Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/8973072 * [to #41669377] docs and tools refinement and release 1. add build_doc linter script 2. add sphinx-docs support 3. add development doc and api doc 4. change version to 0.1.0 for the first internal release version Link: https://code.aone.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/8775307 * [to #41669377] add pipeline tutorial and fix bugs 1. add pipleine tutorial 2. fix bugs when using pipeline with certain model and preprocessor Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/8814301 * refine doc * feat: add audio aec pipeline and preprocessor * feat: add audio aec model classes * feat: add audio aec loss functions * refactor:delete no longer used loss function * [to #42281043] support kwargs in pipeline Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/8949062 * support kwargs in pipeline * update develop doc with CR instruction * Merge branch 'release/0.1' into dev/aec * style: reformat code by pre-commit tools * feat:support maas_lib pipeline auto downloading model * test:add aec test case as sample code * feat:aec pipeline use config from maashub * feat:aec pipeline use feature parameters from maashub * update setup.cfg to disable PEP8 rule W503 in flake8 and yapf * format:fix double quoted strings, indent issues and optimize import * refactor:extract some constant in aec pipeline * refactor: delete no longer used __main__ statement * chore:change all Chinese comments to English * fix: change file name style to lower case * refactor: rename model name * feat:load C++ .so from LD_LIBRARY_PATH * feat:register PROPROCESSOR for LinearAECAndFbank * refactory:move aec process from postprocess() to forward() and update comments * refactory:add more readable error message when audio sample rate is not 16000 * fix: package maas_lib renamed to modelscope in import statement * feat: optimize the error message of audio layer classes * format: delete empty lines * refactor: rename audio preprocessor and optimize error message * refactor: change aec model id to damo/speech_dfsmn_aec_psm_16k * refactor: change sample audio file url to public oss * Merge branch 'master' into dev/aec * feat: add output info for aec pipeline * fix: normalize output audio data to [-1.0, 1.0] * refactor:use constant from ModelFile * feat: AEC pipeline can use c++ lib in current working directory and the test will download it * fix: c++ downloading should work wherever test is triggerdmaster
@@ -0,0 +1,60 @@ | |||
import torch.nn as nn | |||
from .layer_base import LayerBase | |||
class RectifiedLinear(LayerBase): | |||
def __init__(self, input_dim, output_dim): | |||
super(RectifiedLinear, self).__init__() | |||
self.dim = input_dim | |||
self.relu = nn.ReLU() | |||
def forward(self, input): | |||
return self.relu(input) | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<RectifiedLinear> %d %d\n' % (self.dim, self.dim) | |||
return re_str | |||
def load_kaldi_nnet(self, instr): | |||
return instr | |||
class LogSoftmax(LayerBase): | |||
def __init__(self, input_dim, output_dim): | |||
super(LogSoftmax, self).__init__() | |||
self.dim = input_dim | |||
self.ls = nn.LogSoftmax() | |||
def forward(self, input): | |||
return self.ls(input) | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<Softmax> %d %d\n' % (self.dim, self.dim) | |||
return re_str | |||
def load_kaldi_nnet(self, instr): | |||
return instr | |||
class Sigmoid(LayerBase): | |||
def __init__(self, input_dim, output_dim): | |||
super(Sigmoid, self).__init__() | |||
self.dim = input_dim | |||
self.sig = nn.Sigmoid() | |||
def forward(self, input): | |||
return self.sig(input) | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<Sigmoid> %d %d\n' % (self.dim, self.dim) | |||
return re_str | |||
def load_kaldi_nnet(self, instr): | |||
return instr |
@@ -0,0 +1,78 @@ | |||
import numpy as np | |||
import torch as th | |||
import torch.nn as nn | |||
from .layer_base import (LayerBase, expect_kaldi_matrix, expect_token_number, | |||
to_kaldi_matrix) | |||
class AffineTransform(LayerBase): | |||
def __init__(self, input_dim, output_dim): | |||
super(AffineTransform, self).__init__() | |||
self.input_dim = input_dim | |||
self.output_dim = output_dim | |||
self.linear = nn.Linear(input_dim, output_dim) | |||
def forward(self, input): | |||
return self.linear(input) | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<AffineTransform> %d %d\n' % (self.output_dim, | |||
self.input_dim) | |||
re_str += '<LearnRateCoef> 1 <BiasLearnRateCoef> 1 <MaxNorm> 0\n' | |||
linear_weights = self.state_dict()['linear.weight'] | |||
x = linear_weights.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
linear_bias = self.state_dict()['linear.bias'] | |||
x = linear_bias.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
return re_str | |||
def to_raw_nnet(self, fid): | |||
linear_weights = self.state_dict()['linear.weight'] | |||
x = linear_weights.squeeze().numpy() | |||
x.tofile(fid) | |||
linear_bias = self.state_dict()['linear.bias'] | |||
x = linear_bias.squeeze().numpy() | |||
x.tofile(fid) | |||
def load_kaldi_nnet(self, instr): | |||
output = expect_token_number( | |||
instr, | |||
'<LearnRateCoef>', | |||
) | |||
if output is None: | |||
raise Exception('AffineTransform format error for <LearnRateCoef>') | |||
instr, lr = output | |||
output = expect_token_number(instr, '<BiasLearnRateCoef>') | |||
if output is None: | |||
raise Exception( | |||
'AffineTransform format error for <BiasLearnRateCoef>') | |||
instr, lr = output | |||
output = expect_token_number(instr, '<MaxNorm>') | |||
if output is None: | |||
raise Exception('AffineTransform format error for <MaxNorm>') | |||
instr, lr = output | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('AffineTransform format error for parsing matrix') | |||
instr, mat = output | |||
print(mat.shape) | |||
self.linear.weight = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('AffineTransform format error for parsing matrix') | |||
instr, mat = output | |||
mat = np.squeeze(mat) | |||
self.linear.bias = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
return instr |
@@ -0,0 +1,178 @@ | |||
import numpy as np | |||
import torch as th | |||
import torch.nn as nn | |||
import torch.nn.functional as F | |||
from .layer_base import (LayerBase, expect_kaldi_matrix, expect_token_number, | |||
to_kaldi_matrix) | |||
class DeepFsmn(LayerBase): | |||
def __init__(self, | |||
input_dim, | |||
output_dim, | |||
lorder=None, | |||
rorder=None, | |||
hidden_size=None, | |||
layer_norm=False, | |||
dropout=0): | |||
super(DeepFsmn, self).__init__() | |||
self.input_dim = input_dim | |||
self.output_dim = output_dim | |||
if lorder is None: | |||
return | |||
self.lorder = lorder | |||
self.rorder = rorder | |||
self.hidden_size = hidden_size | |||
self.layer_norm = layer_norm | |||
self.linear = nn.Linear(input_dim, hidden_size) | |||
self.norm = nn.LayerNorm(hidden_size) | |||
self.drop1 = nn.Dropout(p=dropout) | |||
self.drop2 = nn.Dropout(p=dropout) | |||
self.project = nn.Linear(hidden_size, output_dim, bias=False) | |||
self.conv1 = nn.Conv2d( | |||
output_dim, | |||
output_dim, [lorder, 1], [1, 1], | |||
groups=output_dim, | |||
bias=False) | |||
self.conv2 = nn.Conv2d( | |||
output_dim, | |||
output_dim, [rorder, 1], [1, 1], | |||
groups=output_dim, | |||
bias=False) | |||
def forward(self, input): | |||
f1 = F.relu(self.linear(input)) | |||
f1 = self.drop1(f1) | |||
if self.layer_norm: | |||
f1 = self.norm(f1) | |||
p1 = self.project(f1) | |||
x = th.unsqueeze(p1, 1) | |||
x_per = x.permute(0, 3, 2, 1) | |||
y = F.pad(x_per, [0, 0, self.lorder - 1, 0]) | |||
yr = F.pad(x_per, [0, 0, 0, self.rorder]) | |||
yr = yr[:, :, 1:, :] | |||
out = x_per + self.conv1(y) + self.conv2(yr) | |||
out = self.drop2(out) | |||
out1 = out.permute(0, 3, 2, 1) | |||
return input + out1.squeeze() | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<UniDeepFsmn> %d %d\n'\ | |||
% (self.output_dim, self.input_dim) | |||
re_str += '<LearnRateCoef> %d <HidSize> %d <LOrder> %d <LStride> %d <MaxNorm> 0\n'\ | |||
% (1, self.hidden_size, self.lorder, 1) | |||
lfiters = self.state_dict()['conv1.weight'] | |||
x = np.flipud(lfiters.squeeze().numpy().T) | |||
re_str += to_kaldi_matrix(x) | |||
proj_weights = self.state_dict()['project.weight'] | |||
x = proj_weights.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
linear_weights = self.state_dict()['linear.weight'] | |||
x = linear_weights.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
linear_bias = self.state_dict()['linear.bias'] | |||
x = linear_bias.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
return re_str | |||
def load_kaldi_nnet(self, instr): | |||
output = expect_token_number( | |||
instr, | |||
'<LearnRateCoef>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LearnRateCoef>') | |||
instr, lr = output | |||
output = expect_token_number( | |||
instr, | |||
'<HidSize>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <HidSize>') | |||
instr, hiddensize = output | |||
self.hidden_size = int(hiddensize) | |||
output = expect_token_number( | |||
instr, | |||
'<LOrder>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LOrder>') | |||
instr, lorder = output | |||
self.lorder = int(lorder) | |||
output = expect_token_number( | |||
instr, | |||
'<LStride>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LStride>') | |||
instr, lstride = output | |||
self.lstride = lstride | |||
output = expect_token_number( | |||
instr, | |||
'<MaxNorm>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <MaxNorm>') | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
mat1 = np.fliplr(mat.T).copy() | |||
self.conv1 = nn.Conv2d( | |||
self.output_dim, | |||
self.output_dim, [self.lorder, 1], [1, 1], | |||
groups=self.output_dim, | |||
bias=False) | |||
mat_th = th.from_numpy(mat1).type(th.FloatTensor) | |||
mat_th = mat_th.unsqueeze(1) | |||
mat_th = mat_th.unsqueeze(3) | |||
self.conv1.weight = th.nn.Parameter(mat_th) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
self.project = nn.Linear(self.hidden_size, self.output_dim, bias=False) | |||
self.linear = nn.Linear(self.input_dim, self.hidden_size) | |||
self.project.weight = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
self.linear.weight = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
self.linear.bias = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
return instr |
@@ -0,0 +1,50 @@ | |||
import abc | |||
import re | |||
import numpy as np | |||
import torch.nn as nn | |||
def expect_token_number(instr, token): | |||
first_token = re.match(r'^\s*' + token, instr) | |||
if first_token is None: | |||
return None | |||
instr = instr[first_token.end():] | |||
lr = re.match(r'^\s*(-?\d+\.?\d*e?-?\d*?)', instr) | |||
if lr is None: | |||
return None | |||
return instr[lr.end():], lr.groups()[0] | |||
def expect_kaldi_matrix(instr): | |||
pos2 = instr.find('[', 0) | |||
pos3 = instr.find(']', pos2) | |||
mat = [] | |||
for stt in instr[pos2 + 1:pos3].split('\n'): | |||
tmp_mat = np.fromstring(stt, dtype=np.float32, sep=' ') | |||
if tmp_mat.size > 0: | |||
mat.append(tmp_mat) | |||
return instr[pos3 + 1:], np.array(mat) | |||
def to_kaldi_matrix(np_mat): | |||
""" | |||
function that transform as str numpy mat to standard kaldi str matrix | |||
:param np_mat: numpy mat | |||
:return: str | |||
""" | |||
np.set_printoptions(threshold=np.inf, linewidth=np.nan, suppress=True) | |||
out_str = str(np_mat) | |||
out_str = out_str.replace('[', '') | |||
out_str = out_str.replace(']', '') | |||
return '[ %s ]\n' % out_str | |||
class LayerBase(nn.Module, metaclass=abc.ABCMeta): | |||
def __init__(self): | |||
super(LayerBase, self).__init__() | |||
@abc.abstractmethod | |||
def to_kaldi_nnet(self): | |||
pass |
@@ -0,0 +1,482 @@ | |||
import numpy as np | |||
import torch as th | |||
import torch.nn as nn | |||
import torch.nn.functional as F | |||
from .layer_base import (LayerBase, expect_kaldi_matrix, expect_token_number, | |||
to_kaldi_matrix) | |||
class SepConv(nn.Module): | |||
def __init__(self, | |||
in_channels, | |||
filters, | |||
out_channels, | |||
kernel_size=(5, 2), | |||
dilation=(1, 1)): | |||
""" :param kernel_size (time, frequency) | |||
""" | |||
super(SepConv, self).__init__() | |||
# depthwise + pointwise | |||
self.dconv = nn.Conv2d( | |||
in_channels, | |||
in_channels * filters, | |||
kernel_size, | |||
dilation=dilation, | |||
groups=in_channels) | |||
self.pconv = nn.Conv2d( | |||
in_channels * filters, out_channels, kernel_size=1) | |||
self.padding = dilation[0] * (kernel_size[0] - 1) | |||
def forward(self, input): | |||
''' input: [B, C, T, F] | |||
''' | |||
x = F.pad(input, [0, 0, self.padding, 0]) | |||
x = self.dconv(x) | |||
x = self.pconv(x) | |||
return x | |||
class Conv2d(nn.Module): | |||
def __init__(self, | |||
input_dim, | |||
output_dim, | |||
lorder=20, | |||
rorder=0, | |||
groups=1, | |||
bias=False, | |||
skip_connect=True): | |||
super(Conv2d, self).__init__() | |||
self.lorder = lorder | |||
self.conv = nn.Conv2d( | |||
input_dim, output_dim, [lorder, 1], groups=groups, bias=bias) | |||
self.rorder = rorder | |||
if self.rorder: | |||
self.conv2 = nn.Conv2d( | |||
input_dim, output_dim, [rorder, 1], groups=groups, bias=bias) | |||
self.skip_connect = skip_connect | |||
def forward(self, input): | |||
# [B, 1, T, F] | |||
x = th.unsqueeze(input, 1) | |||
# [B, F, T, 1] | |||
x_per = x.permute(0, 3, 2, 1) | |||
y = F.pad(x_per, [0, 0, self.lorder - 1, 0]) | |||
out = self.conv(y) | |||
if self.rorder: | |||
yr = F.pad(x_per, [0, 0, 0, self.rorder]) | |||
yr = yr[:, :, 1:, :] | |||
out += self.conv2(yr) | |||
out = out.permute(0, 3, 2, 1).squeeze(1) | |||
if self.skip_connect: | |||
out = out + input | |||
return out | |||
class SelfAttLayer(nn.Module): | |||
def __init__(self, input_dim, output_dim, lorder=None, hidden_size=None): | |||
super(SelfAttLayer, self).__init__() | |||
self.input_dim = input_dim | |||
self.output_dim = output_dim | |||
if lorder is None: | |||
return | |||
self.lorder = lorder | |||
self.hidden_size = hidden_size | |||
self.linear = nn.Linear(input_dim, hidden_size) | |||
self.project = nn.Linear(hidden_size, output_dim, bias=False) | |||
self.att = nn.Linear(input_dim, lorder, bias=False) | |||
def forward(self, input): | |||
f1 = F.relu(self.linear(input)) | |||
p1 = self.project(f1) | |||
x = th.unsqueeze(p1, 1) | |||
x_per = x.permute(0, 3, 2, 1) | |||
y = F.pad(x_per, [0, 0, self.lorder - 1, 0]) | |||
# z [B, F, T, lorder] | |||
z = x_per | |||
for i in range(1, self.lorder): | |||
z = th.cat([z, y[:, :, self.lorder - 1 - i:-i, :]], axis=-1) | |||
# [B, T, lorder] | |||
att = F.softmax(self.att(input), dim=-1) | |||
att = th.unsqueeze(att, 1) | |||
z = th.sum(z * att, axis=-1) | |||
out1 = z.permute(0, 2, 1) | |||
return input + out1 | |||
class TFFsmn(nn.Module): | |||
def __init__(self, | |||
input_dim, | |||
output_dim, | |||
lorder=None, | |||
hidden_size=None, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
skip_connect=True): | |||
super(TFFsmn, self).__init__() | |||
self.skip_connect = skip_connect | |||
self.linear = nn.Linear(input_dim, hidden_size) | |||
self.norm = nn.Identity() | |||
if layer_norm: | |||
self.norm = nn.LayerNorm(input_dim) | |||
self.act = nn.ReLU() | |||
self.project = nn.Linear(hidden_size, output_dim, bias=False) | |||
self.conv1 = nn.Conv2d( | |||
output_dim, | |||
output_dim, [lorder, 1], | |||
dilation=[dilation, 1], | |||
groups=output_dim, | |||
bias=False) | |||
self.padding_left = dilation * (lorder - 1) | |||
dorder = 5 | |||
self.conv2 = nn.Conv2d(1, 1, [dorder, 1], bias=False) | |||
self.padding_freq = dorder - 1 | |||
def forward(self, input): | |||
return self.compute1(input) | |||
def compute1(self, input): | |||
''' linear-dconv-relu(norm)-linear-dconv | |||
''' | |||
x = self.linear(input) | |||
# [B, 1, F, T] | |||
x = th.unsqueeze(x, 1).permute(0, 1, 3, 2) | |||
z = F.pad(x, [0, 0, self.padding_freq, 0]) | |||
z = self.conv2(z) + x | |||
x = z.permute(0, 3, 2, 1).squeeze(-1) | |||
x = self.act(x) | |||
x = self.norm(x) | |||
x = self.project(x) | |||
x = th.unsqueeze(x, 1).permute(0, 3, 2, 1) | |||
# [B, F, T+lorder-1, 1] | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.conv1(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
return input + out | |||
class CNNFsmn(nn.Module): | |||
''' use cnn to reduce parameters | |||
''' | |||
def __init__(self, | |||
input_dim, | |||
output_dim, | |||
lorder=None, | |||
hidden_size=None, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
skip_connect=True): | |||
super(CNNFsmn, self).__init__() | |||
self.input_dim = input_dim | |||
self.output_dim = output_dim | |||
self.skip_connect = skip_connect | |||
if lorder is None: | |||
return | |||
self.lorder = lorder | |||
self.hidden_size = hidden_size | |||
self.linear = nn.Linear(input_dim, hidden_size) | |||
self.act = nn.ReLU() | |||
kernel_size = (3, 8) | |||
stride = (1, 4) | |||
self.conv = nn.Sequential( | |||
nn.ConstantPad2d((stride[1], 0, kernel_size[0] - 1, 0), 0), | |||
nn.Conv2d(1, stride[1], kernel_size=kernel_size, stride=stride)) | |||
self.dconv = nn.Conv2d( | |||
output_dim, | |||
output_dim, [lorder, 1], | |||
dilation=[dilation, 1], | |||
groups=output_dim, | |||
bias=False) | |||
self.padding_left = dilation * (lorder - 1) | |||
def forward(self, input): | |||
return self.compute2(input) | |||
def compute1(self, input): | |||
''' linear-relu(norm)-conv2d-relu?-dconv | |||
''' | |||
# [B, T, F] | |||
x = self.linear(input) | |||
x = self.act(x) | |||
x = th.unsqueeze(x, 1) | |||
x = self.conv(x) | |||
# [B, C, T, F] -> [B, 1, T, F] | |||
b, c, t, f = x.shape | |||
x = x.view([b, 1, t, -1]) | |||
x = x.permute(0, 3, 2, 1) | |||
# [B, F, T+lorder-1, 1] | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.dconv(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
return input + out | |||
def compute2(self, input): | |||
''' conv2d-relu-linear-relu?-dconv | |||
''' | |||
x = th.unsqueeze(input, 1) | |||
x = self.conv(x) | |||
x = self.act(x) | |||
# [B, C, T, F] -> [B, T, F] | |||
b, c, t, f = x.shape | |||
x = x.view([b, t, -1]) | |||
x = self.linear(x) | |||
x = th.unsqueeze(x, 1).permute(0, 3, 2, 1) | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.dconv(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
return input + out | |||
class UniDeepFsmn(LayerBase): | |||
def __init__(self, | |||
input_dim, | |||
output_dim, | |||
lorder=None, | |||
hidden_size=None, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
skip_connect=True): | |||
super(UniDeepFsmn, self).__init__() | |||
self.input_dim = input_dim | |||
self.output_dim = output_dim | |||
self.skip_connect = skip_connect | |||
if lorder is None: | |||
return | |||
self.lorder = lorder | |||
self.hidden_size = hidden_size | |||
self.linear = nn.Linear(input_dim, hidden_size) | |||
self.norm = nn.Identity() | |||
if layer_norm: | |||
self.norm = nn.LayerNorm(input_dim) | |||
self.act = nn.ReLU() | |||
self.project = nn.Linear(hidden_size, output_dim, bias=False) | |||
self.conv1 = nn.Conv2d( | |||
output_dim, | |||
output_dim, [lorder, 1], | |||
dilation=[dilation, 1], | |||
groups=output_dim, | |||
bias=False) | |||
self.padding_left = dilation * (lorder - 1) | |||
def forward(self, input): | |||
return self.compute1(input) | |||
def compute1(self, input): | |||
''' linear-relu(norm)-linear-dconv | |||
''' | |||
# [B, T, F] | |||
x = self.linear(input) | |||
x = self.act(x) | |||
x = self.norm(x) | |||
x = self.project(x) | |||
x = th.unsqueeze(x, 1).permute(0, 3, 2, 1) | |||
# [B, F, T+lorder-1, 1] | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.conv1(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
return input + out | |||
def compute2(self, input): | |||
''' linear-dconv-linear-relu(norm) | |||
''' | |||
x = self.project(input) | |||
x = th.unsqueeze(x, 1).permute(0, 3, 2, 1) | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.conv1(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
x = self.linear(out) | |||
x = self.act(x) | |||
x = self.norm(x) | |||
return input + x | |||
def compute3(self, input): | |||
''' dconv-linear-relu(norm)-linear | |||
''' | |||
x = th.unsqueeze(input, 1).permute(0, 3, 2, 1) | |||
y = F.pad(x, [0, 0, self.padding_left, 0]) | |||
out = self.conv1(y) | |||
if self.skip_connect: | |||
out = out + x | |||
out = out.permute(0, 3, 2, 1).squeeze() | |||
x = self.linear(out) | |||
x = self.act(x) | |||
x = self.norm(x) | |||
x = self.project(x) | |||
return input + x | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<UniDeepFsmn> %d %d\n' \ | |||
% (self.output_dim, self.input_dim) | |||
re_str += '<LearnRateCoef> %d <HidSize> %d <LOrder> %d <LStride> %d <MaxNorm> 0\n' \ | |||
% (1, self.hidden_size, self.lorder, 1) | |||
lfiters = self.state_dict()['conv1.weight'] | |||
x = np.flipud(lfiters.squeeze().numpy().T) | |||
re_str += to_kaldi_matrix(x) | |||
proj_weights = self.state_dict()['project.weight'] | |||
x = proj_weights.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
linear_weights = self.state_dict()['linear.weight'] | |||
x = linear_weights.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
linear_bias = self.state_dict()['linear.bias'] | |||
x = linear_bias.squeeze().numpy() | |||
re_str += to_kaldi_matrix(x) | |||
return re_str | |||
def to_raw_nnet(self, fid): | |||
lfiters = self.state_dict()['conv1.weight'] | |||
x = np.flipud(lfiters.squeeze().numpy().T) | |||
x.tofile(fid) | |||
proj_weights = self.state_dict()['project.weight'] | |||
x = proj_weights.squeeze().numpy() | |||
x.tofile(fid) | |||
linear_weights = self.state_dict()['linear.weight'] | |||
x = linear_weights.squeeze().numpy() | |||
x.tofile(fid) | |||
linear_bias = self.state_dict()['linear.bias'] | |||
x = linear_bias.squeeze().numpy() | |||
x.tofile(fid) | |||
def load_kaldi_nnet(self, instr): | |||
output = expect_token_number( | |||
instr, | |||
'<LearnRateCoef>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LearnRateCoef>') | |||
instr, lr = output | |||
output = expect_token_number( | |||
instr, | |||
'<HidSize>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <HidSize>') | |||
instr, hiddensize = output | |||
self.hidden_size = int(hiddensize) | |||
output = expect_token_number( | |||
instr, | |||
'<LOrder>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LOrder>') | |||
instr, lorder = output | |||
self.lorder = int(lorder) | |||
output = expect_token_number( | |||
instr, | |||
'<LStride>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <LStride>') | |||
instr, lstride = output | |||
self.lstride = lstride | |||
output = expect_token_number( | |||
instr, | |||
'<MaxNorm>', | |||
) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for <MaxNorm>') | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
mat1 = np.fliplr(mat.T).copy() | |||
self.conv1 = nn.Conv2d( | |||
self.output_dim, | |||
self.output_dim, [self.lorder, 1], [1, 1], | |||
groups=self.output_dim, | |||
bias=False) | |||
mat_th = th.from_numpy(mat1).type(th.FloatTensor) | |||
mat_th = mat_th.unsqueeze(1) | |||
mat_th = mat_th.unsqueeze(3) | |||
self.conv1.weight = th.nn.Parameter(mat_th) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
self.project = nn.Linear(self.hidden_size, self.output_dim, bias=False) | |||
self.linear = nn.Linear(self.input_dim, self.hidden_size) | |||
self.project.weight = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
self.linear.weight = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
output = expect_kaldi_matrix(instr) | |||
if output is None: | |||
raise Exception('UniDeepFsmn format error for parsing matrix') | |||
instr, mat = output | |||
mat = np.squeeze(mat) | |||
self.linear.bias = th.nn.Parameter( | |||
th.from_numpy(mat).type(th.FloatTensor)) | |||
return instr |
@@ -0,0 +1,394 @@ | |||
import torch | |||
import torch.nn.functional as F | |||
from .modulation_loss import (GaborSTRFConv, MelScale, | |||
ModulationDomainLossModule) | |||
EPS = 1e-8 | |||
def compute_mask(mixed_spec, clean_spec, mask_type='psmiam', clip=1): | |||
''' | |||
stft: (batch, ..., 2) or complex(batch, ...) | |||
y = x + n | |||
''' | |||
if torch.is_complex(mixed_spec): | |||
yr, yi = mixed_spec.real, mixed_spec.imag | |||
else: | |||
yr, yi = mixed_spec[..., 0], mixed_spec[..., 1] | |||
if torch.is_complex(clean_spec): | |||
xr, xi = clean_spec.real, clean_spec.imag | |||
else: | |||
xr, xi = clean_spec[..., 0], clean_spec[..., 1] | |||
if mask_type == 'iam': | |||
ymag = torch.sqrt(yr**2 + yi**2) | |||
xmag = torch.sqrt(xr**2 + xi**2) | |||
iam = xmag / (ymag + EPS) | |||
return torch.clamp(iam, 0, 1) | |||
elif mask_type == 'psm': | |||
ypow = yr**2 + yi**2 | |||
psm = (xr * yr + xi * yi) / (ypow + EPS) | |||
return torch.clamp(psm, 0, 1) | |||
elif mask_type == 'psmiam': | |||
ypow = yr**2 + yi**2 | |||
psm = (xr * yr + xi * yi) / (ypow + EPS) | |||
ymag = torch.sqrt(yr**2 + yi**2) | |||
xmag = torch.sqrt(xr**2 + xi**2) | |||
iam = xmag / (ymag + EPS) | |||
psmiam = psm * iam | |||
return torch.clamp(psmiam, 0, 1) | |||
elif mask_type == 'crm': | |||
ypow = yr**2 + yi**2 | |||
mr = (xr * yr + xi * yi) / (ypow + EPS) | |||
mi = (xi * yr - xr * yi) / (ypow + EPS) | |||
mr = torch.clamp(mr, -clip, clip) | |||
mi = torch.clamp(mi, -clip, clip) | |||
return mr, mi | |||
def energy_vad(spec, | |||
thdhigh=320 * 600 * 600 * 2, | |||
thdlow=320 * 300 * 300 * 2, | |||
int16=True): | |||
''' | |||
energy based vad should be accurate enough | |||
spec: (batch, bins, frames, 2) | |||
returns (batch, frames) | |||
''' | |||
energy = torch.sum(spec[..., 0]**2 + spec[..., 1]**2, dim=1) | |||
vad = energy > thdhigh | |||
idx = torch.logical_and(vad == 0, energy > thdlow) | |||
vad[idx] = 0.5 | |||
return vad | |||
def modulation_loss_init(n_fft): | |||
gabor_strf_parameters = torch.load( | |||
'./network/gabor_strf_parameters.pt')['state_dict'] | |||
gabor_modulation_kernels = GaborSTRFConv(supn=30, supk=30, nkern=60) | |||
gabor_modulation_kernels.load_state_dict(gabor_strf_parameters) | |||
modulation_loss_module = ModulationDomainLossModule( | |||
gabor_modulation_kernels.eval()) | |||
for param in modulation_loss_module.parameters(): | |||
param.requires_grad = False | |||
stft2mel = MelScale( | |||
n_mels=80, sample_rate=16000, n_stft=n_fft // 2 + 1).cuda() | |||
return modulation_loss_module, stft2mel | |||
def mask_loss_function( | |||
loss_func='psm_loss', | |||
loss_type='mse', # ['mse', 'mae', 'comb'] | |||
mask_type='psmiam', | |||
use_mod_loss=False, | |||
use_wav2vec_loss=False, | |||
n_fft=640, | |||
hop_length=320, | |||
EPS=1e-8, | |||
weight=None): | |||
if weight is not None: | |||
print(f'Use loss weight: {weight}') | |||
winlen = n_fft | |||
window = torch.hamming_window(winlen, periodic=False) | |||
def stft(x, return_complex=False): | |||
# returns [batch, bins, frames, 2] | |||
return torch.stft( | |||
x, | |||
n_fft, | |||
hop_length, | |||
winlen, | |||
window=window.to(x.device), | |||
center=False, | |||
return_complex=return_complex) | |||
def istft(x, slen): | |||
return torch.istft( | |||
x, | |||
n_fft, | |||
hop_length, | |||
winlen, | |||
window=window.to(x.device), | |||
center=False, | |||
length=slen) | |||
def mask_loss(targets, masks, nframes): | |||
''' [Batch, Time, Frequency] | |||
''' | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(targets) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
masks = masks * mask_for_loss | |||
targets = targets * mask_for_loss | |||
if weight is None: | |||
alpha = 1 | |||
else: # for aec ST | |||
alpha = weight - targets | |||
if loss_type == 'mse': | |||
loss = 0.5 * torch.sum(alpha * torch.pow(targets - masks, 2)) | |||
elif loss_type == 'mae': | |||
loss = torch.sum(alpha * torch.abs(targets - masks)) | |||
else: # mse(mask), mae(mask) approx 1:2 | |||
loss = 0.5 * torch.sum(alpha * torch.pow(targets - masks, 2) | |||
+ 0.1 * alpha * torch.abs(targets - masks)) | |||
loss /= torch.sum(nframes) | |||
return loss | |||
def spectrum_loss(targets, spec, nframes): | |||
''' [Batch, Time, Frequency, 2] | |||
''' | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(targets[..., 0]) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
xr = spec[..., 0] * mask_for_loss | |||
xi = spec[..., 1] * mask_for_loss | |||
yr = targets[..., 0] * mask_for_loss | |||
yi = targets[..., 1] * mask_for_loss | |||
xmag = torch.sqrt(spec[..., 0]**2 + spec[..., 1]**2) * mask_for_loss | |||
ymag = torch.sqrt(targets[..., 0]**2 | |||
+ targets[..., 1]**2) * mask_for_loss | |||
loss1 = torch.sum(torch.pow(xr - yr, 2) + torch.pow(xi - yi, 2)) | |||
loss2 = torch.sum(torch.pow(xmag - ymag, 2)) | |||
loss = (loss1 + loss2) / torch.sum(nframes) | |||
return loss | |||
def sa_loss_dlen(mixed, clean, masks, nframes): | |||
yspec = stft(mixed).permute([0, 2, 1, 3]) / 32768 | |||
xspec = stft(clean).permute([0, 2, 1, 3]) / 32768 | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(xspec[..., 0]) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
emag = ((yspec[..., 0]**2 + yspec[..., 1]**2)**0.15) * (masks**0.3) | |||
xmag = (xspec[..., 0]**2 + xspec[..., 1]**2)**0.15 | |||
emag = emag * mask_for_loss | |||
xmag = xmag * mask_for_loss | |||
loss = torch.sum(torch.pow(emag - xmag, 2)) / torch.sum(nframes) | |||
return loss | |||
def psm_vad_loss_dlen(mixed, clean, masks, nframes, subtask=None): | |||
mixed_spec = stft(mixed) | |||
clean_spec = stft(clean) | |||
targets = compute_mask(mixed_spec, clean_spec, mask_type) | |||
# [B, T, F] | |||
targets = targets.permute(0, 2, 1) | |||
loss = mask_loss(targets, masks, nframes) | |||
if subtask is not None: | |||
vadtargets = energy_vad(clean_spec) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(targets[:, :, 0]) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:] = 0 | |||
subtask = subtask[:, :, 0] * mask_for_loss | |||
vadtargets = vadtargets * mask_for_loss | |||
loss_vad = F.binary_cross_entropy(subtask, vadtargets) | |||
return loss + loss_vad | |||
return loss | |||
def modulation_loss(mixed, clean, masks, nframes, subtask=None): | |||
mixed_spec = stft(mixed, True) | |||
clean_spec = stft(clean, True) | |||
enhanced_mag = torch.abs(mixed_spec) | |||
clean_mag = torch.abs(clean_spec) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(clean_mag) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, :, num:] = 0 | |||
clean_mag = clean_mag * mask_for_loss | |||
enhanced_mag = enhanced_mag * mask_for_loss * masks.permute([0, 2, 1]) | |||
# Covert to log-mel representation | |||
# (B,T,#mel_channels) | |||
clean_log_mel = torch.log( | |||
torch.transpose(stft2mel(clean_mag**2), 2, 1) + 1e-8) | |||
enhanced_log_mel = torch.log( | |||
torch.transpose(stft2mel(enhanced_mag**2), 2, 1) + 1e-8) | |||
alpha = compute_mask(mixed_spec, clean_spec, mask_type) | |||
alpha = alpha.permute(0, 2, 1) | |||
loss = 0.05 * modulation_loss_module(enhanced_log_mel, clean_log_mel, | |||
alpha) | |||
loss2 = psm_vad_loss_dlen(mixed, clean, masks, nframes, subtask) | |||
# print(loss.item(), loss2.item()) #approx 1:4 | |||
loss = loss + loss2 | |||
return loss | |||
def wav2vec_loss(mixed, clean, masks, nframes, subtask=None): | |||
mixed /= 32768 | |||
clean /= 32768 | |||
mixed_spec = stft(mixed) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(masks) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
masks_est = masks * mask_for_loss | |||
estimate = mixed_spec * masks_est.permute([0, 2, 1]).unsqueeze(3) | |||
est_clean = istft(estimate, clean.shape[1]) | |||
loss = wav2vec_loss_module(est_clean, clean) | |||
return loss | |||
def sisdr_loss_dlen(mixed, | |||
clean, | |||
masks, | |||
nframes, | |||
subtask=None, | |||
zero_mean=True): | |||
mixed_spec = stft(mixed) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(masks) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
masks_est = masks * mask_for_loss | |||
estimate = mixed_spec * masks_est.permute([0, 2, 1]).unsqueeze(3) | |||
est_clean = istft(estimate, clean.shape[1]) | |||
flen = min(clean.shape[1], est_clean.shape[1]) | |||
clean = clean[:, :flen] | |||
est_clean = est_clean[:, :flen] | |||
# follow asteroid/losses/sdr.py | |||
if zero_mean: | |||
clean = clean - torch.mean(clean, dim=1, keepdim=True) | |||
est_clean = est_clean - torch.mean(est_clean, dim=1, keepdim=True) | |||
dot = torch.sum(est_clean * clean, dim=1, keepdim=True) | |||
s_clean_energy = torch.sum(clean**2, dim=1, keepdim=True) + EPS | |||
scaled_clean = dot * clean / s_clean_energy | |||
e_noise = est_clean - scaled_clean | |||
# [batch] | |||
sisdr = torch.sum( | |||
scaled_clean**2, dim=1) / ( | |||
torch.sum(e_noise**2, dim=1) + EPS) | |||
sisdr = -10 * torch.log10(sisdr + EPS) | |||
loss = sisdr.mean() | |||
return loss | |||
def sisdr_freq_loss_dlen(mixed, clean, masks, nframes, subtask=None): | |||
mixed_spec = stft(mixed) | |||
clean_spec = stft(clean) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(masks) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
masks_est = masks * mask_for_loss | |||
estimate = mixed_spec * masks_est.permute([0, 2, 1]).unsqueeze(3) | |||
dot_real = estimate[..., 0] * clean_spec[..., 0] + \ | |||
estimate[..., 1] * clean_spec[..., 1] | |||
dot_imag = estimate[..., 0] * clean_spec[..., 1] - \ | |||
estimate[..., 1] * clean_spec[..., 0] | |||
dot = torch.cat([dot_real.unsqueeze(3), dot_imag.unsqueeze(3)], dim=-1) | |||
s_clean_energy = clean_spec[..., 0] ** 2 + \ | |||
clean_spec[..., 1] ** 2 + EPS | |||
scaled_clean = dot * clean_spec / s_clean_energy.unsqueeze(3) | |||
e_noise = estimate - scaled_clean | |||
# [batch] | |||
scaled_clean_energy = torch.sum( | |||
scaled_clean[..., 0]**2 + scaled_clean[..., 1]**2, dim=1) | |||
e_noise_energy = torch.sum( | |||
e_noise[..., 0]**2 + e_noise[..., 1]**2, dim=1) | |||
sisdr = torch.sum( | |||
scaled_clean_energy, dim=1) / ( | |||
torch.sum(e_noise_energy, dim=1) + EPS) | |||
sisdr = -10 * torch.log10(sisdr + EPS) | |||
loss = sisdr.mean() | |||
return loss | |||
def crm_loss_dlen(mixed, clean, masks, nframes, subtask=None): | |||
mixed_spec = stft(mixed).permute([0, 2, 1, 3]) | |||
clean_spec = stft(clean).permute([0, 2, 1, 3]) | |||
mixed_spec = mixed_spec / 32768 | |||
clean_spec = clean_spec / 32768 | |||
tgt_mr, tgt_mi = compute_mask(mixed_spec, clean_spec, mask_type='crm') | |||
D = int(masks.shape[2] / 2) | |||
with torch.no_grad(): | |||
mask_for_loss = torch.ones_like(clean_spec[..., 0]) | |||
for idx, num in enumerate(nframes): | |||
mask_for_loss[idx, num:, :] = 0 | |||
mr = masks[..., :D] * mask_for_loss | |||
mi = masks[..., D:] * mask_for_loss | |||
tgt_mr = tgt_mr * mask_for_loss | |||
tgt_mi = tgt_mi * mask_for_loss | |||
if weight is None: | |||
alpha = 1 | |||
else: | |||
alpha = weight - tgt_mr | |||
# signal approximation | |||
yr = mixed_spec[..., 0] | |||
yi = mixed_spec[..., 1] | |||
loss1 = torch.sum(alpha * torch.pow((mr * yr - mi * yi) - clean_spec[..., 0], 2)) \ | |||
+ torch.sum(alpha * torch.pow((mr * yi + mi * yr) - clean_spec[..., 1], 2)) | |||
# mask approximation | |||
loss2 = torch.sum(alpha * torch.pow(mr - tgt_mr, 2)) \ | |||
+ torch.sum(alpha * torch.pow(mi - tgt_mi, 2)) | |||
loss = 0.5 * (loss1 + loss2) / torch.sum(nframes) | |||
return loss | |||
def crm_miso_loss_dlen(mixed, clean, masks, nframes): | |||
return crm_loss_dlen(mixed[..., 0], clean[..., 0], masks, nframes) | |||
def mimo_loss_dlen(mixed, clean, masks, nframes): | |||
chs = mixed.shape[-1] | |||
D = masks.shape[2] // chs | |||
loss = psm_vad_loss_dlen(mixed[..., 0], clean[..., 0], masks[..., :D], | |||
nframes) | |||
for ch in range(1, chs): | |||
loss1 = psm_vad_loss_dlen(mixed[..., ch], clean[..., ch], | |||
masks[..., ch * D:ch * D + D], nframes) | |||
loss = loss + loss1 | |||
return loss / chs | |||
def spec_loss_dlen(mixed, clean, spec, nframes): | |||
clean_spec = stft(clean).permute([0, 2, 1, 3]) | |||
clean_spec = clean_spec / 32768 | |||
D = spec.shape[2] // 2 | |||
spec_est = torch.cat([spec[..., :D, None], spec[..., D:, None]], | |||
dim=-1) | |||
loss = spectrum_loss(clean_spec, spec_est, nframes) | |||
return loss | |||
if loss_func == 'psm_vad_loss_dlen': | |||
return psm_vad_loss_dlen | |||
elif loss_func == 'sisdr_loss_dlen': | |||
return sisdr_loss_dlen | |||
elif loss_func == 'sisdr_freq_loss_dlen': | |||
return sisdr_freq_loss_dlen | |||
elif loss_func == 'crm_loss_dlen': | |||
return crm_loss_dlen | |||
elif loss_func == 'modulation_loss': | |||
return modulation_loss | |||
elif loss_func == 'wav2vec_loss': | |||
return wav2vec_loss | |||
elif loss_func == 'mimo_loss_dlen': | |||
return mimo_loss_dlen | |||
elif loss_func == 'spec_loss_dlen': | |||
return spec_loss_dlen | |||
elif loss_func == 'sa_loss_dlen': | |||
return sa_loss_dlen | |||
else: | |||
print('error loss func') | |||
return None |
@@ -0,0 +1,248 @@ | |||
import math | |||
import torch | |||
import torch.nn as nn | |||
import torch.nn.functional as F | |||
from torchaudio.transforms import MelScale | |||
class ModulationDomainLossModule(torch.nn.Module): | |||
"""Modulation-domain loss function developed in [1] for supervised speech enhancement | |||
In our paper, we used the gabor-based STRF kernels as the modulation kernels and used the log-mel spectrogram | |||
as the input spectrogram representation. | |||
Specific parameter details are in the paper and in the example below | |||
Parameters | |||
---------- | |||
modulation_kernels: nn.Module | |||
Differentiable module that transforms a spectrogram representation to the modulation domain | |||
modulation_domain = modulation_kernels(input_tf_representation) | |||
Input Spectrogram representation (B, T, F) ---> |(M) modulation_kernels|--->Modulation Domain(B, M, T', F') | |||
norm: boolean | |||
Normalizes the modulation domain representation to be 0 mean across time | |||
[1] T. Vuong, Y. Xia, and R. M. Stern, “A modulation-domain lossfor neural-network-based real-time | |||
speech enhancement” | |||
Accepted ICASSP 2021, https://arxiv.org/abs/2102.07330 | |||
""" | |||
def __init__(self, modulation_kernels, norm=True): | |||
super(ModulationDomainLossModule, self).__init__() | |||
self.modulation_kernels = modulation_kernels | |||
self.mse = nn.MSELoss(reduce=False) | |||
self.norm = norm | |||
def forward(self, enhanced_spect, clean_spect, weight=None): | |||
"""Calculate modulation-domain loss | |||
Args: | |||
enhanced_spect (Tensor): spectrogram representation of enhanced signal (B, #frames, #freq_channels). | |||
clean_spect (Tensor): spectrogram representation of clean ground-truth signal (B, #frames, #freq_channels). | |||
Returns: | |||
Tensor: Modulation-domain loss value. | |||
""" | |||
clean_mod = self.modulation_kernels(clean_spect) | |||
enhanced_mod = self.modulation_kernels(enhanced_spect) | |||
if self.norm: | |||
mean_clean_mod = torch.mean(clean_mod, dim=2) | |||
mean_enhanced_mod = torch.mean(enhanced_mod, dim=2) | |||
clean_mod = clean_mod - mean_clean_mod.unsqueeze(2) | |||
enhanced_mod = enhanced_mod - mean_enhanced_mod.unsqueeze(2) | |||
if weight is None: | |||
alpha = 1 | |||
else: # TF-mask weight | |||
alpha = 1 + torch.sum(weight, dim=-1, keepdim=True).unsqueeze(1) | |||
mod_mse_loss = self.mse(enhanced_mod, clean_mod) * alpha | |||
mod_mse_loss = torch.mean( | |||
torch.sum(mod_mse_loss, dim=(1, 2, 3)) | |||
/ torch.sum(clean_mod**2, dim=(1, 2, 3))) | |||
return mod_mse_loss | |||
class ModulationDomainNCCLossModule(torch.nn.Module): | |||
"""Modulation-domain loss function developed in [1] for supervised speech enhancement | |||
# Speech Intelligibility Prediction Using Spectro-Temporal Modulation Analysis - based off of this | |||
In our paper, we used the gabor-based STRF kernels as the modulation kernels and used the log-mel spectrogram | |||
as the input spectrogram representation. | |||
Specific parameter details are in the paper and in the example below | |||
Parameters | |||
---------- | |||
modulation_kernels: nn.Module | |||
Differentiable module that transforms a spectrogram representation to the modulation domain | |||
modulation_domain = modulation_kernels(input_tf_representation) | |||
Input Spectrogram representation(B, T, F) --- (M) modulation_kernels---> Modulation Domain(B, M, T', F') | |||
[1] | |||
""" | |||
def __init__(self, modulation_kernels): | |||
super(ModulationDomainNCCLossModule, self).__init__() | |||
self.modulation_kernels = modulation_kernels | |||
self.mse = nn.MSELoss(reduce=False) | |||
def forward(self, enhanced_spect, clean_spect): | |||
"""Calculate modulation-domain loss | |||
Args: | |||
enhanced_spect (Tensor): spectrogram representation of enhanced signal (B, #frames, #freq_channels). | |||
clean_spect (Tensor): spectrogram representation of clean ground-truth signal (B, #frames, #freq_channels). | |||
Returns: | |||
Tensor: Modulation-domain loss value. | |||
""" | |||
clean_mod = self.modulation_kernels(clean_spect) | |||
enhanced_mod = self.modulation_kernels(enhanced_spect) | |||
mean_clean_mod = torch.mean(clean_mod, dim=2) | |||
mean_enhanced_mod = torch.mean(enhanced_mod, dim=2) | |||
normalized_clean = clean_mod - mean_clean_mod.unsqueeze(2) | |||
normalized_enhanced = enhanced_mod - mean_enhanced_mod.unsqueeze(2) | |||
inner_product = torch.sum( | |||
normalized_clean * normalized_enhanced, dim=2) | |||
normalized_denom = (torch.sum( | |||
normalized_clean * normalized_clean, dim=2))**.5 * (torch.sum( | |||
normalized_enhanced * normalized_enhanced, dim=2))**.5 | |||
ncc = inner_product / normalized_denom | |||
mod_mse_loss = torch.mean((ncc - 1.0)**2) | |||
return mod_mse_loss | |||
class GaborSTRFConv(nn.Module): | |||
"""Gabor-STRF-based cross-correlation kernel.""" | |||
def __init__(self, | |||
supn, | |||
supk, | |||
nkern, | |||
rates=None, | |||
scales=None, | |||
norm_strf=True, | |||
real_only=False): | |||
"""Instantiate a Gabor-based STRF convolution layer. | |||
Parameters | |||
---------- | |||
supn: int | |||
Time support in number of frames. Also the window length. | |||
supk: int | |||
Frequency support in number of channels. Also the window length. | |||
nkern: int | |||
Number of kernels, each with a learnable rate and scale. | |||
rates: list of float, None | |||
Initial values for temporal modulation. | |||
scales: list of float, None | |||
Initial values for spectral modulation. | |||
norm_strf: Boolean | |||
Normalize STRF kernels to be unit length | |||
real_only: Boolean | |||
If True, nkern REAL gabor-STRF kernels | |||
If False, nkern//2 REAL and nkern//2 IMAGINARY gabor-STRF kernels | |||
""" | |||
super(GaborSTRFConv, self).__init__() | |||
self.numN = supn | |||
self.numK = supk | |||
self.numKern = nkern | |||
self.real_only = real_only | |||
self.norm_strf = norm_strf | |||
if not real_only: | |||
nkern = nkern // 2 | |||
if supk % 2 == 0: # force odd number | |||
supk += 1 | |||
self.supk = torch.arange(supk, dtype=torch.float32) | |||
if supn % 2 == 0: # force odd number | |||
supn += 1 | |||
self.supn = torch.arange(supn, dtype=self.supk.dtype) | |||
self.padding = (supn // 2, supk // 2) | |||
# Set up learnable parameters | |||
# for param in (rates, scales): | |||
# assert (not param) or len(param) == nkern | |||
if not rates: | |||
rates = torch.rand(nkern) * math.pi / 2.0 | |||
if not scales: | |||
scales = (torch.rand(nkern) * 2.0 - 1.0) * math.pi / 2.0 | |||
self.rates_ = nn.Parameter(torch.Tensor(rates)) | |||
self.scales_ = nn.Parameter(torch.Tensor(scales)) | |||
def strfs(self): | |||
"""Make STRFs using the current parameters.""" | |||
if self.supn.device != self.rates_.device: # for first run | |||
self.supn = self.supn.to(self.rates_.device) | |||
self.supk = self.supk.to(self.rates_.device) | |||
n0, k0 = self.padding | |||
nwind = .5 - .5 * \ | |||
torch.cos(2 * math.pi * (self.supn + 1) / (len(self.supn) + 1)) | |||
kwind = .5 - .5 * \ | |||
torch.cos(2 * math.pi * (self.supk + 1) / (len(self.supk) + 1)) | |||
new_wind = torch.matmul((nwind).unsqueeze(-1), (kwind).unsqueeze(0)) | |||
n_n_0 = self.supn - n0 | |||
k_k_0 = self.supk - k0 | |||
n_mult = torch.matmul( | |||
n_n_0.unsqueeze(1), | |||
torch.ones((1, len(self.supk))).type(torch.FloatTensor).to( | |||
self.rates_.device)) | |||
k_mult = torch.matmul( | |||
torch.ones((len(self.supn), | |||
1)).type(torch.FloatTensor).to(self.rates_.device), | |||
k_k_0.unsqueeze(0)) | |||
inside = self.rates_.unsqueeze(1).unsqueeze( | |||
1) * n_mult + self.scales_.unsqueeze(1).unsqueeze(1) * k_mult | |||
real_strf = torch.cos(inside) * new_wind.unsqueeze(0) | |||
if self.real_only: | |||
final_strf = real_strf | |||
else: | |||
imag_strf = torch.sin(inside) * new_wind.unsqueeze(0) | |||
final_strf = torch.cat([real_strf, imag_strf], dim=0) | |||
if self.norm_strf: | |||
final_strf = final_strf / (torch.sum( | |||
final_strf**2, dim=(1, 2)).unsqueeze(1).unsqueeze(2))**.5 | |||
return final_strf | |||
def forward(self, sigspec): | |||
"""Forward pass a batch of (real) spectra [Batch x Time x Frequency].""" | |||
if len(sigspec.shape) == 2: # expand batch dimension if single eg | |||
sigspec = sigspec.unsqueeze(0) | |||
strfs = self.strfs().unsqueeze(1).type_as(sigspec) | |||
out = F.conv2d(sigspec.unsqueeze(1), strfs, padding=self.padding) | |||
return out | |||
def __repr__(self): | |||
"""Gabor filter""" | |||
report = """ | |||
+++++ Gabor Filter Kernels [{}], supn[{}], supk[{}] real only [{}] norm strf [{}] +++++ | |||
""".format(self.numKern, self.numN, self.numK, self.real_only, | |||
self.norm_strf) | |||
return report |
@@ -0,0 +1,483 @@ | |||
import torch | |||
import torch.nn as nn | |||
import torch.nn.functional as F | |||
from ..layers.activations import RectifiedLinear, Sigmoid | |||
from ..layers.affine_transform import AffineTransform | |||
from ..layers.deep_fsmn import DeepFsmn | |||
from ..layers.uni_deep_fsmn import Conv2d, UniDeepFsmn | |||
class MaskNet(nn.Module): | |||
def __init__(self, | |||
indim, | |||
outdim, | |||
layers=9, | |||
hidden_dim=128, | |||
hidden_dim2=None, | |||
lorder=20, | |||
rorder=0, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
crm=False, | |||
vad=False, | |||
linearout=False): | |||
super(MaskNet, self).__init__() | |||
self.linear1 = AffineTransform(indim, hidden_dim) | |||
self.relu = RectifiedLinear(hidden_dim, hidden_dim) | |||
if hidden_dim2 is None: | |||
hidden_dim2 = hidden_dim | |||
if rorder == 0: | |||
repeats = [ | |||
UniDeepFsmn( | |||
hidden_dim, | |||
hidden_dim, | |||
lorder, | |||
hidden_dim2, | |||
dilation=dilation, | |||
layer_norm=layer_norm, | |||
dropout=dropout) for i in range(layers) | |||
] | |||
else: | |||
repeats = [ | |||
DeepFsmn( | |||
hidden_dim, | |||
hidden_dim, | |||
lorder, | |||
rorder, | |||
hidden_dim2, | |||
layer_norm=layer_norm, | |||
dropout=dropout) for i in range(layers) | |||
] | |||
self.deepfsmn = nn.Sequential(*repeats) | |||
self.linear2 = AffineTransform(hidden_dim, outdim) | |||
self.crm = crm | |||
if self.crm: | |||
self.sig = nn.Tanh() | |||
else: | |||
self.sig = Sigmoid(outdim, outdim) | |||
self.vad = vad | |||
if self.vad: | |||
self.linear3 = AffineTransform(hidden_dim, 1) | |||
self.layers = layers | |||
self.linearout = linearout | |||
if self.linearout and self.vad: | |||
print('Warning: not supported nnet') | |||
def forward(self, feat, ctl=None): | |||
x1 = self.linear1(feat) | |||
x2 = self.relu(x1) | |||
if ctl is not None: | |||
ctl = min(ctl, self.layers - 1) | |||
for i in range(ctl): | |||
x2 = self.deepfsmn[i](x2) | |||
mask = self.sig(self.linear2(x2)) | |||
if self.vad: | |||
vad = torch.sigmoid(self.linear3(x2)) | |||
return mask, vad | |||
else: | |||
return mask | |||
x3 = self.deepfsmn(x2) | |||
if self.linearout: | |||
return self.linear2(x3) | |||
mask = self.sig(self.linear2(x3)) | |||
if self.vad: | |||
vad = torch.sigmoid(self.linear3(x3)) | |||
return mask, vad | |||
else: | |||
return mask | |||
def to_kaldi_nnet(self): | |||
re_str = '' | |||
re_str += '<Nnet>\n' | |||
re_str += self.linear1.to_kaldi_nnet() | |||
re_str += self.relu.to_kaldi_nnet() | |||
for dfsmn in self.deepfsmn: | |||
re_str += dfsmn.to_kaldi_nnet() | |||
re_str += self.linear2.to_kaldi_nnet() | |||
re_str += self.sig.to_kaldi_nnet() | |||
re_str += '</Nnet>\n' | |||
return re_str | |||
def to_raw_nnet(self, fid): | |||
self.linear1.to_raw_nnet(fid) | |||
for dfsmn in self.deepfsmn: | |||
dfsmn.to_raw_nnet(fid) | |||
self.linear2.to_raw_nnet(fid) | |||
class StageNet(nn.Module): | |||
def __init__(self, | |||
indim, | |||
outdim, | |||
layers=9, | |||
layers2=6, | |||
hidden_dim=128, | |||
lorder=20, | |||
rorder=0, | |||
layer_norm=False, | |||
dropout=0, | |||
crm=False, | |||
vad=False, | |||
linearout=False): | |||
super(StageNet, self).__init__() | |||
self.stage1 = nn.ModuleList() | |||
self.stage2 = nn.ModuleList() | |||
layer = nn.Sequential(nn.Linear(indim, hidden_dim), nn.ReLU()) | |||
self.stage1.append(layer) | |||
for i in range(layers): | |||
layer = UniDeepFsmn( | |||
hidden_dim, | |||
hidden_dim, | |||
lorder, | |||
hidden_dim, | |||
layer_norm=layer_norm, | |||
dropout=dropout) | |||
self.stage1.append(layer) | |||
layer = nn.Sequential(nn.Linear(hidden_dim, 321), nn.Sigmoid()) | |||
self.stage1.append(layer) | |||
# stage2 | |||
layer = nn.Sequential(nn.Linear(321 + indim, hidden_dim), nn.ReLU()) | |||
self.stage2.append(layer) | |||
for i in range(layers2): | |||
layer = UniDeepFsmn( | |||
hidden_dim, | |||
hidden_dim, | |||
lorder, | |||
hidden_dim, | |||
layer_norm=layer_norm, | |||
dropout=dropout) | |||
self.stage2.append(layer) | |||
layer = nn.Sequential( | |||
nn.Linear(hidden_dim, outdim), | |||
nn.Sigmoid() if not crm else nn.Tanh()) | |||
self.stage2.append(layer) | |||
self.crm = crm | |||
self.vad = vad | |||
self.linearout = linearout | |||
self.window = torch.hamming_window(640, periodic=False).cuda() | |||
self.freezed = False | |||
def freeze(self): | |||
if not self.freezed: | |||
for param in self.stage1.parameters(): | |||
param.requires_grad = False | |||
self.freezed = True | |||
print('freezed stage1') | |||
def forward(self, feat, mixture, ctl=None): | |||
if ctl == 'off': | |||
x = feat | |||
for i in range(len(self.stage1)): | |||
x = self.stage1[i](x) | |||
return x | |||
else: | |||
self.freeze() | |||
x = feat | |||
for i in range(len(self.stage1)): | |||
x = self.stage1[i](x) | |||
spec = torch.stft( | |||
mixture / 32768, | |||
640, | |||
320, | |||
640, | |||
self.window, | |||
center=False, | |||
return_complex=True) | |||
spec = torch.view_as_real(spec).permute([0, 2, 1, 3]) | |||
specmag = torch.sqrt(spec[..., 0]**2 + spec[..., 1]**2) | |||
est = x * specmag | |||
y = torch.cat([est, feat], dim=-1) | |||
for i in range(len(self.stage2)): | |||
y = self.stage2[i](y) | |||
return y | |||
class Unet(nn.Module): | |||
def __init__(self, | |||
indim, | |||
outdim, | |||
layers=9, | |||
dims=[256] * 4, | |||
lorder=20, | |||
rorder=0, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
crm=False, | |||
vad=False, | |||
linearout=False): | |||
super(Unet, self).__init__() | |||
self.linear1 = AffineTransform(indim, dims[0]) | |||
self.relu = RectifiedLinear(dims[0], dims[0]) | |||
self.encoder = nn.ModuleList() | |||
self.decoder = nn.ModuleList() | |||
for i in range(len(dims) - 1): | |||
layer = nn.Sequential( | |||
nn.Linear(dims[i], dims[i + 1]), nn.ReLU(), | |||
nn.Linear(dims[i + 1], dims[i + 1], bias=False), | |||
Conv2d( | |||
dims[i + 1], | |||
dims[i + 1], | |||
lorder, | |||
groups=dims[i + 1], | |||
skip_connect=True)) | |||
self.encoder.append(layer) | |||
for i in range(len(dims) - 1, 0, -1): | |||
layer = nn.Sequential( | |||
nn.Linear(dims[i] * 2, dims[i - 1]), nn.ReLU(), | |||
nn.Linear(dims[i - 1], dims[i - 1], bias=False), | |||
Conv2d( | |||
dims[i - 1], | |||
dims[i - 1], | |||
lorder, | |||
groups=dims[i - 1], | |||
skip_connect=True)) | |||
self.decoder.append(layer) | |||
self.tf = nn.ModuleList() | |||
for i in range(layers - 2 * (len(dims) - 1)): | |||
layer = nn.Sequential( | |||
nn.Linear(dims[-1], dims[-1]), nn.ReLU(), | |||
nn.Linear(dims[-1], dims[-1], bias=False), | |||
Conv2d( | |||
dims[-1], | |||
dims[-1], | |||
lorder, | |||
groups=dims[-1], | |||
skip_connect=True)) | |||
self.tf.append(layer) | |||
self.linear2 = AffineTransform(dims[0], outdim) | |||
self.crm = crm | |||
self.act = nn.Tanh() if self.crm else nn.Sigmoid() | |||
self.vad = False | |||
self.layers = layers | |||
self.linearout = linearout | |||
def forward(self, x, ctl=None): | |||
x = self.linear1(x) | |||
x = self.relu(x) | |||
encoder_out = [] | |||
for i in range(len(self.encoder)): | |||
x = self.encoder[i](x) | |||
encoder_out.append(x) | |||
for i in range(len(self.tf)): | |||
x = self.tf[i](x) | |||
for i in range(len(self.decoder)): | |||
x = torch.cat([x, encoder_out[-1 - i]], dim=-1) | |||
x = self.decoder[i](x) | |||
x = self.linear2(x) | |||
if self.linearout: | |||
return x | |||
return self.act(x) | |||
class BranchNet(nn.Module): | |||
def __init__(self, | |||
indim, | |||
outdim, | |||
layers=9, | |||
hidden_dim=256, | |||
lorder=20, | |||
rorder=0, | |||
dilation=1, | |||
layer_norm=False, | |||
dropout=0, | |||
crm=False, | |||
vad=False, | |||
linearout=False): | |||
super(BranchNet, self).__init__() | |||
self.linear1 = AffineTransform(indim, hidden_dim) | |||
self.relu = RectifiedLinear(hidden_dim, hidden_dim) | |||
self.convs = nn.ModuleList() | |||
self.deepfsmn = nn.ModuleList() | |||
self.FREQ = nn.ModuleList() | |||
self.TIME = nn.ModuleList() | |||
self.br1 = nn.ModuleList() | |||
self.br2 = nn.ModuleList() | |||
for i in range(layers): | |||
''' | |||
layer = nn.Sequential( | |||
nn.Linear(hidden_dim, hidden_dim), | |||
nn.ReLU(), | |||
nn.Linear(hidden_dim, hidden_dim, bias=False), | |||
Conv2d(hidden_dim, hidden_dim, lorder, | |||
groups=hidden_dim, skip_connect=True) | |||
) | |||
self.deepfsmn.append(layer) | |||
''' | |||
layer = nn.Sequential(nn.Linear(hidden_dim, hidden_dim), nn.ReLU()) | |||
self.FREQ.append(layer) | |||
''' | |||
layer = nn.GRU(hidden_dim, hidden_dim, | |||
batch_first=True, | |||
bidirectional=False) | |||
self.TIME.append(layer) | |||
layer = nn.Sequential( | |||
nn.Linear(hidden_dim, hidden_dim//2, bias=False), | |||
Conv2d(hidden_dim//2, hidden_dim//2, lorder, | |||
groups=hidden_dim//2, skip_connect=True) | |||
) | |||
self.br1.append(layer) | |||
layer = nn.GRU(hidden_dim, hidden_dim//2, | |||
batch_first=True, | |||
bidirectional=False) | |||
self.br2.append(layer) | |||
''' | |||
self.linear2 = AffineTransform(hidden_dim, outdim) | |||
self.crm = crm | |||
self.act = nn.Tanh() if self.crm else nn.Sigmoid() | |||
self.vad = False | |||
self.layers = layers | |||
self.linearout = linearout | |||
def forward(self, x, ctl=None): | |||
return self.forward_branch(x) | |||
def forward_sepconv(self, x): | |||
x = torch.unsqueeze(x, 1) | |||
for i in range(len(self.convs)): | |||
x = self.convs[i](x) | |||
x = F.relu(x) | |||
B, C, H, W = x.shape | |||
x = x.permute(0, 2, 1, 3) | |||
x = torch.reshape(x, [B, H, C * W]) | |||
x = self.linear1(x) | |||
x = self.relu(x) | |||
for i in range(self.layers): | |||
x = self.deepfsmn[i](x) + x | |||
x = self.linear2(x) | |||
return self.act(x) | |||
def forward_branch(self, x): | |||
x = self.linear1(x) | |||
x = self.relu(x) | |||
for i in range(self.layers): | |||
z = self.FREQ[i](x) | |||
x = z + x | |||
x = self.linear2(x) | |||
if self.linearout: | |||
return x | |||
return self.act(x) | |||
class TACNet(nn.Module): | |||
''' transform average concatenate for ad hoc dr | |||
''' | |||
def __init__(self, | |||
indim, | |||
outdim, | |||
layers=9, | |||
hidden_dim=128, | |||
lorder=20, | |||
rorder=0, | |||
crm=False, | |||
vad=False, | |||
linearout=False): | |||
super(TACNet, self).__init__() | |||
self.linear1 = AffineTransform(indim, hidden_dim) | |||
self.relu = RectifiedLinear(hidden_dim, hidden_dim) | |||
if rorder == 0: | |||
repeats = [ | |||
UniDeepFsmn(hidden_dim, hidden_dim, lorder, hidden_dim) | |||
for i in range(layers) | |||
] | |||
else: | |||
repeats = [ | |||
DeepFsmn(hidden_dim, hidden_dim, lorder, rorder, hidden_dim) | |||
for i in range(layers) | |||
] | |||
self.deepfsmn = nn.Sequential(*repeats) | |||
self.ch_transform = nn.ModuleList([]) | |||
self.ch_average = nn.ModuleList([]) | |||
self.ch_concat = nn.ModuleList([]) | |||
for i in range(layers): | |||
self.ch_transform.append( | |||
nn.Sequential(nn.Linear(hidden_dim, hidden_dim), nn.PReLU())) | |||
self.ch_average.append( | |||
nn.Sequential(nn.Linear(hidden_dim, hidden_dim), nn.PReLU())) | |||
self.ch_concat.append( | |||
nn.Sequential( | |||
nn.Linear(hidden_dim * 2, hidden_dim), nn.PReLU())) | |||
self.linear2 = AffineTransform(hidden_dim, outdim) | |||
self.crm = crm | |||
if self.crm: | |||
self.sig = nn.Tanh() | |||
else: | |||
self.sig = Sigmoid(outdim, outdim) | |||
self.vad = vad | |||
if self.vad: | |||
self.linear3 = AffineTransform(hidden_dim, 1) | |||
self.layers = layers | |||
self.linearout = linearout | |||
if self.linearout and self.vad: | |||
print('Warning: not supported nnet') | |||
def forward(self, feat, ctl=None): | |||
B, T, F = feat.shape | |||
# assume 4ch | |||
ch = 4 | |||
zlist = [] | |||
for c in range(ch): | |||
z = self.linear1(feat[..., c * (F // 4):(c + 1) * (F // 4)]) | |||
z = self.relu(z) | |||
zlist.append(z) | |||
for i in range(self.layers): | |||
# forward | |||
for c in range(ch): | |||
zlist[c] = self.deepfsmn[i](zlist[c]) | |||
# transform | |||
olist = [] | |||
for c in range(ch): | |||
z = self.ch_transform[i](zlist[c]) | |||
olist.append(z) | |||
# average | |||
avg = 0 | |||
for c in range(ch): | |||
avg = avg + olist[c] | |||
avg = avg / ch | |||
avg = self.ch_average[i](avg) | |||
# concate | |||
for c in range(ch): | |||
tac = torch.cat([olist[c], avg], dim=-1) | |||
tac = self.ch_concat[i](tac) | |||
zlist[c] = zlist[c] + tac | |||
for c in range(ch): | |||
zlist[c] = self.sig(self.linear2(zlist[c])) | |||
mask = torch.cat(zlist, dim=-1) | |||
return mask | |||
def to_kaldi_nnet(self): | |||
pass |
@@ -1,4 +1,4 @@ | |||
from .audio import * # noqa F403 | |||
from .audio import LinearAECPipeline | |||
from .base import Pipeline | |||
from .builder import pipeline | |||
from .cv import * # noqa F403 | |||
@@ -0,0 +1 @@ | |||
from .linear_aec_pipeline import LinearAECPipeline |
@@ -0,0 +1,160 @@ | |||
import importlib | |||
import os | |||
from typing import Any, Dict | |||
import numpy as np | |||
import scipy.io.wavfile as wav | |||
import torch | |||
import yaml | |||
from modelscope.preprocessors.audio import LinearAECAndFbank | |||
from modelscope.utils.constant import ModelFile, Tasks | |||
from ..base import Pipeline | |||
from ..builder import PIPELINES | |||
FEATURE_MVN = 'feature.DEY.mvn.txt' | |||
CONFIG_YAML = 'dey_mini.yaml' | |||
def initialize_config(module_cfg): | |||
r"""According to config items, load specific module dynamically with params. | |||
1. Load the module corresponding to the "module" param. | |||
2. Call function (or instantiate class) corresponding to the "main" param. | |||
3. Send the param (in "args") into the function (or class) when calling ( or instantiating). | |||
Args: | |||
module_cfg (dict): config items, eg: | |||
{ | |||
"module": "models.model", | |||
"main": "Model", | |||
"args": {...} | |||
} | |||
Returns: | |||
the module loaded. | |||
""" | |||
module = importlib.import_module(module_cfg['module']) | |||
return getattr(module, module_cfg['main'])(**module_cfg['args']) | |||
@PIPELINES.register_module( | |||
Tasks.speech_signal_process, module_name=r'speech_dfsmn_aec_psm_16k') | |||
class LinearAECPipeline(Pipeline): | |||
r"""AEC Inference Pipeline only support 16000 sample rate. | |||
When invoke the class with pipeline.__call__(), you should provide two params: | |||
Dict[str, Any] | |||
the path of wav files,eg:{ | |||
"nearend_mic": "/your/data/near_end_mic_audio.wav", | |||
"farend_speech": "/your/data/far_end_speech_audio.wav"} | |||
output_path (str, optional): "/your/output/audio_after_aec.wav" | |||
the file path to write generate audio. | |||
""" | |||
def __init__(self, model): | |||
r""" | |||
Args: | |||
model: model id on modelscope hub. | |||
""" | |||
super().__init__(model=model) | |||
self.use_cuda = torch.cuda.is_available() | |||
with open( | |||
os.path.join(self.model, CONFIG_YAML), encoding='utf-8') as f: | |||
self.config = yaml.full_load(f.read()) | |||
self.config['io']['mvn'] = os.path.join(self.model, FEATURE_MVN) | |||
self._init_model() | |||
self.preprocessor = LinearAECAndFbank(self.config['io']) | |||
n_fft = self.config['loss']['args']['n_fft'] | |||
hop_length = self.config['loss']['args']['hop_length'] | |||
winlen = n_fft | |||
window = torch.hamming_window(winlen, periodic=False) | |||
def stft(x): | |||
return torch.stft( | |||
x, | |||
n_fft, | |||
hop_length, | |||
winlen, | |||
center=False, | |||
window=window.to(x.device), | |||
return_complex=False) | |||
def istft(x, slen): | |||
return torch.istft( | |||
x, | |||
n_fft, | |||
hop_length, | |||
winlen, | |||
window=window.to(x.device), | |||
center=False, | |||
length=slen) | |||
self.stft = stft | |||
self.istft = istft | |||
def _init_model(self): | |||
checkpoint = torch.load( | |||
os.path.join(self.model, ModelFile.TORCH_MODEL_BIN_FILE), | |||
map_location='cpu') | |||
self.model = initialize_config(self.config['nnet']) | |||
if self.use_cuda: | |||
self.model = self.model.cuda() | |||
self.model.load_state_dict(checkpoint) | |||
def forward(self, inputs: Dict[str, Any]) -> Dict[str, Any]: | |||
r"""The AEC process. | |||
Args: | |||
inputs: dict={'feature': Tensor, 'base': Tensor} | |||
'feature' feature of input audio. | |||
'base' the base audio to mask. | |||
Returns: | |||
dict: | |||
{ | |||
'output_pcm': generated audio array | |||
} | |||
""" | |||
output_data = self._process(inputs['feature'], inputs['base']) | |||
return {'output_pcm': output_data} | |||
def postprocess(self, inputs: Dict[str, Any], **kwargs) -> Dict[str, Any]: | |||
r"""The post process. Will save audio to file, if the output_path is given. | |||
Args: | |||
inputs: dict: | |||
{ | |||
'output_pcm': generated audio array | |||
} | |||
kwargs: accept 'output_path' which is the path to write generated audio | |||
Returns: | |||
dict: | |||
{ | |||
'output_pcm': generated audio array | |||
} | |||
""" | |||
if 'output_path' in kwargs.keys(): | |||
wav.write(kwargs['output_path'], self.preprocessor.SAMPLE_RATE, | |||
inputs['output_pcm'].astype(np.int16)) | |||
inputs['output_pcm'] = inputs['output_pcm'] / 32768.0 | |||
return inputs | |||
def _process(self, fbanks, mixture): | |||
if self.use_cuda: | |||
fbanks = fbanks.cuda() | |||
mixture = mixture.cuda() | |||
if self.model.vad: | |||
with torch.no_grad(): | |||
masks, vad = self.model(fbanks.unsqueeze(0)) | |||
masks = masks.permute([2, 1, 0]) | |||
else: | |||
with torch.no_grad(): | |||
masks = self.model(fbanks.unsqueeze(0)) | |||
masks = masks.permute([2, 1, 0]) | |||
spectrum = self.stft(mixture) | |||
masked_spec = spectrum * masks | |||
masked_sig = self.istft(masked_spec, len(mixture)).cpu().numpy() | |||
return masked_sig |
@@ -84,6 +84,12 @@ TASK_OUTPUTS = { | |||
# ============ audio tasks =================== | |||
# audio processed for single file in PCM format | |||
# { | |||
# "output_pcm": np.array with shape(samples,) and dtype float32 | |||
# } | |||
Tasks.speech_signal_process: ['output_pcm'], | |||
# ============ multi-modal tasks =================== | |||
# image caption result for single sample | |||
@@ -1,5 +1,6 @@ | |||
# Copyright (c) Alibaba, Inc. and its affiliates. | |||
from .audio import LinearAECAndFbank | |||
from .base import Preprocessor | |||
from .builder import PREPROCESSORS, build_preprocessor | |||
from .common import Compose | |||
@@ -0,0 +1,230 @@ | |||
import ctypes | |||
import os | |||
from typing import Any, Dict | |||
import numpy as np | |||
import scipy.io.wavfile as wav | |||
import torch | |||
import torchaudio.compliance.kaldi as kaldi | |||
from numpy.ctypeslib import ndpointer | |||
from modelscope.utils.constant import Fields | |||
from .builder import PREPROCESSORS | |||
def load_wav(path): | |||
samp_rate, data = wav.read(path) | |||
return np.float32(data), samp_rate | |||
def load_library(libaec): | |||
libaec_in_cwd = os.path.join('.', libaec) | |||
if os.path.exists(libaec_in_cwd): | |||
libaec = libaec_in_cwd | |||
mitaec = ctypes.cdll.LoadLibrary(libaec) | |||
fe_process = mitaec.fe_process_inst | |||
fe_process.argtypes = [ | |||
ndpointer(ctypes.c_float, flags='C_CONTIGUOUS'), | |||
ndpointer(ctypes.c_float, flags='C_CONTIGUOUS'), ctypes.c_int, | |||
ndpointer(ctypes.c_float, flags='C_CONTIGUOUS'), | |||
ndpointer(ctypes.c_float, flags='C_CONTIGUOUS'), | |||
ndpointer(ctypes.c_float, flags='C_CONTIGUOUS') | |||
] | |||
return fe_process | |||
def do_linear_aec(fe_process, mic, ref, int16range=True): | |||
mic = np.float32(mic) | |||
ref = np.float32(ref) | |||
if len(mic) > len(ref): | |||
mic = mic[:len(ref)] | |||
out_mic = np.zeros_like(mic) | |||
out_linear = np.zeros_like(mic) | |||
out_echo = np.zeros_like(mic) | |||
out_ref = np.zeros_like(mic) | |||
if int16range: | |||
mic /= 32768 | |||
ref /= 32768 | |||
fe_process(mic, ref, len(mic), out_mic, out_linear, out_echo) | |||
# out_ref not in use here | |||
if int16range: | |||
out_mic *= 32768 | |||
out_linear *= 32768 | |||
out_echo *= 32768 | |||
return out_mic, out_ref, out_linear, out_echo | |||
def load_kaldi_feature_transform(filename): | |||
fp = open(filename, 'r') | |||
all_str = fp.read() | |||
pos1 = all_str.find('AddShift') | |||
pos2 = all_str.find('[', pos1) | |||
pos3 = all_str.find(']', pos2) | |||
mean = np.fromstring(all_str[pos2 + 1:pos3], dtype=np.float32, sep=' ') | |||
pos1 = all_str.find('Rescale') | |||
pos2 = all_str.find('[', pos1) | |||
pos3 = all_str.find(']', pos2) | |||
scale = np.fromstring(all_str[pos2 + 1:pos3], dtype=np.float32, sep=' ') | |||
fp.close() | |||
return mean, scale | |||
class Feature: | |||
r"""Extract feat from one utterance. | |||
""" | |||
def __init__(self, | |||
fbank_config, | |||
feat_type='spec', | |||
mvn_file=None, | |||
cuda=False): | |||
r""" | |||
Args: | |||
fbank_config (dict): | |||
feat_type (str): | |||
raw: do nothing | |||
fbank: use kaldi.fbank | |||
spec: Real/Imag | |||
logpow: log(1+|x|^2) | |||
mvn_file (str): the path of data file for mean variance normalization | |||
cuda: | |||
""" | |||
self.fbank_config = fbank_config | |||
self.feat_type = feat_type | |||
self.n_fft = fbank_config['frame_length'] * fbank_config[ | |||
'sample_frequency'] // 1000 | |||
self.hop_length = fbank_config['frame_shift'] * fbank_config[ | |||
'sample_frequency'] // 1000 | |||
self.window = torch.hamming_window(self.n_fft, periodic=False) | |||
self.mvn = False | |||
if mvn_file is not None and os.path.exists(mvn_file): | |||
print(f'loading mvn file: {mvn_file}') | |||
shift, scale = load_kaldi_feature_transform(mvn_file) | |||
self.shift = torch.from_numpy(shift) | |||
self.scale = torch.from_numpy(scale) | |||
self.mvn = True | |||
if cuda: | |||
self.window = self.window.cuda() | |||
if self.mvn: | |||
self.shift = self.shift.cuda() | |||
self.scale = self.scale.cuda() | |||
def compute(self, utt): | |||
r""" | |||
Args: | |||
utt: in [-32768, 32767] range | |||
Returns: | |||
[..., T, F] | |||
""" | |||
if self.feat_type == 'raw': | |||
return utt | |||
elif self.feat_type == 'fbank': | |||
if len(utt.shape) == 1: | |||
utt = utt.unsqueeze(0) | |||
feat = kaldi.fbank(utt, **self.fbank_config) | |||
elif self.feat_type == 'spec': | |||
spec = torch.stft( | |||
utt / 32768, | |||
self.n_fft, | |||
self.hop_length, | |||
self.n_fft, | |||
self.window, | |||
center=False, | |||
return_complex=True) | |||
feat = torch.cat([spec.real, spec.imag], dim=-2).permute(-1, -2) | |||
elif self.feat_type == 'logpow': | |||
spec = torch.stft( | |||
utt, | |||
self.n_fft, | |||
self.hop_length, | |||
self.n_fft, | |||
self.window, | |||
center=False, | |||
return_complex=True) | |||
abspow = torch.abs(spec)**2 | |||
feat = torch.log(1 + abspow).permute(-1, -2) | |||
return feat | |||
def normalize(self, feat): | |||
if self.mvn: | |||
feat = feat + self.shift | |||
feat = feat * self.scale | |||
return feat | |||
@PREPROCESSORS.register_module(Fields.audio) | |||
class LinearAECAndFbank: | |||
SAMPLE_RATE = 16000 | |||
def __init__(self, io_config): | |||
self.trunc_length = 7200 * self.SAMPLE_RATE | |||
self.linear_aec_delay = io_config['linear_aec_delay'] | |||
self.feature = Feature(io_config['fbank_config'], | |||
io_config['feat_type'], io_config['mvn']) | |||
self.mitaec = load_library(io_config['mitaec_library']) | |||
self.mask_on_mic = io_config['mask_on'] == 'nearend_mic' | |||
def __call__(self, data: Dict[str, Any]) -> Dict[str, Any]: | |||
""" linear filtering the near end mic and far end audio, then extract the feature | |||
:param data: dict with two keys and correspond audios: "nearend_mic" and "farend_speech" | |||
:return: dict with two keys and Tensor values: "base" linear filtered audio,and "feature" | |||
""" | |||
# read files | |||
nearend_mic, fs = load_wav(data['nearend_mic']) | |||
assert fs == self.SAMPLE_RATE, f'The sample rate should be {self.SAMPLE_RATE}' | |||
farend_speech, fs = load_wav(data['farend_speech']) | |||
assert fs == self.SAMPLE_RATE, f'The sample rate should be {self.SAMPLE_RATE}' | |||
if 'nearend_speech' in data: | |||
nearend_speech, fs = load_wav(data['nearend_speech']) | |||
assert fs == self.SAMPLE_RATE, f'The sample rate should be {self.SAMPLE_RATE}' | |||
else: | |||
nearend_speech = np.zeros_like(nearend_mic) | |||
out_mic, out_ref, out_linear, out_echo = do_linear_aec( | |||
self.mitaec, nearend_mic, farend_speech) | |||
# fix 20ms linear aec delay by delaying the target speech | |||
extra_zeros = np.zeros([int(self.linear_aec_delay * fs)]) | |||
nearend_speech = np.concatenate([extra_zeros, nearend_speech]) | |||
# truncate files to the same length | |||
flen = min( | |||
len(out_mic), len(out_ref), len(out_linear), len(out_echo), | |||
len(nearend_speech)) | |||
fstart = 0 | |||
flen = min(flen, self.trunc_length) | |||
nearend_mic, out_ref, out_linear, out_echo, nearend_speech = ( | |||
out_mic[fstart:flen], out_ref[fstart:flen], | |||
out_linear[fstart:flen], out_echo[fstart:flen], | |||
nearend_speech[fstart:flen]) | |||
# extract features (frames, [mic, linear, ref, aes?]) | |||
feat = torch.FloatTensor() | |||
nearend_mic = torch.from_numpy(np.float32(nearend_mic)) | |||
fbank_nearend_mic = self.feature.compute(nearend_mic) | |||
feat = torch.cat([feat, fbank_nearend_mic], dim=1) | |||
out_linear = torch.from_numpy(np.float32(out_linear)) | |||
fbank_out_linear = self.feature.compute(out_linear) | |||
feat = torch.cat([feat, fbank_out_linear], dim=1) | |||
out_echo = torch.from_numpy(np.float32(out_echo)) | |||
fbank_out_echo = self.feature.compute(out_echo) | |||
feat = torch.cat([feat, fbank_out_echo], dim=1) | |||
# feature transform | |||
feat = self.feature.normalize(feat) | |||
# prepare target | |||
if nearend_speech is not None: | |||
nearend_speech = torch.from_numpy(np.float32(nearend_speech)) | |||
if self.mask_on_mic: | |||
base = nearend_mic | |||
else: | |||
base = out_linear | |||
out_data = {'base': base, 'target': nearend_speech, 'feature': feat} | |||
return out_data |
@@ -7,6 +7,7 @@ opencv-python-headless | |||
Pillow>=6.2.0 | |||
pyyaml | |||
requests | |||
scipy | |||
tokenizers<=0.10.3 | |||
transformers<=4.16.2 | |||
yapf |
@@ -11,6 +11,7 @@ default_section = THIRDPARTY | |||
BASED_ON_STYLE = pep8 | |||
BLANK_LINE_BEFORE_NESTED_CLASS_OR_DEF = true | |||
SPLIT_BEFORE_EXPRESSION_AFTER_OPENING_PAREN = true | |||
SPLIT_BEFORE_ARITHMETIC_OPERATOR = true | |||
[codespell] | |||
skip = *.ipynb | |||
@@ -20,5 +21,5 @@ ignore-words-list = patten,nd,ty,mot,hist,formating,winn,gool,datas,wan,confids | |||
[flake8] | |||
select = B,C,E,F,P,T4,W,B9 | |||
max-line-length = 120 | |||
ignore = F401,F821 | |||
ignore = F401,F821,W503 | |||
exclude = docs/src,*.pyi,.git |
@@ -0,0 +1,56 @@ | |||
import os.path | |||
import shutil | |||
import unittest | |||
from modelscope.fileio import File | |||
from modelscope.pipelines import pipeline | |||
from modelscope.utils.constant import Tasks | |||
from modelscope.utils.hub import get_model_cache_dir | |||
NEAREND_MIC_URL = 'https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/AEC/sample_audio/nearend_mic.wav' | |||
FAREND_SPEECH_URL = 'https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/AEC/sample_audio/farend_speech.wav' | |||
NEAREND_MIC_FILE = 'nearend_mic.wav' | |||
FAREND_SPEECH_FILE = 'farend_speech.wav' | |||
AEC_LIB_URL = 'http://isv-data.oss-cn-hangzhou.aliyuncs.com/ics%2FMaaS%2FAEC%2Flib%2Flibmitaec_pyio.so' \ | |||
'?Expires=1664085465&OSSAccessKeyId=LTAIxjQyZNde90zh&Signature=Y7gelmGEsQAJRK4yyHSYMrdWizk%3D' | |||
AEC_LIB_FILE = 'libmitaec_pyio.so' | |||
def download(remote_path, local_path): | |||
local_dir = os.path.dirname(local_path) | |||
if len(local_dir) > 0: | |||
if not os.path.exists(local_dir): | |||
os.makedirs(local_dir) | |||
with open(local_path, 'wb') as ofile: | |||
ofile.write(File.read(remote_path)) | |||
class SpeechSignalProcessTest(unittest.TestCase): | |||
def setUp(self) -> None: | |||
self.model_id = 'damo/speech_dfsmn_aec_psm_16k' | |||
# switch to False if downloading everytime is not desired | |||
purge_cache = True | |||
if purge_cache: | |||
shutil.rmtree( | |||
get_model_cache_dir(self.model_id), ignore_errors=True) | |||
# A temporary hack to provide c++ lib. Download it first. | |||
download(AEC_LIB_URL, AEC_LIB_FILE) | |||
def test_run(self): | |||
download(NEAREND_MIC_URL, NEAREND_MIC_FILE) | |||
download(FAREND_SPEECH_URL, FAREND_SPEECH_FILE) | |||
input = { | |||
'nearend_mic': NEAREND_MIC_FILE, | |||
'farend_speech': FAREND_SPEECH_FILE | |||
} | |||
aec = pipeline( | |||
Tasks.speech_signal_process, | |||
model=self.model_id, | |||
pipeline_name=r'speech_dfsmn_aec_psm_16k') | |||
aec(input, output_path='output.wav') | |||
if __name__ == '__main__': | |||
unittest.main() |