speaker-embedding-extractor-nemo-model.cc
4.1 KB
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
// sherpa-onnx/csrc/speaker-embedding-extractor-nemo-model.cc
//
// Copyright (c) 2024 Xiaomi Corporation
#include "sherpa-onnx/csrc/speaker-embedding-extractor-nemo-model.h"
#include <string>
#include <utility>
#include <vector>
#include "sherpa-onnx/csrc/macros.h"
#include "sherpa-onnx/csrc/onnx-utils.h"
#include "sherpa-onnx/csrc/session.h"
#include "sherpa-onnx/csrc/speaker-embedding-extractor-nemo-model-meta-data.h"
namespace sherpa_onnx {
class SpeakerEmbeddingExtractorNeMoModel::Impl {
public:
explicit Impl(const SpeakerEmbeddingExtractorConfig &config)
: config_(config),
env_(ORT_LOGGING_LEVEL_ERROR),
sess_opts_(GetSessionOptions(config)),
allocator_{} {
{
auto buf = ReadFile(config.model);
Init(buf.data(), buf.size());
}
}
Ort::Value Compute(Ort::Value x, Ort::Value x_lens) const {
std::array<Ort::Value, 2> inputs = {std::move(x), std::move(x_lens)};
// output_names_ptr_[0] is logits
// output_names_ptr_[1] is embeddings
// so we use output_names_ptr_.data() + 1 here to extract only the
// embeddings
auto outputs = sess_->Run({}, input_names_ptr_.data(), inputs.data(),
inputs.size(), output_names_ptr_.data() + 1, 1);
return std::move(outputs[0]);
}
OrtAllocator *Allocator() const { return allocator_; }
const SpeakerEmbeddingExtractorNeMoModelMetaData &GetMetaData() const {
return meta_data_;
}
private:
void Init(void *model_data, size_t model_data_length) {
sess_ = std::make_unique<Ort::Session>(env_, model_data, model_data_length,
sess_opts_);
GetInputNames(sess_.get(), &input_names_, &input_names_ptr_);
GetOutputNames(sess_.get(), &output_names_, &output_names_ptr_);
// get meta data
Ort::ModelMetadata meta_data = sess_->GetModelMetadata();
if (config_.debug) {
std::ostringstream os;
PrintModelMetadata(os, meta_data);
SHERPA_ONNX_LOGE("%s", os.str().c_str());
}
Ort::AllocatorWithDefaultOptions allocator; // used in the macro below
SHERPA_ONNX_READ_META_DATA(meta_data_.output_dim, "output_dim");
SHERPA_ONNX_READ_META_DATA(meta_data_.feat_dim, "feat_dim");
SHERPA_ONNX_READ_META_DATA(meta_data_.sample_rate, "sample_rate");
SHERPA_ONNX_READ_META_DATA(meta_data_.window_size_ms, "window_size_ms");
SHERPA_ONNX_READ_META_DATA(meta_data_.window_stride_ms, "window_stride_ms");
SHERPA_ONNX_READ_META_DATA_STR(meta_data_.language, "language");
SHERPA_ONNX_READ_META_DATA_STR_WITH_DEFAULT(
meta_data_.feature_normalize_type, "feature_normalize_type", "");
SHERPA_ONNX_READ_META_DATA_STR_WITH_DEFAULT(meta_data_.window_type,
"window_type", "povey");
std::string framework;
SHERPA_ONNX_READ_META_DATA_STR(framework, "framework");
if (framework != "nemo") {
SHERPA_ONNX_LOGE("Expect a NeMo model, given: %s", framework.c_str());
exit(-1);
}
}
private:
SpeakerEmbeddingExtractorConfig config_;
Ort::Env env_;
Ort::SessionOptions sess_opts_;
Ort::AllocatorWithDefaultOptions allocator_;
std::unique_ptr<Ort::Session> sess_;
std::vector<std::string> input_names_;
std::vector<const char *> input_names_ptr_;
std::vector<std::string> output_names_;
std::vector<const char *> output_names_ptr_;
SpeakerEmbeddingExtractorNeMoModelMetaData meta_data_;
};
SpeakerEmbeddingExtractorNeMoModel::SpeakerEmbeddingExtractorNeMoModel(
const SpeakerEmbeddingExtractorConfig &config)
: impl_(std::make_unique<Impl>(config)) {}
SpeakerEmbeddingExtractorNeMoModel::~SpeakerEmbeddingExtractorNeMoModel() =
default;
const SpeakerEmbeddingExtractorNeMoModelMetaData &
SpeakerEmbeddingExtractorNeMoModel::GetMetaData() const {
return impl_->GetMetaData();
}
Ort::Value SpeakerEmbeddingExtractorNeMoModel::Compute(
Ort::Value x, Ort::Value x_lens) const {
return impl_->Compute(std::move(x), std::move(x_lens));
}
OrtAllocator *SpeakerEmbeddingExtractorNeMoModel::Allocator() const {
return impl_->Allocator();
}
} // namespace sherpa_onnx