1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
|
/*
* Copyright (c) 2018 The WebRTC project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#ifndef MODULES_AUDIO_PROCESSING_AGC2_RNN_VAD_RNN_H_
#define MODULES_AUDIO_PROCESSING_AGC2_RNN_VAD_RNN_H_
#include <stddef.h>
#include <sys/types.h>
#include <array>
#include <vector>
#include "api/array_view.h"
#include "api/function_view.h"
#include "modules/audio_processing/agc2/rnn_vad/common.h"
#include "rtc_base/system/arch.h"
namespace webrtc {
namespace rnn_vad {
// Maximum number of units for a fully-connected layer. This value is used to
// over-allocate space for fully-connected layers output vectors (implemented as
// std::array). The value should equal the number of units of the largest
// fully-connected layer.
constexpr size_t kFullyConnectedLayersMaxUnits = 24;
// Maximum number of units for a recurrent layer. This value is used to
// over-allocate space for recurrent layers state vectors (implemented as
// std::array). The value should equal the number of units of the largest
// recurrent layer.
constexpr size_t kRecurrentLayersMaxUnits = 24;
// Fully-connected layer.
class FullyConnectedLayer {
public:
FullyConnectedLayer(size_t input_size,
size_t output_size,
rtc::ArrayView<const int8_t> bias,
rtc::ArrayView<const int8_t> weights,
rtc::FunctionView<float(float)> activation_function,
Optimization optimization);
FullyConnectedLayer(const FullyConnectedLayer&) = delete;
FullyConnectedLayer& operator=(const FullyConnectedLayer&) = delete;
~FullyConnectedLayer();
size_t input_size() const { return input_size_; }
size_t output_size() const { return output_size_; }
Optimization optimization() const { return optimization_; }
rtc::ArrayView<const float> GetOutput() const;
// Computes the fully-connected layer output.
void ComputeOutput(rtc::ArrayView<const float> input);
private:
const size_t input_size_;
const size_t output_size_;
const std::vector<float> bias_;
const std::vector<float> weights_;
rtc::FunctionView<float(float)> activation_function_;
// The output vector of a recurrent layer has length equal to |output_size_|.
// However, for efficiency, over-allocation is used.
std::array<float, kFullyConnectedLayersMaxUnits> output_;
const Optimization optimization_;
};
// Recurrent layer with gated recurrent units (GRUs) with sigmoid and ReLU as
// activation functions for the update/reset and output gates respectively.
class GatedRecurrentLayer {
public:
GatedRecurrentLayer(size_t input_size,
size_t output_size,
rtc::ArrayView<const int8_t> bias,
rtc::ArrayView<const int8_t> weights,
rtc::ArrayView<const int8_t> recurrent_weights,
Optimization optimization);
GatedRecurrentLayer(const GatedRecurrentLayer&) = delete;
GatedRecurrentLayer& operator=(const GatedRecurrentLayer&) = delete;
~GatedRecurrentLayer();
size_t input_size() const { return input_size_; }
size_t output_size() const { return output_size_; }
Optimization optimization() const { return optimization_; }
rtc::ArrayView<const float> GetOutput() const;
void Reset();
// Computes the recurrent layer output and updates the status.
void ComputeOutput(rtc::ArrayView<const float> input);
private:
const size_t input_size_;
const size_t output_size_;
const std::vector<float> bias_;
const std::vector<float> weights_;
const std::vector<float> recurrent_weights_;
// The state vector of a recurrent layer has length equal to |output_size_|.
// However, to avoid dynamic allocation, over-allocation is used.
std::array<float, kRecurrentLayersMaxUnits> state_;
const Optimization optimization_;
};
// Recurrent network based VAD.
class RnnBasedVad {
public:
RnnBasedVad();
RnnBasedVad(const RnnBasedVad&) = delete;
RnnBasedVad& operator=(const RnnBasedVad&) = delete;
~RnnBasedVad();
void Reset();
// Compute and returns the probability of voice (range: [0.0, 1.0]).
float ComputeVadProbability(
rtc::ArrayView<const float, kFeatureVectorSize> feature_vector,
bool is_silence);
private:
FullyConnectedLayer input_layer_;
GatedRecurrentLayer hidden_layer_;
FullyConnectedLayer output_layer_;
};
} // namespace rnn_vad
} // namespace webrtc
#endif // MODULES_AUDIO_PROCESSING_AGC2_RNN_VAD_RNN_H_
|
