ def keys(self):
+ return self.__dict__.keys()
+
+ def items(self):
+ return self.__dict__.items()
+
+ def values(self):
+ return self.__dict__.values()
+
+ def __len__(self):
+ return len(self.__dict__)
+
+ def __getitem__(self, key):
+ return getattr(self, key)
+
+ def __setitem__(self, key, value):
+ return setattr(self, key, value)
+
+ def __contains__(self, key):
+ return key in self.__dict__
+
+ def __repr__(self):
+ return self.__dict__.__repr__()