forked from mingj2021/segment-anything-tensorrt
-
Notifications
You must be signed in to change notification settings - Fork 0
/
baseModel.h
84 lines (71 loc) · 2.66 KB
/
baseModel.h
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
#ifndef BASEMODEL_H
#define BASEMODEL_H
#include "sam_utils.h"
class BaseModel
{
public:
cudaStream_t stream;
std::shared_ptr<nvinfer1::ICudaEngine> mEngine;
std::unique_ptr<nvinfer1::IExecutionContext> context;
std::vector<void *> mDeviceBindings;
std::map<std::string, std::unique_ptr<algorithms::DeviceBuffer>> mInOut;
std::vector<std::string> mInputsName, mOutputsName;
public:
BaseModel(std::string modelFile);
~BaseModel();
void read_engine_file(std::string modelFile);
// std::vector<std::string> get_inputs_name();
// std::vector<std::string> get_outputs_name();
// std::vector<void *> get_device_buffer();
};
BaseModel::BaseModel(std::string modelFile)
{
read_engine_file(modelFile);
context = std::unique_ptr<nvinfer1::IExecutionContext>(mEngine->createExecutionContext());
if (!context)
{
std::cerr << "create context error" << std::endl;
}
CHECK(cudaStreamCreate(&stream));
for (int i = 0; i < mEngine->getNbBindings(); i++)
{
auto dims = mEngine->getBindingDimensions(i);
auto tensor_name = mEngine->getBindingName(i);
bool isInput = mEngine->bindingIsInput(i);
if (isInput)
mInputsName.emplace_back(tensor_name);
else
mOutputsName.emplace_back(tensor_name);
std::cout << "tensor_name: " << tensor_name << std::endl;
dims2str(dims);
nvinfer1::DataType type = mEngine->getBindingDataType(i);
index2srt(type);
auto vol = std::accumulate(dims.d, dims.d + dims.nbDims, int64_t{1}, std::multiplies<int64_t>{});
std::unique_ptr<algorithms::DeviceBuffer> device_buffer{new algorithms::DeviceBuffer(vol, type)};
mDeviceBindings.emplace_back(device_buffer->data());
mInOut[tensor_name] = std::move(device_buffer);
}
}
BaseModel::~BaseModel()
{
CHECK(cudaStreamDestroy(stream));
}
void BaseModel::read_engine_file(std::string modelFile)
{
std::ifstream engineFile(modelFile.c_str(), std::ifstream::binary);
assert(engineFile);
int fsize;
engineFile.seekg(0, engineFile.end);
fsize = engineFile.tellg();
engineFile.seekg(0, engineFile.beg);
std::vector<char> engineData(fsize);
engineFile.read(engineData.data(), fsize);
if (engineFile)
std::cout << "all characters read successfully." << std::endl;
else
std::cout << "error: only " << engineFile.gcount() << " could be read" << std::endl;
engineFile.close();
std::unique_ptr<nvinfer1::IRuntime> runtime(nvinfer1::createInferRuntime(logger));
mEngine = std::shared_ptr<nvinfer1::ICudaEngine>(runtime->deserializeCudaEngine(engineData.data(), fsize, nullptr));
}
#endif