online-model-config.cc
3.9 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
127
// sherpa-onnx/csrc/online-model-config.cc
//
// Copyright (c) 2023 Xiaomi Corporation
#include "sherpa-onnx/csrc/online-model-config.h"
#include <string>
#include "sherpa-onnx/csrc/file-utils.h"
#include "sherpa-onnx/csrc/macros.h"
namespace sherpa_onnx {
void OnlineModelConfig::Register(ParseOptions *po) {
transducer.Register(po);
paraformer.Register(po);
wenet_ctc.Register(po);
zipformer2_ctc.Register(po);
nemo_ctc.Register(po);
provider_config.Register(po);
po->Register("tokens", &tokens, "Path to tokens.txt");
po->Register("num-threads", &num_threads,
"Number of threads to run the neural network");
po->Register("warm-up", &warm_up,
"Number of warm-up to run the onnxruntime"
"Valid vales are: zipformer2");
po->Register("debug", &debug,
"true to print model information while loading it.");
po->Register("modeling-unit", &modeling_unit,
"The modeling unit of the model, commonly used units are bpe, "
"cjkchar, cjkchar+bpe, etc. Currently, it is needed only when "
"hotwords are provided, we need it to encode the hotwords into "
"token sequence.");
po->Register("bpe-vocab", &bpe_vocab,
"The vocabulary generated by google's sentencepiece program. "
"It is a file has two columns, one is the token, the other is "
"the log probability, you can get it from the directory where "
"your bpe model is generated. Only used when hotwords provided "
"and the modeling unit is bpe or cjkchar+bpe");
po->Register("model-type", &model_type,
"Specify it to reduce model initialization time. "
"Valid values are: conformer, lstm, zipformer, zipformer2, "
"wenet_ctc, nemo_ctc. "
"All other values lead to loading the model twice.");
}
bool OnlineModelConfig::Validate() const {
if (num_threads < 1) {
SHERPA_ONNX_LOGE("num_threads should be > 0. Given %d", num_threads);
return false;
}
if (!tokens_buf.empty() && FileExists(tokens)) {
SHERPA_ONNX_LOGE(
"you can not provide a tokens_buf and a tokens file: '%s', "
"at the same time, which is confusing",
tokens.c_str());
return false;
}
if (tokens_buf.empty() && !FileExists(tokens)) {
SHERPA_ONNX_LOGE(
"tokens: '%s' does not exist, you should provide "
"either a tokens buffer or a tokens file",
tokens.c_str());
return false;
}
if (!modeling_unit.empty() &&
(modeling_unit == "bpe" || modeling_unit == "cjkchar+bpe")) {
if (!FileExists(bpe_vocab)) {
SHERPA_ONNX_LOGE("bpe_vocab: '%s' does not exist", bpe_vocab.c_str());
return false;
}
}
if (!paraformer.encoder.empty()) {
return paraformer.Validate();
}
if (!wenet_ctc.model.empty()) {
return wenet_ctc.Validate();
}
if (!zipformer2_ctc.model.empty()) {
return zipformer2_ctc.Validate();
}
if (!nemo_ctc.model.empty()) {
return nemo_ctc.Validate();
}
if (!provider_config.Validate()) {
return false;
}
return transducer.Validate();
}
std::string OnlineModelConfig::ToString() const {
std::ostringstream os;
os << "OnlineModelConfig(";
os << "transducer=" << transducer.ToString() << ", ";
os << "paraformer=" << paraformer.ToString() << ", ";
os << "wenet_ctc=" << wenet_ctc.ToString() << ", ";
os << "zipformer2_ctc=" << zipformer2_ctc.ToString() << ", ";
os << "nemo_ctc=" << nemo_ctc.ToString() << ", ";
os << "provider_config=" << provider_config.ToString() << ", ";
os << "tokens=\"" << tokens << "\", ";
os << "num_threads=" << num_threads << ", ";
os << "warm_up=" << warm_up << ", ";
os << "debug=" << (debug ? "True" : "False") << ", ";
os << "model_type=\"" << model_type << "\", ";
os << "modeling_unit=\"" << modeling_unit << "\", ";
os << "bpe_vocab=\"" << bpe_vocab << "\")";
return os.str();
}
} // namespace sherpa_onnx