sherpa-onnx-wasm-nodejs.cc
6.8 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
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
// wasm/sherpa-onnx-wasm-main-nodejs.cc
//
// Copyright (c) 2024 Xiaomi Corporation
#include <stdio.h>
#include <algorithm>
#include <memory>
#include "sherpa-onnx/c-api/c-api.h"
extern "C" {
static_assert(sizeof(SherpaOnnxOfflineTransducerModelConfig) == 3 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineParaformerModelConfig) == 4, "");
static_assert(sizeof(SherpaOnnxOfflineDolphinModelConfig) == 4, "");
static_assert(sizeof(SherpaOnnxOfflineNemoEncDecCtcModelConfig) == 4, "");
static_assert(sizeof(SherpaOnnxOfflineWhisperModelConfig) == 5 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineFireRedAsrModelConfig) == 2 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineMoonshineModelConfig) == 4 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineTdnnModelConfig) == 4, "");
static_assert(sizeof(SherpaOnnxOfflineSenseVoiceModelConfig) == 3 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineLMConfig) == 2 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineModelConfig) ==
sizeof(SherpaOnnxOfflineTransducerModelConfig) +
sizeof(SherpaOnnxOfflineParaformerModelConfig) +
sizeof(SherpaOnnxOfflineNemoEncDecCtcModelConfig) +
sizeof(SherpaOnnxOfflineWhisperModelConfig) +
sizeof(SherpaOnnxOfflineTdnnModelConfig) + 8 * 4 +
sizeof(SherpaOnnxOfflineSenseVoiceModelConfig) +
sizeof(SherpaOnnxOfflineMoonshineModelConfig) +
sizeof(SherpaOnnxOfflineFireRedAsrModelConfig) +
sizeof(SherpaOnnxOfflineDolphinModelConfig),
"");
static_assert(sizeof(SherpaOnnxFeatureConfig) == 2 * 4, "");
static_assert(sizeof(SherpaOnnxOfflineRecognizerConfig) ==
sizeof(SherpaOnnxFeatureConfig) +
sizeof(SherpaOnnxOfflineLMConfig) +
sizeof(SherpaOnnxOfflineModelConfig) + 7 * 4,
"");
void PrintOfflineTtsConfig(SherpaOnnxOfflineTtsConfig *tts_config) {
auto tts_model_config = &tts_config->model;
auto vits_model_config = &tts_model_config->vits;
fprintf(stdout, "----------vits model config----------\n");
fprintf(stdout, "model: %s\n", vits_model_config->model);
fprintf(stdout, "lexicon: %s\n", vits_model_config->lexicon);
fprintf(stdout, "tokens: %s\n", vits_model_config->tokens);
fprintf(stdout, "data_dir: %s\n", vits_model_config->data_dir);
fprintf(stdout, "noise scale: %.3f\n", vits_model_config->noise_scale);
fprintf(stdout, "noise scale w: %.3f\n", vits_model_config->noise_scale_w);
fprintf(stdout, "length scale: %.3f\n", vits_model_config->length_scale);
fprintf(stdout, "dict_dir: %s\n", vits_model_config->dict_dir);
fprintf(stdout, "----------tts model config----------\n");
fprintf(stdout, "num threads: %d\n", tts_model_config->num_threads);
fprintf(stdout, "debug: %d\n", tts_model_config->debug);
fprintf(stdout, "provider: %s\n", tts_model_config->provider);
fprintf(stdout, "----------tts config----------\n");
fprintf(stdout, "rule_fsts: %s\n", tts_config->rule_fsts);
fprintf(stdout, "max num sentences: %d\n", tts_config->max_num_sentences);
}
void PrintOfflineRecognizerConfig(SherpaOnnxOfflineRecognizerConfig *config) {
auto model_config = &config->model_config;
auto feat = &config->feat_config;
auto transducer = &model_config->transducer;
auto paraformer = &model_config->paraformer;
auto nemo_ctc = &model_config->nemo_ctc;
auto whisper = &model_config->whisper;
auto tdnn = &model_config->tdnn;
auto sense_voice = &model_config->sense_voice;
auto moonshine = &model_config->moonshine;
auto fire_red_asr = &model_config->fire_red_asr;
auto dolphin = &model_config->dolphin;
fprintf(stdout, "----------offline transducer model config----------\n");
fprintf(stdout, "encoder: %s\n", transducer->encoder);
fprintf(stdout, "decoder: %s\n", transducer->decoder);
fprintf(stdout, "joiner: %s\n", transducer->joiner);
fprintf(stdout, "----------offline paraformer model config----------\n");
fprintf(stdout, "model: %s\n", paraformer->model);
fprintf(stdout, "----------offline nemo_ctc model config----------\n");
fprintf(stdout, "model: %s\n", nemo_ctc->model);
fprintf(stdout, "----------offline whisper model config----------\n");
fprintf(stdout, "encoder: %s\n", whisper->encoder);
fprintf(stdout, "decoder: %s\n", whisper->decoder);
fprintf(stdout, "language: %s\n", whisper->language);
fprintf(stdout, "task: %s\n", whisper->task);
fprintf(stdout, "tail_paddings: %d\n", whisper->tail_paddings);
fprintf(stdout, "----------offline tdnn model config----------\n");
fprintf(stdout, "model: %s\n", tdnn->model);
fprintf(stdout, "----------offline sense_voice model config----------\n");
fprintf(stdout, "model: %s\n", sense_voice->model);
fprintf(stdout, "language: %s\n", sense_voice->language);
fprintf(stdout, "use_itn: %d\n", sense_voice->use_itn);
fprintf(stdout, "----------offline moonshine model config----------\n");
fprintf(stdout, "preprocessor: %s\n", moonshine->preprocessor);
fprintf(stdout, "encoder: %s\n", moonshine->encoder);
fprintf(stdout, "uncached_decoder: %s\n", moonshine->uncached_decoder);
fprintf(stdout, "cached_decoder: %s\n", moonshine->cached_decoder);
fprintf(stdout, "----------offline FireRedAsr model config----------\n");
fprintf(stdout, "encoder: %s\n", fire_red_asr->encoder);
fprintf(stdout, "decoder: %s\n", fire_red_asr->decoder);
fprintf(stdout, "----------offline Dolphin model config----------\n");
fprintf(stdout, "model: %s\n", dolphin->model);
fprintf(stdout, "tokens: %s\n", model_config->tokens);
fprintf(stdout, "num_threads: %d\n", model_config->num_threads);
fprintf(stdout, "provider: %s\n", model_config->provider);
fprintf(stdout, "debug: %d\n", model_config->debug);
fprintf(stdout, "model type: %s\n", model_config->model_type);
fprintf(stdout, "modeling unit: %s\n", model_config->modeling_unit);
fprintf(stdout, "bpe vocab: %s\n", model_config->bpe_vocab);
fprintf(stdout, "telespeech_ctc: %s\n", model_config->telespeech_ctc);
fprintf(stdout, "----------feat config----------\n");
fprintf(stdout, "sample rate: %d\n", feat->sample_rate);
fprintf(stdout, "feat dim: %d\n", feat->feature_dim);
fprintf(stdout, "----------recognizer config----------\n");
fprintf(stdout, "decoding method: %s\n", config->decoding_method);
fprintf(stdout, "max active paths: %d\n", config->max_active_paths);
fprintf(stdout, "hotwords_file: %s\n", config->hotwords_file);
fprintf(stdout, "hotwords_score: %.2f\n", config->hotwords_score);
fprintf(stdout, "rule_fsts: %s\n", config->rule_fsts);
fprintf(stdout, "rule_fars: %s\n", config->rule_fars);
fprintf(stdout, "blank_penalty: %f\n", config->blank_penalty);
}
void CopyHeap(const char *src, int32_t num_bytes, char *dst) {
std::copy(src, src + num_bytes, dst);
}
}