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
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
|
/*
* Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef __CONTEXT_H__
#define __CONTEXT_H__
#include "Convert.h"
#include "TensorBags.h"
#include <coco/IR/Data.h>
#include <coco/IR/Module.h>
#include <schema_generated.h>
#include <map>
using namespace nncc::core::ADT;
namespace tflimport
{
/**
* @brief Extracts and holds operand(tensor) information such as name, shape, and type
*/
class TensorContext
{
public:
void prepare(const tflite::SubGraph *graph);
const std::string &name(uint32_t tensor_id) { return _name_ctx[tensor_id]; }
const tensor::Shape &shape(uint32_t tensor_id) { return _shape_ctx[tensor_id]; }
const tflite::TensorType &type(uint32_t tensor_id) { return _type_ctx[tensor_id]; }
private:
std::map<uint32_t, std::string> _name_ctx;
std::map<uint32_t, tensor::Shape> _shape_ctx;
std::map<uint32_t, tflite::TensorType> _type_ctx;
};
/**
* @brief Class that holds operator codes and related methods
*/
class TflOpCodeContext
{
public:
TflOpCodeContext(const flatbuffers::Vector<flatbuffers::Offset<tflite::OperatorCode>> *opcodes);
/**
* @brief Returns BuiltinOperator value of the operator
*/
tflite::BuiltinOperator builtin_code(const tflite::Operator *op) const;
/**
* @brief Returns human readable name of the operator code of the operator
*
* @note TF lite InterpreterBuilder sets an error state and returns error code
* for invalid opcode. Here we just return human readable message as
* this method returns a name for the operator code.
*/
std::string opcode_name(const tflite::Operator *op) const;
public:
static bool is_valid(const tflite::OperatorCode *opcode);
static bool is_custom(const tflite::OperatorCode *opcode);
private:
std::vector<const tflite::OperatorCode *> _opcodes;
};
/**
* @brief Class to read and provide buffer information of tflite
*/
class TflBufferContext
{
public:
template <typename T> struct TflBuffer
{
TflBuffer(const T *p, size_t s) : ptr{p}, len{s} {};
const T *ptr;
size_t len;
};
public:
explicit TflBufferContext(const tflite::Model *tfl_model);
public:
template <typename T>
TflBuffer<T> tensor_buffer(const tflite::SubGraph *graph, uint32_t tensor_idx) const
{
TflBufferContext::TflBuffer<T> res{nullptr, 0};
const auto *tensor = graph->tensors()->Get(tensor_idx);
uint32_t tfl_buf_id = tensor->buffer();
assert(_buffer_ctx.size() > tfl_buf_id);
const tflite::Buffer *tfl_buffer = _buffer_ctx.at(tfl_buf_id);
if (auto *array = tfl_buffer->data())
{
if (size_t size = array->size())
{
assert(size % sizeof(T) == 0);
res.len = size / sizeof(T);
res.ptr = reinterpret_cast<const T *>(array->data());
}
}
return res;
}
private:
std::map<uint32_t /* Buffer ID */, const tflite::Buffer *> _buffer_ctx;
};
/**
* @brief Class to store context to build IR from tflite
*/
class GraphBuilderContext
{
public:
explicit GraphBuilderContext(coco::Module *m, coco::Data *d, coco::Block *block,
TensorBags &tensor_bags, TensorContext &tensor_context,
TflBufferContext &buffer_context, const tflite::SubGraph *graph)
: _m(m), _d(d), _block(block), _tensor_bags(tensor_bags), _tensor_context(tensor_context),
_buffer_context(buffer_context), _graph(graph)
{
// DO NOTHING
}
GraphBuilderContext() = delete;
GraphBuilderContext(const GraphBuilderContext &) = delete;
GraphBuilderContext(GraphBuilderContext &&) = delete;
public:
coco::Module *m() { return _m; }
coco::Data *d() { return _d; }
coco::Block *block() { return _block; }
TensorContext &tensor() { return _tensor_context; }
TensorBags &bags() { return _tensor_bags; }
TflBufferContext &buffer() { return _buffer_context; }
const tflite::SubGraph *graph() { return _graph; }
private:
coco::Module *_m;
coco::Data *_d;
coco::Block *_block;
TensorContext &_tensor_context;
TensorBags &_tensor_bags;
TflBufferContext &_buffer_context;
const tflite::SubGraph *_graph;
};
} // namespace tflimport
#endif // __CONTEXT_H__
|
