onnx.cpp 25.2 KB
Newer Older
Paul's avatar
Paul committed
1
2
3
4
5
6
7
8
#include <google/protobuf/text_format.h>
#include <google/protobuf/io/zero_copy_stream_impl.h>
#include <onnx.pb.h>
#include <iostream>
#include <fstream>
#include <unordered_map>
#include <functional>
#include <array>
Paul's avatar
Paul committed
9
#include <utility>
10
#include <vector>
Paul's avatar
Paul committed
11

Paul's avatar
Paul committed
12
13
14
15
#include <migraph/fallthrough.hpp>
#include <migraph/program.hpp>
#include <migraph/operators.hpp>
#include <migraph/ranges.hpp>
16
#include <migraph/instruction.hpp>
17
#include <migraph/config.hpp>
Paul's avatar
Paul committed
18

Paul's avatar
Paul committed
19
namespace migraph {
20
inline namespace MIGRAPH_INLINE_NS {
Paul's avatar
Paul committed
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
struct unknown
{
    std::string op;
    std::string name() const { return "unknown:" + op; }
    shape compute_shape(std::vector<shape> input) const
    {
        if(input.empty())
            return {};
        else
            return input.front();
    }
    friend std::ostream& operator<<(std::ostream& os, const unknown& x)
    {
        os << x.name();
        return os;
    }
};

struct onnx_parser
{
    using attribute_map = std::unordered_map<std::string, onnx::AttributeProto>;
    using node_map      = std::unordered_map<std::string, onnx::NodeProto>;
Paul's avatar
Paul committed
43
    using op_func = std::function<instruction_ref(attribute_map, std::vector<instruction_ref>)>;
Paul's avatar
Paul committed
44
45
46
47
48
49
50
51
    node_map nodes;
    std::unordered_map<std::string, instruction_ref> instructions;
    program prog = program();

    std::unordered_map<std::string, op_func> ops;

    onnx_parser()
    {
52
53
        add_generic_op("Add", op::add{});
        add_generic_op("Div", op::div{});
Shucai Xiao's avatar
Shucai Xiao committed
54
        add_generic_op("MatMul", op::dot{});
55
        add_generic_op("Mul", op::mul{});
Khalique's avatar
Khalique committed
56
        add_generic_op("Relu", op::relu{});
57
58
        add_generic_op("Sub", op::sub{});
        add_generic_op("Sum", op::add{});
Khalique's avatar
Khalique committed
59
60
        // disable dropout for inference
        add_generic_op("Dropout", op::identity{});
Paul's avatar
Paul committed
61

Khalique's avatar
Khalique committed
62
        add_mem_op("LRN", &onnx_parser::parse_lrn);
Khalique's avatar
Khalique committed
63
        add_mem_op("ImageScaler", &onnx_parser::parse_imagescaler);
64
        add_mem_op("LeakyRelu", &onnx_parser::parse_leaky_relu);
Paul's avatar
Paul committed
65
66
        add_mem_op("Constant", &onnx_parser::parse_constant);
        add_mem_op("Conv", &onnx_parser::parse_conv);
Paul's avatar
Paul committed
67
68
        add_mem_op("MaxPool", &onnx_parser::parse_pooling);
        add_mem_op("AveragePool", &onnx_parser::parse_pooling);
69
70
        add_mem_op("GlobalMaxPool", &onnx_parser::parse_pooling);
        add_mem_op("GlobalAveragePool", &onnx_parser::parse_pooling);
Paul's avatar
Paul committed
71
        add_mem_op("Reshape", &onnx_parser::parse_reshape);
Paul's avatar
Paul committed
72
73
        add_mem_op("Flatten", &onnx_parser::parse_flatten);
        add_mem_op("Gemm", &onnx_parser::parse_gemm);
74
        add_mem_op("BatchNormalization", &onnx_parser::parse_batchnorm);
Paul's avatar
Paul committed
75
        add_mem_op("Softmax", &onnx_parser::parse_softmax);
76
77
78
        add_mem_op("Squeeze", &onnx_parser::parse_squeeze);
        add_mem_op("Unsqueeze", &onnx_parser::parse_unsqueeze);
        add_mem_op("Slice", &onnx_parser::parse_slice);
Scott Thornton's avatar
Scott Thornton committed
79
        add_mem_op("Concat", &onnx_parser::parse_concat);
Khalique's avatar
Khalique committed
80
        add_mem_op("Transpose", &onnx_parser::parse_transpose);
Paul's avatar
Paul committed
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
    }

    template <class F>
    void add_op(std::string name, F f)
    {
        ops.emplace(name, f);
    }

    template <class F>
    void add_mem_op(std::string name, F f)
    {
        ops.emplace(name, [=](auto&&... xs) {
            return std::mem_fn(f)(*this, name, std::forward<decltype(xs)>(xs)...);
        });
    }

Paul's avatar
Paul committed
97
    template <class T>
Paul's avatar
Paul committed
98
99
100
101
102
103
104
105
106
107
108
    void add_generic_op(std::string name, T x)
    {
        ops.emplace(name, [this, x](attribute_map attributes, std::vector<instruction_ref> args) {
            if(args.size() == 2 and contains(attributes, "broadcast"))
            {
                uint64_t broadcasted = parse_value(attributes.at("broadcast")).at<uint64_t>();
                if(broadcasted != 0)
                {
                    uint64_t axis = (contains(attributes, "axis"))
                                        ? parse_value(attributes.at("axis")).at<uint64_t>()
                                        : 0;
Scott Thornton's avatar
Scott Thornton committed
109
110
                    auto l =
                        prog.add_instruction(op::broadcast{axis, args[0]->get_shape()}, args[1]);
Paul's avatar
Paul committed
111
112
113
114
115
116
117
                    return prog.add_instruction(x, args[0], l);
                }
            }
            return prog.add_instruction(x, args);
        });
    }

Paul's avatar
Paul committed
118
    instruction_ref
Paul's avatar
Paul committed
119
    parse_softmax(const std::string&, const attribute_map&, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
120
121
    {
        auto dims = args.front()->get_shape().lens();
Scott Thornton's avatar
Scott Thornton committed
122
123
        auto r =
            prog.add_instruction(op::reshape{{long(dims[0]), long(dims[1]), 1, 1}}, args.front());
124
125
        auto s = prog.add_instruction(op::softmax{}, r);
        return prog.add_instruction(op::reshape{{long(dims[0]), long(dims[1])}}, s);
Paul's avatar
Paul committed
126
127
    }

Paul's avatar
Paul committed
128
    instruction_ref
Paul's avatar
Paul committed
129
    parse_conv(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
130
    {
131
        op::convolution op;
Paul's avatar
Paul committed
132
133
134
        if(contains(attributes, "pads"))
        {
            copy(attributes["pads"].ints(), op.padding.begin());
Paul's avatar
Paul committed
135
        }
Paul's avatar
Paul committed
136
137
138
139
140
141
142
143
144
145
146
147
        if(contains(attributes, "strides"))
        {
            copy(attributes["strides"].ints(), op.stride.begin());
        }
        if(contains(attributes, "dilations"))
        {
            copy(attributes["dilations"].ints(), op.dilation.begin());
        }
        if(args.size() == 3)
        {
            uint64_t axis = 1;
            auto l1       = prog.add_instruction(op, args[0], args[1]);
Scott Thornton's avatar
Scott Thornton committed
148
            auto l2       = prog.add_instruction(op::broadcast{axis, l1->get_shape()}, args[2]);
149
            return prog.add_instruction(op::add{}, l1, l2);
Paul's avatar
Paul committed
150
        }
Paul's avatar
Paul committed
151
152
        return prog.add_instruction(op, args);
    }
Paul's avatar
Paul committed
153

Paul's avatar
Paul committed
154
155
156
    instruction_ref parse_pooling(const std::string& name,
                                  attribute_map attributes,
                                  std::vector<instruction_ref> args)
Paul's avatar
Paul committed
157
    {
Khalique's avatar
Khalique committed
158
159
        op::pooling op{ends_with(name, "MaxPool") ? "max" : "average"};
        if(starts_with(name, "Global"))
160
        {
Khalique's avatar
Khalique committed
161
162
            auto lens  = args.front()->get_shape().lens();
            op.lengths = {lens[2], lens[3]};
163
        }
Paul's avatar
Paul committed
164
165
166
167
168
169
170
171
172
173
174
175
        if(contains(attributes, "pads"))
        {
            copy(attributes["pads"].ints(), op.padding.begin());
        }
        if(contains(attributes, "strides"))
        {
            copy(attributes["strides"].ints(), op.stride.begin());
        }
        if(contains(attributes, "kernel_shape"))
        {
            copy(attributes["kernel_shape"].ints(), op.lengths.begin());
        }
Paul's avatar
Paul committed
176
        return prog.add_instruction(op, std::move(args));
Paul's avatar
Paul committed
177
178
    }

Paul's avatar
Paul committed
179
    instruction_ref
Paul's avatar
Paul committed
180
    parse_reshape(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
181
    {
182
        op::reshape op;
Paul's avatar
Paul committed
183
184
185
186
187
188
189
        if(args.size() == 1)
        {
            literal s = parse_value(attributes.at("shape"));
            s.visit([&](auto v) { copy(v, std::back_inserter(op.dims)); });
        }
        if(args.size() == 2)
        {
Paul's avatar
Paul committed
190
            literal s = args[1]->get_literal();
Paul's avatar
Paul committed
191
            s.visit([&](auto v) { copy(v, std::back_inserter(op.dims)); });
Paul's avatar
Paul committed
192
        }
Paul's avatar
Paul committed
193
194
195
        return prog.add_instruction(op, args[0]);
    }

Paul's avatar
Paul committed
196
    instruction_ref
Paul's avatar
Paul committed
197
    parse_flatten(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
198
199
    {
        uint64_t axis = 0;
Paul's avatar
Paul committed
200
201
202
203
        if(contains(attributes, "axis"))
        {
            axis = parse_value(attributes.at("axis")).at<int>();
        }
204
        return prog.add_instruction(op::flatten{axis}, args[0]);
Paul's avatar
Paul committed
205
206
    }

207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
    instruction_ref
    parse_squeeze(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        op::squeeze op;
        literal s = parse_value(attributes.at("axes"));
        s.visit([&](auto v) { copy(v, std::back_inserter(op.axes)); });
        return prog.add_instruction(op, args[0]);
    }

    instruction_ref
    parse_unsqueeze(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        op::unsqueeze op;
        literal s = parse_value(attributes.at("axes"));
        s.visit([&](auto v) { copy(v, std::back_inserter(op.axes)); });
        return prog.add_instruction(op, args[0]);
    }

Scott Thornton's avatar
Scott Thornton committed
225
226
227
228
229
230
231
    instruction_ref
    parse_concat(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        std::size_t axis = parse_value(attributes.at("axis")).at<int>();
        op::concat op{axis};
        return prog.add_instruction(op, std::move(args));
    }
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252

    instruction_ref
    parse_slice(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        op::slice op;
        if(contains(attributes, "axes"))
        {
            literal s = parse_value(attributes.at("axes"));
            s.visit([&](auto v) { copy(v, std::back_inserter(op.axes)); });
        }
        {
            literal s = parse_value(attributes.at("ends"));
            s.visit([&](auto v) { copy(v, std::back_inserter(op.ends)); });
        }
        {
            literal s = parse_value(attributes.at("starts"));
            s.visit([&](auto v) { copy(v, std::back_inserter(op.starts)); });
        }
        return prog.add_instruction(op, args[0]);
    }

Paul's avatar
Paul committed
253
254
255
    instruction_ref parse_constant(const std::string&,
                                   attribute_map attributes,
                                   const std::vector<instruction_ref>&)
Paul's avatar
Paul committed
256
257
258
259
    {
        literal v = parse_value(attributes.at("value"));
        return prog.add_literal(v);
    }
Paul's avatar
Paul committed
260

Paul's avatar
Paul committed
261
    instruction_ref
Paul's avatar
Paul committed
262
    parse_gemm(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
    {
        float alpha = 1.0f;
        float beta  = 0.0f;
        bool transa = false;
        bool transb = false;
        if(contains(attributes, "alpha"))
        {
            alpha = parse_value(attributes.at("alpha")).at<float>();
        }
        if(contains(attributes, "beta"))
        {
            alpha = parse_value(attributes.at("beta")).at<float>();
        }
        if(contains(attributes, "transA"))
        {
            transa = parse_value(attributes.at("transA")).at<bool>();
        }
        if(contains(attributes, "transB"))
        {
            transb = parse_value(attributes.at("transB")).at<bool>();
        }
        std::vector<int64_t> perm = {1, 0};
285
286
        auto l1 = (transa) ? prog.add_instruction(op::transpose{perm}, args[0]) : args[0];
        auto l2 = (transb) ? prog.add_instruction(op::transpose{perm}, args[1]) : args[1];
Paul's avatar
Paul committed
287
288
289
        if(args.size() == 3)
        {
            uint64_t axis = 1;
Shucai Xiao's avatar
Shucai Xiao committed
290
            auto l3       = prog.add_instruction(op::dot{alpha, beta}, l1, l2);
Scott Thornton's avatar
Scott Thornton committed
291
            auto l4       = prog.add_instruction(op::broadcast{axis, l3->get_shape()}, args[2]);
292
            return prog.add_instruction(op::add{}, l3, l4);
Paul's avatar
Paul committed
293
        }
Shucai Xiao's avatar
Shucai Xiao committed
294
        return prog.add_instruction(op::dot{alpha, beta}, l1, l2);
Paul's avatar
Paul committed
295
296
    }

297
    instruction_ref
Paul's avatar
Paul committed
298
    parse_batchnorm(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
299
    {
Scott Thornton's avatar
Scott Thornton committed
300
301
        float epsilon                                     = 1e-5f;
        float momentum                                    = 0.9f;
302
        op::batch_norm_inference::bn_infer_mode_t bn_mode = op::batch_norm_inference::spatial;
Scott Thornton's avatar
Scott Thornton committed
303
        bool is_test                                      = false;
304
305
306
307
308
309
        if(contains(attributes, "epsilon"))
        {
            epsilon = parse_value(attributes.at("epsilon")).at<float>();
        }
        if(contains(attributes, "momentum"))
        {
310
            momentum = parse_value(attributes.at("momentum")).at<float>();
311
312
313
        }
        if(contains(attributes, "is_test"))
        {
wsttiger's avatar
wsttiger committed
314
            is_test = parse_value(attributes.at("is_test")).at<uint64_t>() > 0;
315
316
317
        }
        if(contains(attributes, "spatial"))
        {
318
            bn_mode = (parse_value(attributes.at("spatial")).at<uint64_t>() > 0)
319
320
                          ? op::batch_norm_inference::spatial
                          : op::batch_norm_inference::per_activation;
321
        }
Paul's avatar
Paul committed
322
        (void)is_test;
Paul's avatar
Paul committed
323
        op::batch_norm_inference op{epsilon, momentum, bn_mode};
Paul's avatar
Paul committed
324
        return prog.add_instruction(op, std::move(args));
325
326
    }

327
328
329
330
    instruction_ref parse_leaky_relu(const std::string&,
                                     attribute_map attributes,
                                     std::vector<instruction_ref> args)
    {
Khalique's avatar
Khalique committed
331
        float alpha = 0.01; // default alpha val for leaky relu
332
333
334
335
336
337
338
339
        if(contains(attributes, "alpha"))
        {
            alpha = parse_value(attributes.at("alpha")).at<float>();
        }
        op::leaky_relu op{alpha};
        return prog.add_instruction(op, args.front());
    }

Khalique's avatar
Khalique committed
340
341
    instruction_ref
    parse_lrn(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Khalique's avatar
Khalique committed
342
343
    {
        float alpha = 0.0001;
Khalique's avatar
Khalique committed
344
345
346
        float beta  = 0.75;
        float bias  = 1.0;
        int size    = 1;
Khalique's avatar
Khalique committed
347
348
349
350
351
352
353
354
355
356
357
358
        if(contains(attributes, "alpha"))
            alpha = parse_value(attributes.at("alpha")).at<float>();
        if(contains(attributes, "beta"))
            beta = parse_value(attributes.at("beta")).at<float>();
        if(contains(attributes, "bias"))
            bias = parse_value(attributes.at("bias")).at<float>();
        if(contains(attributes, "size"))
            size = parse_value(attributes.at("size")).at<int>();
        op::lrn op{alpha, beta, bias, size};
        return prog.add_instruction(op, args.front());
    }

Khalique's avatar
Khalique committed
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
    instruction_ref parse_imagescaler(const std::string&,
                                      attribute_map attributes,
                                      std::vector<instruction_ref> args)
    {
        float scale = 1.0;
        std::vector<float> bias{};
        if(contains(attributes, "scale"))
        {
            scale = parse_value(attributes.at("scale")).at<float>();
        }

        if(contains(attributes, "bias"))
        {
            auto&& bias_floats = attributes["bias"].floats();
            bias               = std::vector<float>(bias_floats.begin(), bias_floats.end());
        }
        auto input_shape = args.front()->get_shape();
Khalique's avatar
Khalique committed
376

Khalique's avatar
Khalique committed
377
378
379
380
381
382
383
384
385
        auto scale_val = prog.add_literal(scale);
        auto bias_vals = prog.add_literal(
            migraph::literal{migraph::shape{migraph::shape::float_type, {bias.size()}}, bias});

        auto scale_tensor = prog.add_instruction(migraph::op::scalar{input_shape}, scale_val);
        auto img_scaled   = prog.add_instruction(migraph::op::mul{}, args.front(), scale_tensor);
        auto bias_bcast   = prog.add_instruction(migraph::op::broadcast{1, input_shape}, bias_vals);
        return prog.add_instruction(migraph::op::add{}, img_scaled, bias_bcast);
    }
Khalique's avatar
Khalique committed
386

Khalique's avatar
Khalique committed
387
388
    instruction_ref
    parse_transpose(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Khalique's avatar
Khalique committed
389
390
391
392
393
394
395
396
397
398
    {
        std::vector<int64_t> perm{};
        if(contains(attributes, "perm"))
        {
            auto&& perm_vals = attributes["perm"].ints();
            perm             = std::vector<int64_t>(perm_vals.begin(), perm_vals.end());
        }
        return prog.add_instruction(migraph::op::transpose{perm}, args.front());
    }

Paul's avatar
Paul committed
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
    void parse_from(std::istream& is)
    {
        onnx::ModelProto model;
        if(model.ParseFromIstream(&is))
        {
            if(model.has_graph())
            {
                this->parse_graph(model.graph());
            }
        }
        else
        {
            throw std::runtime_error("Failed reading");
        }
    }

    void parse_graph(const onnx::GraphProto& graph)
    {
        nodes = get_nodes(graph);
418
419
420
421
422
        std::unordered_map<std::string, onnx::TensorProto> initializer_data;
        for(auto&& f : graph.initializer())
        {
            initializer_data[f.name()] = f;
        }
Paul's avatar
Paul committed
423
424
425
        for(auto&& input : graph.input())
        {
            const std::string& name = input.name();
426
427
428
429
430
431
432
433
434
435
436
437
            // Does the input have an initializer?
            if(contains(initializer_data, name))
            {
                auto t             = initializer_data[name];
                instructions[name] = prog.add_literal(parse_tensor(t));
            }
            else
            {
                // TODO: Get shape of input parameter
                shape s            = parse_type(input.type());
                instructions[name] = prog.add_parameter(name, s);
            }
Paul's avatar
Paul committed
438
439
440
        }
        for(auto&& p : nodes)
        {
441
            this->parse_node(get_name(p.second));
Paul's avatar
Paul committed
442
443
444
        }
    }

Paul's avatar
Paul committed
445
    void parse_node(const std::string& name)
Paul's avatar
Paul committed
446
    {
Paul's avatar
Paul committed
447
        if(name.empty())
Paul's avatar
Paul committed
448
            MIGRAPH_THROW("Onnx node must have a name");
Paul's avatar
Paul committed
449
450
451
452
453
454
455
456
        if(instructions.count(name) == 0)
        {
            auto&& node = nodes.at(name);
            std::vector<instruction_ref> args;
            for(auto&& input : node.input())
            {
                if(nodes.count(input) > 0)
                {
457
                    auto&& iname = get_name(nodes.at(input));
Paul's avatar
Paul committed
458
                    assert(name != iname);
Paul's avatar
Paul committed
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
                    this->parse_node(iname);
                    args.push_back(instructions.at(iname));
                }
                else
                {
                    args.push_back(instructions.at(input));
                }
            }
            if(ops.count(node.op_type()) == 0)
            {
                instructions[name] = prog.add_instruction(unknown{node.op_type()}, args);
            }
            else
            {
                instructions[name] = ops[node.op_type()](get_attributes(node), args);
            }
        }
    }

    static attribute_map get_attributes(const onnx::NodeProto& node)
    {
        std::unordered_map<std::string, onnx::AttributeProto> result;
        for(auto&& attr : node.attribute())
        {
            result[attr.name()] = attr;
        }
        return result;
    }

488
489
490
491
492
    static std::string get_name(const onnx::NodeProto& node)
    {
        if(node.name().empty())
        {
            std::string generated = "migraph_unnamed_node";
Paul's avatar
Paul committed
493
494
495
496
            return std::accumulate(node.output().begin(),
                                   node.output().end(),
                                   generated,
                                   [](auto x, auto y) { return x + "_" + y; });
497
498
499
500
        }
        return node.name();
    }

Paul's avatar
Paul committed
501
502
503
504
505
    static node_map get_nodes(const onnx::GraphProto& graph)
    {
        std::unordered_map<std::string, onnx::NodeProto> result;
        for(auto&& node : graph.node())
        {
506
            result[get_name(node)] = node;
Paul's avatar
Paul committed
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
            for(auto&& output : node.output())
            {
                result[output] = node;
            }
        }
        return result;
    }

    template <class T>
    static literal from_repeated(shape::type_t t, const T& r)
    {
        std::size_t size = r.size();
        return literal{{t, {size}}, r.begin(), r.end()};
    }

    static literal parse_value(const onnx::AttributeProto& attr)
    {
        switch(attr.type())
        {
        case onnx::AttributeProto::UNDEFINED: return {};
        case onnx::AttributeProto::FLOAT: return literal{attr.f()};
        case onnx::AttributeProto::INT: return literal{attr.i()};
        case onnx::AttributeProto::STRING: return {};
        case onnx::AttributeProto::TENSOR: return parse_tensor(attr.t());
        case onnx::AttributeProto::GRAPH: return {};
Paul's avatar
Paul committed
532
        case onnx::AttributeProto::FLOATS: return from_repeated(shape::float_type, attr.floats());
Paul's avatar
Paul committed
533
534
535
536
537
        case onnx::AttributeProto::INTS: return from_repeated(shape::int64_type, attr.ints());
        case onnx::AttributeProto::STRINGS: return {};
        case onnx::AttributeProto::TENSORS: return {};
        case onnx::AttributeProto::GRAPHS: return {};
        }
Paul's avatar
Paul committed
538
        MIGRAPH_THROW("Invalid attribute type");
Paul's avatar
Paul committed
539
540
541
542
543
    }

    static literal parse_tensor(const onnx::TensorProto& t)
    {
        std::vector<std::size_t> dims(t.dims().begin(), t.dims().end());
544
545
        if(t.has_raw_data())
        {
wsttiger's avatar
wsttiger committed
546
            const std::string& s = t.raw_data();
Scott Thornton's avatar
Scott Thornton committed
547
548
549
550
551
552
553
554
555
556
557
558
            switch(t.data_type())
            {
            case onnx::TensorProto::UNDEFINED: throw std::runtime_error("");
            case onnx::TensorProto::FLOAT: return literal{{shape::float_type, dims}, s.data()};
            case onnx::TensorProto::UINT8: throw std::runtime_error("");
            case onnx::TensorProto::INT8: return literal{{shape::int32_type, dims}, s.data()};
            case onnx::TensorProto::UINT16: return literal{{shape::int32_type, dims}, s.data()};
            case onnx::TensorProto::INT16: return literal{{shape::int32_type, dims}, s.data()};
            case onnx::TensorProto::INT32: return literal{{shape::int32_type, dims}, s.data()};
            case onnx::TensorProto::INT64: return literal{{shape::int64_type, dims}, s.data()};
            case onnx::TensorProto::STRING: throw std::runtime_error("");
            case onnx::TensorProto::BOOL: return literal{{shape::int32_type, dims}, s.data()};
Paul's avatar
Paul committed
559
            case onnx::TensorProto::FLOAT16: return literal{{shape::half_type, dims}, s.data()};
Scott Thornton's avatar
Scott Thornton committed
560
561
562
563
564
565
566
            case onnx::TensorProto::DOUBLE: return literal{{shape::double_type, dims}, s.data()};
            case onnx::TensorProto::UINT32: throw std::runtime_error("");
            case onnx::TensorProto::UINT64: throw std::runtime_error("");
            case onnx::TensorProto::COMPLEX64: throw std::runtime_error("");
            case onnx::TensorProto::COMPLEX128: throw std::runtime_error("");
            }
            MIGRAPH_THROW("Invalid tensor type");
567
        }
Paul's avatar
Paul committed
568
569
570
571
        switch(t.data_type())
        {
        case onnx::TensorProto::UNDEFINED: throw std::runtime_error("");
        case onnx::TensorProto::FLOAT:
Paul's avatar
Paul committed
572
            return literal{{shape::float_type, dims}, t.float_data().begin(), t.float_data().end()};
Paul's avatar
Paul committed
573
574
        case onnx::TensorProto::UINT8: throw std::runtime_error("");
        case onnx::TensorProto::INT8:
Paul's avatar
Paul committed
575
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
576
        case onnx::TensorProto::UINT16:
Paul's avatar
Paul committed
577
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
578
        case onnx::TensorProto::INT16:
Paul's avatar
Paul committed
579
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
580
        case onnx::TensorProto::INT32:
Paul's avatar
Paul committed
581
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
582
        case onnx::TensorProto::INT64:
Paul's avatar
Paul committed
583
            return literal{{shape::int64_type, dims}, t.int64_data().begin(), t.int64_data().end()};
Paul's avatar
Paul committed
584
585
        case onnx::TensorProto::STRING: throw std::runtime_error("");
        case onnx::TensorProto::BOOL:
Paul's avatar
Paul committed
586
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
587
588
        case onnx::TensorProto::FLOAT16:
            return literal{{shape::half_type, dims}, t.float_data().begin(), t.float_data().end()};
Paul's avatar
Paul committed
589
590
591
592
593
594
595
596
        case onnx::TensorProto::DOUBLE:
            return literal{
                {shape::double_type, dims}, t.double_data().begin(), t.double_data().end()};
        case onnx::TensorProto::UINT32: throw std::runtime_error("");
        case onnx::TensorProto::UINT64: throw std::runtime_error("");
        case onnx::TensorProto::COMPLEX64: throw std::runtime_error("");
        case onnx::TensorProto::COMPLEX128: throw std::runtime_error("");
        }
Paul's avatar
Paul committed
597
        MIGRAPH_THROW("Invalid tensor type");
Paul's avatar
Paul committed
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
    }

    static shape parse_type(const onnx::TypeProto& t)
    {
        shape::type_t shape_type{};
        switch(t.tensor_type().elem_type())
        {
        case onnx::TensorProto::UNDEFINED:
            break; // throw std::runtime_error("Unsupported type UNDEFINED");
        case onnx::TensorProto::FLOAT: shape_type = shape::float_type; break;
        case onnx::TensorProto::UINT8:
            break; // throw std::runtime_error("Unsupported type UINT8");
        case onnx::TensorProto::INT8: shape_type = shape::int8_type; break;
        case onnx::TensorProto::UINT16: shape_type = shape::uint16_type; break;
        case onnx::TensorProto::INT16: shape_type = shape::int16_type; break;
        case onnx::TensorProto::INT32: shape_type = shape::int32_type; break;
        case onnx::TensorProto::INT64: shape_type = shape::int64_type; break;
        case onnx::TensorProto::STRING:
            break; // throw std::runtime_error("Unsupported type STRING");
        case onnx::TensorProto::BOOL:
            break; // throw std::runtime_error("Unsupported type BOOL");
Paul's avatar
Paul committed
619
        case onnx::TensorProto::FLOAT16: shape_type = shape::half_type; break;
Paul's avatar
Paul committed
620
621
622
623
624
625
626
627
628
        case onnx::TensorProto::DOUBLE: shape_type = shape::double_type; break;
        case onnx::TensorProto::UINT32: shape_type = shape::uint32_type; break;
        case onnx::TensorProto::UINT64: shape_type = shape::uint64_type; break;
        case onnx::TensorProto::COMPLEX64:
            break; // throw std::runtime_error("Unsupported type COMPLEX64");
        case onnx::TensorProto::COMPLEX128:
            break; // throw std::runtime_error("Unsupported type COMPLEX128");
        }
        std::vector<std::size_t> dims;
Paul's avatar
Paul committed
629
        auto&& tensor_dims = t.tensor_type().shape().dim();
630
631
632
633
634
635
636
637
638
        std::transform(
            tensor_dims.begin(), tensor_dims.end(), std::back_inserter(dims), [](auto&& d) {
                if(not d.has_dim_value())
                {
                    long default_batch_size = 1; // FIXME
                    return default_batch_size;
                }
                return d.dim_value();
            });
Paul's avatar
Paul committed
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
        return {shape_type, dims};
    }
};

program parse_onnx(const std::string& name)
{
    std::fstream input(name.c_str(), std::ios::in | std::ios::binary);
    onnx_parser parser;
#ifndef NDEBUG
    // Log the program when it can't be parsed
    try
    {
        parser.parse_from(input);
    }
    catch(...)
    {
        std::cerr << parser.prog << std::endl;
        throw;
    }
#else
    parser.parse_from(input);
#endif
    return std::move(parser.prog);
}

664
} // namespace MIGRAPH_INLINE_NS
Paul's avatar
Paul committed
665
} // namespace migraph