offline-sense-voice-model.h
1.8 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
// sherpa-onnx/csrc/offline-sense-voice-model.h
//
// Copyright (c) 2022-2023 Xiaomi Corporation
#ifndef SHERPA_ONNX_CSRC_OFFLINE_SENSE_VOICE_MODEL_H_
#define SHERPA_ONNX_CSRC_OFFLINE_SENSE_VOICE_MODEL_H_
#include <memory>
#include <vector>
#if __ANDROID_API__ >= 9
#include "android/asset_manager.h"
#include "android/asset_manager_jni.h"
#endif
#include "onnxruntime_cxx_api.h" // NOLINT
#include "sherpa-onnx/csrc/offline-model-config.h"
#include "sherpa-onnx/csrc/offline-sense-voice-model-meta-data.h"
namespace sherpa_onnx {
class OfflineSenseVoiceModel {
public:
explicit OfflineSenseVoiceModel(const OfflineModelConfig &config);
#if __ANDROID_API__ >= 9
OfflineSenseVoiceModel(AAssetManager *mgr, const OfflineModelConfig &config);
#endif
~OfflineSenseVoiceModel();
/** Run the forward method of the model.
*
* @param features A tensor of shape (N, T, C). It is changed in-place.
* @param features_length A 1-D tensor of shape (N,) containing number of
* valid frames in `features` before padding.
* Its dtype is int32_t.
* @param language A 1-D tensor of shape (N,) with dtype int32_t
* @param text_norm A 1-D tensor of shape (N,) with dtype int32_t
*
* @return Return logits of shape (N, T, C) with dtype float
*
* Note: The subsampling factor is 1 for SenseVoice, so there is
* no need to output logits_length.
*/
Ort::Value Forward(Ort::Value features, Ort::Value features_length,
Ort::Value language, Ort::Value text_norm) const;
const OfflineSenseVoiceModelMetaData &GetModelMetadata() const;
/** Return an allocator for allocating memory
*/
OrtAllocator *Allocator() const;
private:
class Impl;
std::unique_ptr<Impl> impl_;
};
} // namespace sherpa_onnx
#endif // SHERPA_ONNX_CSRC_OFFLINE_SENSE_VOICE_MODEL_H_