onnx.cpp 53 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
16
17
#include <migraphx/fallthrough.hpp>
#include <migraphx/program.hpp>
#include <migraphx/operators.hpp>
#include <migraphx/ranges.hpp>
#include <migraphx/instruction.hpp>
#include <migraphx/config.hpp>
18
#include <migraphx/onnx.hpp>
Paul's avatar
Paul committed
19
20

namespace migraphx {
Paul's avatar
Paul committed
21
inline namespace MIGRAPHX_INLINE_NS {
Paul's avatar
Paul committed
22
23
24
25
26

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
27
28
    using op_func =
        std::function<std::vector<instruction_ref>(attribute_map, std::vector<instruction_ref>)>;
Paul's avatar
Paul committed
29
30
    node_map nodes;
    std::unordered_map<std::string, instruction_ref> instructions;
Scott Thornton's avatar
Scott Thornton committed
31
    program prog    = program();
32
    bool is_pytorch = false;
Paul's avatar
Paul committed
33
34

    std::unordered_map<std::string, op_func> ops;
35
    std::unordered_map<std::string, operation> map_actv_funcs;
Paul's avatar
Paul committed
36
37
38

    onnx_parser()
    {
Shucai Xiao's avatar
Shucai Xiao committed
39
        add_generic_op("MatMul", op::dot{});
Khalique's avatar
Khalique committed
40
        add_generic_op("Relu", op::relu{});
Khalique's avatar
Khalique committed
41
42
        add_generic_op("Sigmoid", op::sigmoid{});
        add_generic_op("Abs", op::abs{});
Shucai Xiao's avatar
Shucai Xiao committed
43
44
        add_generic_op("Exp", op::exp{});
        add_generic_op("Log", op::log{});
Khalique's avatar
Khalique committed
45
46
        // disable dropout for inference
        add_generic_op("Dropout", op::identity{});
Khalique's avatar
Khalique committed
47
        add_generic_op("Identity", op::identity{});
Shucai Xiao's avatar
Shucai Xiao committed
48
49
50
        add_generic_op("Sin", op::sin{});
        add_generic_op("Cos", op::cos{});
        add_generic_op("Tan", op::tan{});
51
52
        add_generic_op("Sinh", op::sinh{});
        add_generic_op("Cosh", op::cosh{});
53
        add_generic_op("Tanh", op::tanh{});
54
55
56
        add_generic_op("Asin", op::asin{});
        add_generic_op("Acos", op::acos{});
        add_generic_op("Atan", op::atan{});
Paul's avatar
Paul committed
57

Khalique's avatar
Khalique committed
58
59
60
61
62
        add_binary_op("Add", op::add{});
        add_binary_op("Div", op::div{});
        add_binary_op("Mul", op::mul{});
        add_binary_op("Sub", op::sub{});

Khalique's avatar
Khalique committed
63
64
65
        add_variadic_op("Sum", op::add{});
        add_variadic_op("Max", op::max{});
        add_variadic_op("Min", op::min{});
Paul's avatar
Paul committed
66

Khalique's avatar
Khalique committed
67
        add_mem_op("LRN", &onnx_parser::parse_lrn);
Khalique's avatar
Khalique committed
68
        add_mem_op("ImageScaler", &onnx_parser::parse_imagescaler);
69
        add_mem_op("LeakyRelu", &onnx_parser::parse_leaky_relu);
Khalique's avatar
Khalique committed
70
        add_mem_op("Elu", &onnx_parser::parse_elu);
Paul's avatar
Paul committed
71
72
        add_mem_op("Constant", &onnx_parser::parse_constant);
        add_mem_op("Conv", &onnx_parser::parse_conv);
Paul's avatar
Paul committed
73
74
        add_mem_op("MaxPool", &onnx_parser::parse_pooling);
        add_mem_op("AveragePool", &onnx_parser::parse_pooling);
75
76
        add_mem_op("GlobalMaxPool", &onnx_parser::parse_pooling);
        add_mem_op("GlobalAveragePool", &onnx_parser::parse_pooling);
Paul's avatar
Paul committed
77
        add_mem_op("Reshape", &onnx_parser::parse_reshape);
Paul's avatar
Paul committed
78
79
        add_mem_op("Flatten", &onnx_parser::parse_flatten);
        add_mem_op("Gemm", &onnx_parser::parse_gemm);
80
        add_mem_op("BatchNormalization", &onnx_parser::parse_batchnorm);
Paul's avatar
Paul committed
81
        add_mem_op("Softmax", &onnx_parser::parse_softmax);
82
83
84
        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
85
        add_mem_op("Concat", &onnx_parser::parse_concat);
86
87
88
        add_mem_op("Gather", &onnx_parser::parse_gather);
        add_mem_op("Shape", &onnx_parser::parse_shape);
        add_mem_op("ConstantFill", &onnx_parser::parse_constant_fill);
Khalique's avatar
Khalique committed
89
        add_mem_op("Transpose", &onnx_parser::parse_transpose);
Shucai Xiao's avatar
Shucai Xiao committed
90
        add_mem_op("RNN", &onnx_parser::parse_rnn);
91
        add_mem_op("GRU", &onnx_parser::parse_gru);
Shucai Xiao's avatar
Shucai Xiao committed
92
        add_mem_op("LSTM", &onnx_parser::parse_lstm);
Khalique's avatar
Khalique committed
93
        add_mem_op("Pad", &onnx_parser::parse_pad);
94
95
96
97
98
99
100

        // init the activation function map
        init_actv_func();
    }

    void init_actv_func()
    {
101
102
103
104
105
        map_actv_funcs.insert(std::make_pair("tanh", op::tanh{}));
        map_actv_funcs.insert(std::make_pair("relu", op::relu{}));
        map_actv_funcs.insert(std::make_pair("sigmoid", op::sigmoid{}));
        map_actv_funcs.insert(std::make_pair("leakyrelu", op::leaky_relu{}));
        map_actv_funcs.insert(std::make_pair("elu", op::elu{}));
Paul's avatar
Paul committed
106
107
108
109
    }

    template <class F>
    void add_op(std::string name, F f)
Paul's avatar
Paul committed
110
111
112
113
114
115
116
117
118
    {
        ops.emplace(name, [=](auto&&... xs) {
            return std::vector<instruction_ref>{f(std::forward<decltype(xs)>(xs)...)};
        });
    }

    // Multi output op
    template <class F>
    void add_multi_op(std::string name, F f)
Paul's avatar
Paul committed
119
120
121
122
123
124
125
    {
        ops.emplace(name, f);
    }

    template <class F>
    void add_mem_op(std::string name, F f)
    {
Paul's avatar
Paul committed
126
        add_op(name, [=](auto&&... xs) {
Paul's avatar
Paul committed
127
128
129
            return std::mem_fn(f)(*this, name, std::forward<decltype(xs)>(xs)...);
        });
    }
Khalique's avatar
Khalique committed
130

131
    template <class T>
Khalique's avatar
Khalique committed
132
    void add_binary_op(std::string name, T x)
133
    {
Paul's avatar
Paul committed
134
        add_op(name, [this, x](attribute_map attributes, std::vector<instruction_ref> args) {
Scott Thornton's avatar
Scott Thornton committed
135
            if(args.size() != 2)
Paul's avatar
Paul committed
136
                MIGRAPHX_THROW("binary operators should have 2 operands");
137
            if(contains(attributes, "broadcast") and contains(attributes, "axis"))
138
139
140
141
            {
                uint64_t broadcasted = parse_value(attributes.at("broadcast")).at<uint64_t>();
                if(broadcasted != 0)
                {
142
                    uint64_t axis = parse_value(attributes.at("axis")).at<uint64_t>();
143
144
145
146
                    auto l =
                        prog.add_instruction(op::broadcast{axis, args[0]->get_shape()}, args[1]);
                    return prog.add_instruction(x, args[0], l);
                }
147
                return prog.add_instruction(x, args);
148
            }
Paul's avatar
Paul committed
149
            else
150
            {
Khalique's avatar
Khalique committed
151
                return add_broadcastable_binary_op(args[0], args[1], x);
152
153
154
155
            }
        });
    }

Khalique's avatar
Khalique committed
156
157
158
159
160
    template <class T>
    instruction_ref add_broadcastable_binary_op(instruction_ref arg0, instruction_ref arg1, T x)
    {
        if(arg0->get_shape() != arg1->get_shape())
        {
Khalique's avatar
Khalique committed
161
162
163
164
165
166
167
168
169
170
171
172
173
            // Example:
            // s0 = (3,2,4,5) and s1 = (2,1,1)
            //
            // In this case we need to broadcast (:,1,1) portion of
            // s1 plus broadcast the 1st dimension of s1
            // giving output_lens = (3,2,4,5)
            //
            // Another example:
            // s0 = (3,2,1,5) and s1 = (2,7,5)
            // In this case we need to broadcast the (:,:,1:,:) axis
            // of s0 plus the 1st dimension of s1 giving
            // output_lens = (3,2,7,5)
            //
Khalique's avatar
Khalique committed
174
175
176
177
178
179
180
181
            // Get lengths for both arguments
            const std::vector<std::size_t>* s0 = &arg0->get_shape().lens();
            const std::vector<std::size_t>* s1 = &arg1->get_shape().lens();

            // Make sure s0 is the smaller size
            if(s0->size() > s1->size())
                std::swap(s0, s1);

Khalique's avatar
Khalique committed
182
            std::vector<std::size_t> output_lens(*s1);
Khalique's avatar
Khalique committed
183
184
            auto offset = s1->size() - s0->size();
            std::transform(s0->begin(),
Khalique's avatar
Khalique committed
185
186
187
188
                           s0->end(),
                           s1->begin() + offset,
                           output_lens.begin() + offset,
                           [](auto a, auto b) { return std::max(a, b); });
Khalique's avatar
Khalique committed
189
190
191
192
193
194
195
196
197

            auto l0 = prog.add_instruction(op::multibroadcast{output_lens}, arg0);
            auto l1 = prog.add_instruction(op::multibroadcast{output_lens}, arg1);
            return prog.add_instruction(x, l0, l1);
        }
        else
        {
            return prog.add_instruction(x, {arg0, arg1});
        }
198
199
    }

Paul's avatar
Paul committed
200
    template <class T>
Paul's avatar
Paul committed
201
202
    void add_generic_op(std::string name, T x)
    {
Paul's avatar
Paul committed
203
        add_op(name, [this, x](attribute_map, std::vector<instruction_ref> args) {
Paul's avatar
Paul committed
204
205
206
207
            return prog.add_instruction(x, args);
        });
    }

Khalique's avatar
Khalique committed
208
    template <class T>
Khalique's avatar
Khalique committed
209
    void add_variadic_op(std::string name, T x)
Khalique's avatar
Khalique committed
210
    {
Paul's avatar
Paul committed
211
        add_op(name, [this, x](attribute_map, std::vector<instruction_ref> args) {
Khalique's avatar
Khalique committed
212
            return std::accumulate(std::next(args.begin()),
Khalique's avatar
Khalique committed
213
214
215
216
217
                                   args.end(),
                                   args.front(),
                                   [this, x](instruction_ref a, instruction_ref b) {
                                       return add_broadcastable_binary_op(a, b, x);
                                   });
Khalique's avatar
Khalique committed
218
        });
Khalique's avatar
Khalique committed
219
220
    }

Paul's avatar
Paul committed
221
    instruction_ref
Paul's avatar
Paul committed
222
    parse_softmax(const std::string&, const attribute_map&, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
223
224
    {
        auto dims = args.front()->get_shape().lens();
Scott Thornton's avatar
Scott Thornton committed
225
226
        auto r =
            prog.add_instruction(op::reshape{{long(dims[0]), long(dims[1]), 1, 1}}, args.front());
227
228
        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
229
230
    }

Paul's avatar
Paul committed
231
    instruction_ref
Paul's avatar
Paul committed
232
    parse_conv(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
233
    {
234
        op::convolution op;
235
        auto l0 = args[0];
Paul's avatar
Paul committed
236
237
        if(contains(attributes, "pads"))
        {
Scott Thornton's avatar
Scott Thornton committed
238
            if(contains(attributes, "auto_pad"))
239
            {
Paul's avatar
Paul committed
240
                MIGRAPHX_THROW("auto_pad and padding cannot be specified simultaneously");
241
            }
242
243
            std::vector<std::int64_t> padding;
            copy(attributes["pads"].ints(), std::back_inserter(padding));
Scott Thornton's avatar
Scott Thornton committed
244
            if(padding.size() != 4)
245
            {
Paul's avatar
Paul committed
246
                MIGRAPHX_THROW("padding should have 4 values");
247
            }
Scott Thornton's avatar
Scott Thornton committed
248
            if(padding[0] != padding[2] || padding[1] != padding[3])
249
            {
250
251
                // insert zeros for pad op (args[0] has 4 dims)
                padding = {0, 0, padding[0], padding[1], 0, 0, padding[2], padding[3]};
Khalique's avatar
Khalique committed
252
                l0      = prog.add_instruction(op::pad{padding}, l0);
253
            }
254
255
256
257
            else
            {
                op.padding[0] = padding[0];
                op.padding[1] = padding[1];
258
            }
Paul's avatar
Paul committed
259
        }
Paul's avatar
Paul committed
260
261
262
263
264
265
266
267
        if(contains(attributes, "strides"))
        {
            copy(attributes["strides"].ints(), op.stride.begin());
        }
        if(contains(attributes, "dilations"))
        {
            copy(attributes["dilations"].ints(), op.dilation.begin());
        }
Scott Thornton's avatar
Scott Thornton committed
268
        if(contains(attributes, "auto_pad"))
269
270
        {
            auto s = attributes["auto_pad"].s();
Scott Thornton's avatar
Scott Thornton committed
271
            if(contains(attributes, "pads") and to_upper(s) != "NOTSET")
272
            {
Paul's avatar
Paul committed
273
                MIGRAPHX_THROW("auto_pad and padding cannot be specified simultaneously");
274
275
            }

wsttiger's avatar
fixes  
wsttiger committed
276
            if(s.find("SAME") != std::string::npos)
277
            {
278
                op.padding_mode = op::padding_mode_t::same;
279
280
            }
        }
Khalique's avatar
Khalique committed
281
282
283
284
        if(contains(attributes, "group"))
        {
            op.group = parse_value(attributes.at("group")).at<int>();
        }
Paul's avatar
Paul committed
285
286
287
288
        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
289
            auto l2       = prog.add_instruction(op::broadcast{axis, l1->get_shape()}, args[2]);
290
            return prog.add_instruction(op::add{}, l1, l2);
Paul's avatar
Paul committed
291
        }
292
        return prog.add_instruction(op, l0, args[1]);
Paul's avatar
Paul committed
293
    }
Paul's avatar
Paul committed
294

Paul's avatar
Paul committed
295
296
297
    instruction_ref parse_pooling(const std::string& name,
                                  attribute_map attributes,
                                  std::vector<instruction_ref> args)
Paul's avatar
Paul committed
298
    {
Khalique's avatar
Khalique committed
299
        op::pooling op{ends_with(name, "MaxPool") ? "max" : "average"};
300
        auto l0 = args[0];
Khalique's avatar
Khalique committed
301
        if(starts_with(name, "Global"))
302
        {
Khalique's avatar
Khalique committed
303
304
            auto lens  = args.front()->get_shape().lens();
            op.lengths = {lens[2], lens[3]};
305
        }
Paul's avatar
Paul committed
306
307
        if(contains(attributes, "pads"))
        {
308
309
            std::vector<std::int64_t> padding;
            copy(attributes["pads"].ints(), std::back_inserter(padding));
Scott Thornton's avatar
Scott Thornton committed
310
            if(padding.size() != 4)
311
            {
Paul's avatar
Paul committed
312
                MIGRAPHX_THROW("padding should have 4 values");
313
            }
Scott Thornton's avatar
Scott Thornton committed
314
            if(padding[0] != padding[2] || padding[1] != padding[3])
315
            {
316
317
                // insert zeros for pad op (args[0] has 4 dims)
                padding = {0, 0, padding[0], padding[1], 0, 0, padding[2], padding[3]};
Khalique's avatar
Khalique committed
318
                l0      = prog.add_instruction(op::pad{padding}, l0);
319
320
321
322
323
            }
            else
            {
                op.padding[0] = padding[0];
                op.padding[1] = padding[1];
324
            }
Paul's avatar
Paul committed
325
326
327
328
329
330
331
332
333
        }
        if(contains(attributes, "strides"))
        {
            copy(attributes["strides"].ints(), op.stride.begin());
        }
        if(contains(attributes, "kernel_shape"))
        {
            copy(attributes["kernel_shape"].ints(), op.lengths.begin());
        }
Scott Thornton's avatar
Scott Thornton committed
334
        if(contains(attributes, "auto_pad"))
335
336
        {
            auto s = attributes["auto_pad"].s();
337
            if(s.find("SAME_UPPER") == std::string::npos)
338
            {
339
                MIGRAPHX_THROW("auto_pad only supports SAME_UPPER for pooling");
340
            }
341
            op.padding_mode = op::padding_mode_t::same;
342
343
        }

344
        return prog.add_instruction(op, l0);
Paul's avatar
Paul committed
345
346
    }

Paul's avatar
Paul committed
347
    instruction_ref
Paul's avatar
Paul committed
348
    parse_reshape(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
349
    {
350
        op::reshape op;
Paul's avatar
Paul committed
351
352
353
354
355
356
357
        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
358
            literal s = args[1]->get_literal();
Paul's avatar
Paul committed
359
            s.visit([&](auto v) { copy(v, std::back_inserter(op.dims)); });
Paul's avatar
Paul committed
360
        }
Paul's avatar
Paul committed
361
362
363
        return prog.add_instruction(op, args[0]);
    }

Paul's avatar
Paul committed
364
    instruction_ref
Paul's avatar
Paul committed
365
    parse_flatten(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
366
    {
367
        uint64_t axis = 1;
Paul's avatar
Paul committed
368
369
370
371
        if(contains(attributes, "axis"))
        {
            axis = parse_value(attributes.at("axis")).at<int>();
        }
372
        return prog.add_instruction(op::flatten{axis}, args[0]);
Paul's avatar
Paul committed
373
374
    }

375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
    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
393
394
395
396
397
398
399
    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));
    }
400

401
402
403
    instruction_ref
    parse_gather(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
404
        int axis = 0;
405
406
407
408
        if(contains(attributes, "axis"))
        {
            axis = parse_value(attributes.at("axis")).at<int>();
        }
409
        op::gather op{axis};
410
411
412
        return prog.add_instruction(op, std::move(args));
    }

413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
    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
433
434
435
    instruction_ref parse_constant(const std::string&,
                                   attribute_map attributes,
                                   const std::vector<instruction_ref>&)
Paul's avatar
Paul committed
436
    {
Shucai Xiao's avatar
Shucai Xiao committed
437
        literal v     = parse_value(attributes.at("value"));
438
439
440
        auto dim_size = attributes.at("value").t().dims_size();
        // if dim_size is 0, it is a scalar
        if(dim_size == 0)
441
        {
442
            migraphx::shape scalar_shape{v.get_shape().type(), {1}, {0}};
443
444
445
            return prog.add_literal(migraphx::literal{scalar_shape, v.data()});
        }

Paul's avatar
Paul committed
446
447
        return prog.add_literal(v);
    }
Paul's avatar
Paul committed
448

Paul's avatar
Paul committed
449
    instruction_ref
Paul's avatar
Paul committed
450
    parse_gemm(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Paul's avatar
Paul committed
451
452
    {
        float alpha = 1.0f;
Khalique's avatar
Khalique committed
453
        float beta  = 1.0f;
Paul's avatar
Paul committed
454
455
456
457
458
459
460
461
        bool transa = false;
        bool transb = false;
        if(contains(attributes, "alpha"))
        {
            alpha = parse_value(attributes.at("alpha")).at<float>();
        }
        if(contains(attributes, "beta"))
        {
462
            beta = parse_value(attributes.at("beta")).at<float>();
Paul's avatar
Paul committed
463
464
465
466
467
468
469
470
471
        }
        if(contains(attributes, "transA"))
        {
            transa = parse_value(attributes.at("transA")).at<bool>();
        }
        if(contains(attributes, "transB"))
        {
            transb = parse_value(attributes.at("transB")).at<bool>();
        }
472
473
474
475
476
477

        std::vector<int64_t> perm(args[0]->get_shape().lens().size());
        std::iota(perm.begin(), perm.end(), int64_t{0});
        // swap the last two elements
        std::swap(*perm.rbegin(), *(perm.rbegin() + 1));

478
479
        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
480
481
        if(args.size() == 3)
        {
Khalique's avatar
Khalique committed
482
            if(beta != 0.f)
483
            {
Khalique's avatar
Khalique committed
484
                auto l3 = prog.add_instruction(op::dot{alpha}, l1, l2);
Khalique's avatar
Khalique committed
485
                auto l4 = args[2];
Khalique's avatar
Khalique committed
486
                if(l4->get_shape().scalar()) // ignore args[2] (no C value added to alpha*A*B)
Khalique's avatar
Khalique committed
487
                    return l3;
Khalique's avatar
Khalique committed
488
                if(beta != 1.f)
Khalique's avatar
Khalique committed
489
490
                {
                    auto beta_val = prog.add_literal(beta);
Khalique's avatar
Khalique committed
491
492
                    auto l5 = prog.add_instruction(op::scalar{args[2]->get_shape()}, beta_val);
                    l4      = prog.add_instruction(op::mul{}, args[2], l5);
Khalique's avatar
Khalique committed
493
494
                }
                return add_broadcastable_binary_op(l3, l4, op::add{});
495
            }
Paul's avatar
Paul committed
496
        }
497
        return prog.add_instruction(op::dot{alpha, beta}, l1, l2);
Paul's avatar
Paul committed
498
499
    }

500
    instruction_ref
Paul's avatar
Paul committed
501
    parse_batchnorm(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
502
    {
Scott Thornton's avatar
Scott Thornton committed
503
504
        float epsilon                                     = 1e-5f;
        float momentum                                    = 0.9f;
505
        op::batch_norm_inference::bn_infer_mode_t bn_mode = op::batch_norm_inference::spatial;
Scott Thornton's avatar
Scott Thornton committed
506
        bool is_test                                      = false;
507
508
509
510
511
512
        if(contains(attributes, "epsilon"))
        {
            epsilon = parse_value(attributes.at("epsilon")).at<float>();
        }
        if(contains(attributes, "momentum"))
        {
513
            momentum = parse_value(attributes.at("momentum")).at<float>();
514
515
516
        }
        if(contains(attributes, "is_test"))
        {
wsttiger's avatar
wsttiger committed
517
            is_test = parse_value(attributes.at("is_test")).at<uint64_t>() > 0;
518
519
520
        }
        if(contains(attributes, "spatial"))
        {
521
            bn_mode = (parse_value(attributes.at("spatial")).at<uint64_t>() > 0)
522
523
                          ? op::batch_norm_inference::spatial
                          : op::batch_norm_inference::per_activation;
524
        }
Paul's avatar
Paul committed
525
        (void)is_test;
Paul's avatar
Paul committed
526
        op::batch_norm_inference op{epsilon, momentum, bn_mode};
Paul's avatar
Paul committed
527
        return prog.add_instruction(op, std::move(args));
528
529
    }

530
531
532
533
    instruction_ref parse_leaky_relu(const std::string&,
                                     attribute_map attributes,
                                     std::vector<instruction_ref> args)
    {
Khalique's avatar
Khalique committed
534
        float alpha = 0.01; // default alpha val for leaky relu
535
536
537
538
539
540
541
542
        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
543
544
    instruction_ref
    parse_elu(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Khalique's avatar
Khalique committed
545
546
547
548
549
550
551
552
553
554
    {
        float alpha = 1.0; // default alpha val for elu
        if(contains(attributes, "alpha"))
        {
            alpha = parse_value(attributes.at("alpha")).at<float>();
        }
        op::elu op{alpha};
        return prog.add_instruction(op, args.front());
    }

Khalique's avatar
Khalique committed
555
556
    instruction_ref
    parse_lrn(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Khalique's avatar
Khalique committed
557
558
    {
        float alpha = 0.0001;
Khalique's avatar
Khalique committed
559
560
561
        float beta  = 0.75;
        float bias  = 1.0;
        int size    = 1;
Khalique's avatar
Khalique committed
562
563
564
565
566
567
568
569
570
571
572
573
        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
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
    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
591

Khalique's avatar
Khalique committed
592
593
        auto scale_val = prog.add_literal(scale);
        auto bias_vals = prog.add_literal(
Paul's avatar
Paul committed
594
            migraphx::literal{migraphx::shape{migraphx::shape::float_type, {bias.size()}}, bias});
Khalique's avatar
Khalique committed
595

Paul's avatar
Paul committed
596
597
        auto scale_tensor = prog.add_instruction(migraphx::op::scalar{input_shape}, scale_val);
        auto img_scaled   = prog.add_instruction(migraphx::op::mul{}, args.front(), scale_tensor);
Paul's avatar
Paul committed
598
        auto bias_bcast = prog.add_instruction(migraphx::op::broadcast{1, input_shape}, bias_vals);
Paul's avatar
Paul committed
599
        return prog.add_instruction(migraphx::op::add{}, img_scaled, bias_bcast);
Khalique's avatar
Khalique committed
600
    }
Khalique's avatar
Khalique committed
601

Khalique's avatar
Khalique committed
602
603
    instruction_ref
    parse_transpose(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
Khalique's avatar
Khalique committed
604
605
606
607
608
609
610
    {
        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());
        }
Paul's avatar
Paul committed
611
        return prog.add_instruction(migraphx::op::transpose{perm}, args.front());
Khalique's avatar
Khalique committed
612
613
    }

Khalique's avatar
Khalique committed
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
    instruction_ref
    parse_pad(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        std::vector<int64_t> pads{};
        float value = 0.0f;
        if(contains(attributes, "pads"))
        {
            auto&& pad_vals = attributes["pads"].ints();
            pads            = std::vector<int64_t>(pad_vals.begin(), pad_vals.end());
        }
        if(contains(attributes, "value"))
        {
            value = parse_value(attributes.at("value")).at<float>();
        }
        if(contains(attributes, "mode"))
        {
            auto mode = attributes.at("mode").s();
            if(mode != "constant")
                MIGRAPHX_THROW("migraphx currently only supports constant padding");
        }
        return prog.add_instruction(migraphx::op::pad{pads, value}, args.front());
    }
636
637
638
    // Use a literal instruction to replace the shape since, output of
    // shape operator are literals in migraphx
    instruction_ref
Shucai Xiao's avatar
Shucai Xiao committed
639
    parse_shape(const std::string&, const attribute_map&, std::vector<instruction_ref> args)
640
641
    {
        if(args.size() != 1)
642
            MIGRAPHX_THROW("Shape: operator should have 1 operand");
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
        std::vector<std::size_t> arg_shape = args[0]->get_shape().lens();
        std::vector<int64_t> vec_shape(arg_shape.size());
        migraphx::shape s(migraphx::shape::int64_type, {arg_shape.size()});
        std::transform(arg_shape.begin(), arg_shape.end(), vec_shape.begin(), [](auto i) {
            return int64_t(i);
        });
        return prog.add_literal(migraphx::literal{s, vec_shape});
    }

    // Use a literal instruction to replace the constantFill operator. In RNN, input shape
    // and value are fixed, so no need to do the actual computation for the constantFill
    // operator
    instruction_ref parse_constant_fill(const std::string&,
                                        attribute_map attributes,
                                        std::vector<instruction_ref> args)
    {
        int input_as_shape = 0;
        int dtype          = 1;
        float value        = 0.0f;

        if(contains(attributes, "dtype"))
        {
            dtype = parse_value(attributes.at("dtype")).at<int>();
        }
        migraphx::shape::type_t type = get_type(dtype);

        if(contains(attributes, "input_as_shape"))
        {
            input_as_shape = parse_value(attributes.at("input_as_shape")).at<int>();
        }

        if(contains(attributes, "value"))
        {
            value = parse_value(attributes.at("value")).at<float>();
        }

Shucai Xiao's avatar
Shucai Xiao committed
679
680
        if(contains(attributes, "extra_shape"))
        {
681
            MIGRAPHX_THROW("ConstantFill: cannot handle extra shape attribute");
682
683
        }

684
685
        if(input_as_shape == 1)
        {
Shucai Xiao's avatar
Shucai Xiao committed
686
            if(args.size() != 1)
687
            {
688
                MIGRAPHX_THROW("ConstantFill: need an input argument as output shape");
689
690
            }

Shucai Xiao's avatar
Shucai Xiao committed
691
692
            if(contains(attributes, "shape"))
            {
693
                MIGRAPHX_THROW("ConstantFill: cannot set the shape argument and pass in an input "
Shucai Xiao's avatar
Shucai Xiao committed
694
                               "at the same time");
695
696
            }

697
698
699
            migraphx::argument in = args[0]->eval();
            if(in.empty())
            {
700
                MIGRAPHX_THROW("ConstantFill: cannot handle dynamic shape as input");
701
            }
702

703
704
705
            std::vector<std::size_t> dims;
            in.visit([&](auto input) { dims.assign(input.begin(), input.end()); });
            migraphx::shape s(type, dims);
706
707
            std::vector<float> values(s.elements(), value);
            return prog.add_literal(migraphx::literal(s, values));
708
709
710
        }
        else if(input_as_shape == 0)
        {
Shucai Xiao's avatar
Shucai Xiao committed
711
712
            if(!contains(attributes, "shape"))
            {
713
                MIGRAPHX_THROW("ConstantFill: attribute output shape is needed");
714
715
716
            }

            literal ls = parse_value(attributes.at("shape"));
717
            std::vector<std::size_t> dims;
Shucai Xiao's avatar
Shucai Xiao committed
718
            ls.visit([&](auto s) { dims.assign(s.begin(), s.end()); });
719
            migraphx::shape s{type, dims};
720
721
            std::vector<float> values(s.elements(), value);
            return prog.add_literal(migraphx::literal(s, values));
722
723
724
        }
        else
        {
725
            MIGRAPHX_THROW("ConstantFill: wrong value of attribute input_as_shape");
726
727
728
        }
    }

Shucai Xiao's avatar
Shucai Xiao committed
729
    std::vector<instruction_ref>
Shucai Xiao's avatar
Shucai Xiao committed
730
731
732
    parse_rnn(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        migraphx::shape input_shape = args[0]->get_shape();
733
        std::size_t hidden_size     = args[1]->get_shape().lens()[1];
Shucai Xiao's avatar
Shucai Xiao committed
734
735
736

        if(contains(attributes, "hidden_size"))
        {
Shucai Xiao's avatar
Shucai Xiao committed
737
            std::size_t hidden_size_att = parse_value(attributes.at("hidden_size")).at<int>();
Shucai Xiao's avatar
Shucai Xiao committed
738
            if(hidden_size != hidden_size_att)
Shucai Xiao's avatar
Shucai Xiao committed
739
740
741
            {
                MIGRAPHX_THROW("RNN: hidden size mismatch in input and attribute");
            }
Shucai Xiao's avatar
Shucai Xiao committed
742
743
744
745
746
747
748
749
750
        }

        // Handling of direction to be added later
        std::string direction{"forward"};
        if(contains(attributes, "direction"))
        {
            direction = attributes.at("direction").s();
        }

751
        op::rnn_direction dirct = op::rnn_direction::forward;
Shucai Xiao's avatar
Shucai Xiao committed
752
753
        if(direction == "bidirectional")
        {
754
            dirct = op::rnn_direction::bidirectional;
Shucai Xiao's avatar
Shucai Xiao committed
755
756
757
        }
        else if(direction == "reverse")
        {
758
            dirct = op::rnn_direction::reverse;
Shucai Xiao's avatar
Shucai Xiao committed
759
760
        }

761
762
763
764
765
        std::vector<std::string> vec_names{"tanh"};
        if(contains(attributes, "activations"))
        {
            auto names = attributes.at("activations").strings();
            vec_names.clear();
766
            vec_names.resize(names.size());
767
            std::copy(names.begin(), names.end(), vec_names.begin());
768
769
        }

770
771
772
        auto name_it = std::find_if(vec_names.begin(), vec_names.end(), [&](auto& name) {
            return (map_actv_funcs.count(name) == 0);
        });
Shucai Xiao's avatar
Shucai Xiao committed
773
        if(name_it != vec_names.end())
774
775
776
        {
            MIGRAPHX_THROW("RNN: activation function " + std::string(*name_it) + " not supported");
        }
777

Shucai Xiao's avatar
Shucai Xiao committed
778
        // bidirectional case should have two activation functions.
Shucai Xiao's avatar
Shucai Xiao committed
779
        // one is for forward, and the other is for reverse.
Shucai Xiao's avatar
Shucai Xiao committed
780
        // if only one actv function is provided, we use it in both
781
        // forward and reverse direction
782
        if(dirct == op::rnn_direction::bidirectional)
783
        {
Shucai Xiao's avatar
Shucai Xiao committed
784
            if(vec_names.size() == 1)
785
786
787
788
789
            {
                vec_names.push_back(vec_names.at(0));
            }
        }

Shucai Xiao's avatar
Shucai Xiao committed
790
791
792
        std::vector<operation> vec_actv_funcs(vec_names.size());
        std::transform(vec_names.begin(), vec_names.end(), vec_actv_funcs.begin(), [&](auto& fn) {
            return map_actv_funcs[fn];
793
        });
Shucai Xiao's avatar
Shucai Xiao committed
794

Shucai Xiao's avatar
Shucai Xiao committed
795
796
797
798
799
800
801
        // To be added later
        float clip = 0.0;
        if(contains(attributes, "clip"))
        {
            clip = parse_value(attributes.at("clip")).at<float>();
        }

802
803
        // if the number of arguments is less than 6, append
        // undefined operator to have 6 arguments
Shucai Xiao's avatar
Shucai Xiao committed
804
        if(args.size() < 6)
805
806
807
808
809
        {
            auto ins = prog.add_instruction(op::undefined{});
            args.insert(args.end(), (6 - args.size()), ins);
        }

Shucai Xiao's avatar
Shucai Xiao committed
810
811
        // first output for the concatenation of hidden states
        auto hidden_states = prog.add_instruction(op::rnn{hidden_size, vec_actv_funcs, dirct, clip},
Shucai Xiao's avatar
Shucai Xiao committed
812
                                                  std::move(args));
Shucai Xiao's avatar
Shucai Xiao committed
813

814
        // second output for the last hidden state
Shucai Xiao's avatar
Shucai Xiao committed
815
        auto last_output = prog.add_instruction(op::rnn_last_output{}, hidden_states);
Shucai Xiao's avatar
Shucai Xiao committed
816

Shucai Xiao's avatar
Shucai Xiao committed
817
        return {hidden_states, last_output};
Shucai Xiao's avatar
Shucai Xiao committed
818
819
    }

820
    std::vector<instruction_ref>
821
822
823
824
825
826
827
    parse_gru(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        migraphx::shape input_shape = args[0]->get_shape();
        std::size_t hidden_size     = args[2]->get_shape().lens()[2];

        if(contains(attributes, "hidden_size"))
        {
Shucai Xiao's avatar
Shucai Xiao committed
828
            std::size_t hidden_size_att = parse_value(attributes.at("hidden_size")).at<int>();
Shucai Xiao's avatar
Shucai Xiao committed
829
            if(hidden_size != hidden_size_att)
Shucai Xiao's avatar
Shucai Xiao committed
830
831
832
            {
                MIGRAPHX_THROW("GRU: hidden size mismatch in input and attribute");
            }
833
834
835
836
837
838
839
840
841
        }

        // Handling of direction to be added later
        std::string direction{"forward"};
        if(contains(attributes, "direction"))
        {
            direction = attributes.at("direction").s();
        }

842
        op::rnn_direction dirct = op::rnn_direction::forward;
843
844
        if(direction == "bidirectional")
        {
845
            dirct = op::rnn_direction::bidirectional;
846
847
848
        }
        else if(direction == "reverse")
        {
849
            dirct = op::rnn_direction::reverse;
850
851
        }

852
        std::vector<std::string> vec_names = {"sigmoid", "tanh"};
853
854
        if(contains(attributes, "activations"))
        {
855
            auto names = attributes.at("activations").strings();
856
            vec_names.clear();
Shucai Xiao's avatar
Shucai Xiao committed
857
            vec_names.resize(names.size());
858
            std::copy(names.begin(), names.end(), vec_names.begin());
859
860
        }

861
        // need 4 activation functions
862
        if(dirct == op::rnn_direction::bidirectional)
863
        {
Shucai Xiao's avatar
Shucai Xiao committed
864
            // 4 activation functions are used in the bidirectional
865
            // scenario. No spec is provided in onnx::operator. we
Shucai Xiao's avatar
Shucai Xiao committed
866
867
            // use the algorithm that: if 1 actv function is provided,
            // repeat 1 four times. If 2 actv functins are provided,
868
869
            // assume forward and reverse use the same pair of actv
            // functions. For the case of 3 actv functions provided,
Shucai Xiao's avatar
Shucai Xiao committed
870
871
872
            // assume the 3rd one is repeated once and used by the
            // reverse direction.
            // This may need change later
873
            if(vec_names.size() == 1)
874
            {
875
                vec_names.insert(vec_names.end(), 3, vec_names.at(0));
876
            }
877
            else if(vec_names.size() == 2)
878
            {
879
880
881
                // repeat the activation functions
                vec_names.push_back(vec_names.at(0));
                vec_names.push_back(vec_names.at(1));
882
            }
883
            else if(vec_names.size() == 3)
884
            {
885
                vec_names.push_back(vec_names.at(2));
886
887
            }
        }
Shucai Xiao's avatar
Shucai Xiao committed
888
        else
889
        {
890
            if(vec_names.size() == 1)
891
            {
892
                vec_names.push_back(vec_names.at(0));
893
894
895
            }
        }

896
897
898
        auto name_it = std::find_if(vec_names.begin(), vec_names.end(), [&](auto& name) {
            return (map_actv_funcs.count(name) == 0);
        });
Shucai Xiao's avatar
Shucai Xiao committed
899
        if(name_it != vec_names.end())
900
901
902
        {
            MIGRAPHX_THROW("GRU: activation function " + std::string(*name_it) + " not supported");
        }
903

Shucai Xiao's avatar
Shucai Xiao committed
904
905
906
        std::vector<operation> vec_actv_funcs(vec_names.size());
        std::transform(vec_names.begin(), vec_names.end(), vec_actv_funcs.begin(), [&](auto& name) {
            return map_actv_funcs[name];
Shucai Xiao's avatar
Shucai Xiao committed
907
        });
908
909
910
911
912
913
914
915

        float clip = 0.0;
        if(contains(attributes, "clip"))
        {
            clip = parse_value(attributes.at("clip")).at<float>();
        }

        int linear_before_reset = 0;
Shucai Xiao's avatar
Shucai Xiao committed
916
        if(contains(attributes, "linear_before_reset"))
917
918
919
920
        {
            linear_before_reset = parse_value(attributes.at("linear_before_reset")).at<int>();
        }

Shucai Xiao's avatar
Shucai Xiao committed
921
        // append undefined opeator to make 6 arguments
Shucai Xiao's avatar
Shucai Xiao committed
922
        if(args.size() < 6)
Shucai Xiao's avatar
Shucai Xiao committed
923
924
925
926
927
        {
            auto ins = prog.add_instruction(op::undefined{});
            args.insert(args.end(), 6 - args.size(), ins);
        }

928
929
        // first output for concatenation of hidden states
        auto hidden_states = prog.add_instruction(
Shucai Xiao's avatar
Shucai Xiao committed
930
            op::gru{hidden_size, vec_actv_funcs, dirct, clip, linear_before_reset},
Shucai Xiao's avatar
Shucai Xiao committed
931
            std::move(args));
932
933

        // second output for last gru output
934
        auto last_output = prog.add_instruction(op::rnn_last_output{}, hidden_states);
935

Shucai Xiao's avatar
Shucai Xiao committed
936
        return {hidden_states, last_output};
937
938
    }

Shucai Xiao's avatar
Shucai Xiao committed
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
    std::vector<instruction_ref>
    parse_lstm(const std::string&, attribute_map attributes, std::vector<instruction_ref> args)
    {
        migraphx::shape input_shape = args[0]->get_shape();
        std::size_t hidden_size     = args[2]->get_shape().lens()[2];

        if(contains(attributes, "hidden_size"))
        {
            std::size_t hidden_size_att = parse_value(attributes.at("hidden_size")).at<int>();
            if(hidden_size != hidden_size_att)
            {
                MIGRAPHX_THROW("LSTM: hidden size mismatch in input and attribute");
            }
        }

        // Handling of direction to be added later
        std::string direction{"forward"};
        if(contains(attributes, "direction"))
        {
            direction = attributes.at("direction").s();
        }

Shucai Xiao's avatar
Shucai Xiao committed
961
        op::rnn_direction dirct = op::rnn_direction::forward;
Shucai Xiao's avatar
Shucai Xiao committed
962
963
        if(direction == "bidirectional")
        {
Shucai Xiao's avatar
Shucai Xiao committed
964
            dirct = op::rnn_direction::bidirectional;
Shucai Xiao's avatar
Shucai Xiao committed
965
966
967
        }
        else if(direction == "reverse")
        {
Shucai Xiao's avatar
Shucai Xiao committed
968
            dirct = op::rnn_direction::reverse;
Shucai Xiao's avatar
Shucai Xiao committed
969
        }
Shucai Xiao's avatar
Shucai Xiao committed
970
        else if(direction == "forward")
Shucai Xiao's avatar
Shucai Xiao committed
971
        {
Shucai Xiao's avatar
Shucai Xiao committed
972
            dirct = op::rnn_direction::forward;
Shucai Xiao's avatar
Shucai Xiao committed
973
974
975
976
977
978
979
980
981
982
983
984
        }
        else
        {
            MIGRAPHX_THROW("LSTM: incorrect direction attribute");
        }

        std::vector<std::string> vec_names = {"sigmoid", "tanh", "tanh"};
        if(contains(attributes, "activations"))
        {
            auto names = attributes.at("activations").strings();
            vec_names.clear();
            vec_names.resize(names.size());
985
            std::copy(names.begin(), names.end(), vec_names.begin());
Shucai Xiao's avatar
Shucai Xiao committed
986
987
988
        }

        // need 6 activation functions for bidirectional directions
Shucai Xiao's avatar
Shucai Xiao committed
989
        if(dirct == op::rnn_direction::bidirectional)
Shucai Xiao's avatar
Shucai Xiao committed
990
991
992
993
994
995
        {
            // 6 activation functions are used in the bidirectional
            // scenario. No spec is provided in onnx::operator. we
            // use the algorithm that: if 1 actv function is provided,
            // repeat 1st six times. If 2 actv functins are provided,
            // repeat 2nd once, then repeat all three once
Shucai Xiao's avatar
Shucai Xiao committed
996
            // if 3 actv funcs are provide, repeat all three once.
Shucai Xiao's avatar
Shucai Xiao committed
997
998
999
1000
            // the same algorithm is used for 4, 5, and 6 actv funcions
            // provided. This may need change later
            switch(vec_names.size())
            {
1001
            case 1:
Shucai Xiao's avatar
Shucai Xiao committed
1002
1003
1004
1005
1006
1007
                vec_names = {vec_names.at(0),
                             vec_names.at(0),
                             vec_names.at(0),
                             vec_names.at(0),
                             vec_names.at(0),
                             vec_names.at(0)};
1008
                break;
Shucai Xiao's avatar
Shucai Xiao committed
1009
1010
1011

            case 2:
                // repeat the 2nd actv func once, then repeat all three another time
Shucai Xiao's avatar
Shucai Xiao committed
1012
1013
1014
1015
1016
1017
                vec_names = {vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(1),
                             vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(1)};
Shucai Xiao's avatar
Shucai Xiao committed
1018
1019
1020
1021
                break;

            case 3:
                // repeat all three actv funcs once
Shucai Xiao's avatar
Shucai Xiao committed
1022
1023
1024
1025
1026
1027
                vec_names = {vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(2),
                             vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(2)};
Shucai Xiao's avatar
Shucai Xiao committed
1028
1029
                break;

Shucai Xiao's avatar
Shucai Xiao committed
1030
1031
1032
1033
1034
1035
1036
            case 4:
                vec_names = {vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(2),
                             vec_names.at(3),
                             vec_names.at(3),
                             vec_names.at(3)};
1037
                break;
Shucai Xiao's avatar
Shucai Xiao committed
1038

Shucai Xiao's avatar
Shucai Xiao committed
1039
1040
1041
1042
1043
1044
1045
            case 5:
                vec_names = {vec_names.at(0),
                             vec_names.at(1),
                             vec_names.at(2),
                             vec_names.at(3),
                             vec_names.at(4),
                             vec_names.at(4)};
1046
                break;
Shucai Xiao's avatar
Shucai Xiao committed
1047

Shucai Xiao's avatar
Shucai Xiao committed
1048
            default: break;
Shucai Xiao's avatar
Shucai Xiao committed
1049
1050
1051
1052
1053
1054
            }
        }
        else
        {
            switch(vec_names.size())
            {
Shucai Xiao's avatar
Shucai Xiao committed
1055
            case 1: vec_names = {vec_names.at(0), vec_names.at(0), vec_names.at(0)}; break;
Shucai Xiao's avatar
Shucai Xiao committed
1056
1057
1058

            case 2:
                // repeat the 2nd actv func once, so we have 3 actv funcs
Shucai Xiao's avatar
Shucai Xiao committed
1059
                vec_names = {vec_names.at(0), vec_names.at(1), vec_names.at(1)};
Shucai Xiao's avatar
Shucai Xiao committed
1060
1061
                break;

Shucai Xiao's avatar
Shucai Xiao committed
1062
            default: break;
Shucai Xiao's avatar
Shucai Xiao committed
1063
1064
1065
            }
        }

1066
1067
1068
        auto name_it = std::find_if(vec_names.begin(), vec_names.end(), [&](auto& name) {
            return (map_actv_funcs.count(name) == 0);
        });
Shucai Xiao's avatar
Shucai Xiao committed
1069
        if(name_it != vec_names.end())
1070
1071
1072
        {
            MIGRAPHX_THROW("LSTM: activation function " + std::string(*name_it) + " not supported");
        }
Shucai Xiao's avatar
Shucai Xiao committed
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094

        std::vector<operation> vec_actv_funcs(vec_names.size());
        std::transform(vec_names.begin(), vec_names.end(), vec_actv_funcs.begin(), [&](auto& name) {
            return map_actv_funcs[name];
        });

        float clip = 0.0;
        if(contains(attributes, "clip"))
        {
            clip = parse_value(attributes.at("clip")).at<float>();
        }

        int input_forget = 0;
        if(contains(attributes, "input_forget"))
        {
            input_forget = parse_value(attributes.at("input_forget")).at<int>();
        }

        // append undefined opeator to make 6 arguments
        if(args.size() < 8)
        {
            auto ins = prog.add_instruction(op::undefined{});
Shucai Xiao's avatar
Shucai Xiao committed
1095
            args.insert(args.end(), 8 - args.size(), ins);
Shucai Xiao's avatar
Shucai Xiao committed
1096
1097
1098
1099
        }

        // first output for concatenation of hidden states
        auto hidden_states = prog.add_instruction(
Shucai Xiao's avatar
Shucai Xiao committed
1100
            op::lstm{hidden_size, vec_actv_funcs, dirct, clip, input_forget}, std::move(args));
Shucai Xiao's avatar
Shucai Xiao committed
1101
1102

        // second output for last lstm output
Shucai Xiao's avatar
Shucai Xiao committed
1103
        auto last_output = prog.add_instruction(op::rnn_last_output{}, hidden_states);
Shucai Xiao's avatar
Shucai Xiao committed
1104
1105
1106
1107
1108
1109
1110

        // third output for last cell output
        auto last_cell_output = prog.add_instruction(op::lstm_last_cell_output{}, hidden_states);

        return {hidden_states, last_output, last_cell_output};
    }

Paul's avatar
Paul committed
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
    void parse_from(std::istream& is)
    {
        onnx::ModelProto model;
        if(model.ParseFromIstream(&is))
        {
            if(model.has_graph())
            {
                this->parse_graph(model.graph());
            }
        }
        else
        {
Paul's avatar
Paul committed
1123
            MIGRAPHX_THROW("Failed reading onnx file.");
Paul's avatar
Paul committed
1124
1125
1126
1127
1128
1129
        }
    }

    void parse_graph(const onnx::GraphProto& graph)
    {
        nodes = get_nodes(graph);
1130
1131
1132
1133
1134
        std::unordered_map<std::string, onnx::TensorProto> initializer_data;
        for(auto&& f : graph.initializer())
        {
            initializer_data[f.name()] = f;
        }
Paul's avatar
Paul committed
1135
1136
1137
        for(auto&& input : graph.input())
        {
            const std::string& name = input.name();
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
            // 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
1150
1151
1152
        }
        for(auto&& p : nodes)
        {
Paul's avatar
Paul committed
1153
            this->parse_node(p.first);
Paul's avatar
Paul committed
1154
1155
1156
        }
    }

Shucai Xiao's avatar
Shucai Xiao committed
1157
    void parse_undefined(const std::string& name)
1158
    {
Shucai Xiao's avatar
Shucai Xiao committed
1159
        auto ins           = prog.add_instruction(op::undefined{});
1160
1161
1162
        instructions[name] = ins;
    }

Paul's avatar
Paul committed
1163
    void parse_node(const std::string& name)
Paul's avatar
Paul committed
1164
    {
Paul's avatar
Paul committed
1165
        if(name.empty())
Paul's avatar
Paul committed
1166
            MIGRAPHX_THROW("Onnx node must have a name");
Paul's avatar
Paul committed
1167
1168
1169
1170
1171
1172
1173
1174
        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)
                {
Paul's avatar
Paul committed
1175
1176
                    assert(name != input);
                    this->parse_node(input);
Paul's avatar
Paul committed
1177
                }
Shucai Xiao's avatar
Shucai Xiao committed
1178
                else if(input.empty())
Paul's avatar
Paul committed
1179
                {
1180
                    this->parse_undefined(input);
Paul's avatar
Paul committed
1181
                }
1182
                args.push_back(instructions.at(input));
Paul's avatar
Paul committed
1183
            }
Paul's avatar
Paul committed
1184
            std::vector<instruction_ref> result;
Paul's avatar
Paul committed
1185
1186
            if(ops.count(node.op_type()) == 0)
            {
Paul's avatar
Paul committed
1187
                result.push_back(prog.add_instruction(unknown{node.op_type()}, args));
Paul's avatar
Paul committed
1188
1189
1190
            }
            else
            {
Paul's avatar
Paul committed
1191
                result = ops[node.op_type()](get_attributes(node), args);
Paul's avatar
Paul committed
1192
            }
Paul's avatar
Paul committed
1193
            // Even no output nodes produce output in migraphx
Paul's avatar
Paul committed
1194
            if(node.output().empty() and result.size() == 1)
Paul's avatar
Paul committed
1195
1196
            {
                instructions[name] = result.front();
Paul's avatar
Paul committed
1197
1198
1199
            }
            else
            {
Paul's avatar
Paul committed
1200
1201
1202
1203
1204
1205
                assert(node.output().size() >= result.size());
                std::transform(result.begin(),
                               result.end(),
                               node.output().begin(),
                               std::inserter(instructions, instructions.end()),
                               [](auto&& x, auto&& y) { return std::make_pair(y, x); });
Paul's avatar
Paul committed
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
            }
        }
    }

    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;
    }

    static node_map get_nodes(const onnx::GraphProto& graph)
    {
        std::unordered_map<std::string, onnx::NodeProto> result;
Paul's avatar
Paul committed
1223
        std::size_t n = 0;
Paul's avatar
Paul committed
1224
1225
        for(auto&& node : graph.node())
        {
Paul's avatar
Paul committed
1226
            if(node.output().empty())
Paul's avatar
Paul committed
1227
            {
Paul's avatar
Paul committed
1228
                if(node.name().empty())
Paul's avatar
Paul committed
1229
1230
1231
1232
1233
1234
1235
1236
1237
                {
                    result["migraphx_unamed_node_" + std::to_string(n)] = node;
                    n++;
                }
                else
                {
                    result[node.name()] = node;
                }
            }
Paul's avatar
Paul committed
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
            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
1263
        case onnx::AttributeProto::FLOATS: return from_repeated(shape::float_type, attr.floats());
Paul's avatar
Paul committed
1264
1265
1266
1267
1268
        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
1269
        MIGRAPHX_THROW("Invalid attribute type");
Paul's avatar
Paul committed
1270
1271
1272
1273
1274
    }

    static literal parse_tensor(const onnx::TensorProto& t)
    {
        std::vector<std::size_t> dims(t.dims().begin(), t.dims().end());
Khalique's avatar
Khalique committed
1275
        // in case of scalar constants in onnx file, use dims=1 to fill initializer data
1276
        if(dims.empty())
Khalique's avatar
Khalique committed
1277
1278
1279
        {
            dims = {1};
        }
1280
1281
        if(t.has_raw_data())
        {
wsttiger's avatar
wsttiger committed
1282
            const std::string& s = t.raw_data();
Scott Thornton's avatar
Scott Thornton committed
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
            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
1295
            case onnx::TensorProto::FLOAT16: return literal{{shape::half_type, dims}, s.data()};
Scott Thornton's avatar
Scott Thornton committed
1296
1297
1298
1299
1300
1301
            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("");
            }
Paul's avatar
Paul committed
1302
            MIGRAPHX_THROW("Invalid tensor type");
1303
        }
Paul's avatar
Paul committed
1304
1305
1306
1307
        switch(t.data_type())
        {
        case onnx::TensorProto::UNDEFINED: throw std::runtime_error("");
        case onnx::TensorProto::FLOAT:
Paul's avatar
Paul committed
1308
            return literal{{shape::float_type, dims}, t.float_data().begin(), t.float_data().end()};
Paul's avatar
Paul committed
1309
1310
        case onnx::TensorProto::UINT8: throw std::runtime_error("");
        case onnx::TensorProto::INT8:
Paul's avatar
Paul committed
1311
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
1312
        case onnx::TensorProto::UINT16:
Paul's avatar
Paul committed
1313
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
1314
        case onnx::TensorProto::INT16:
Paul's avatar
Paul committed
1315
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
1316
        case onnx::TensorProto::INT32:
Paul's avatar
Paul committed
1317
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
1318
        case onnx::TensorProto::INT64:
Paul's avatar
Paul committed
1319
            return literal{{shape::int64_type, dims}, t.int64_data().begin(), t.int64_data().end()};
Paul's avatar
Paul committed
1320
1321
        case onnx::TensorProto::STRING: throw std::runtime_error("");
        case onnx::TensorProto::BOOL:
Paul's avatar
Paul committed
1322
            return literal{{shape::int32_type, dims}, t.int32_data().begin(), t.int32_data().end()};
Paul's avatar
Paul committed
1323
        case onnx::TensorProto::FLOAT16:
Khalique's avatar
Khalique committed
1324
        {
Khalique's avatar
Khalique committed
1325
            std::vector<uint16_t> data_uint16(t.int32_data().begin(), t.int32_data().end());
1326
            std::vector<half> data_half;
Khalique's avatar
Khalique committed
1327
1328
1329
            std::transform(data_uint16.begin(),
                           data_uint16.end(),
                           std::back_inserter(data_half),
1330
                           [](uint16_t raw_val) { return *reinterpret_cast<half*>(&raw_val); });
1331
            return literal{{shape::half_type, dims}, data_half.begin(), data_half.end()};
Khalique's avatar
Khalique committed
1332
        }
Paul's avatar
Paul committed
1333
1334
1335
1336
1337
1338
1339
1340
        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
1341
        MIGRAPHX_THROW("Invalid tensor type");
Paul's avatar
Paul committed
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
    }

    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
1363
        case onnx::TensorProto::FLOAT16: shape_type = shape::half_type; break;
Paul's avatar
Paul committed
1364
1365
1366
1367
1368
1369
1370
1371
1372
        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
1373
        auto&& tensor_dims = t.tensor_type().shape().dim();
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
        std::transform(tensor_dims.begin(),
                       tensor_dims.end(),
                       std::back_inserter(dims),
                       [](auto&& d) -> std::size_t {
                           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
1385
1386
        return {shape_type, dims};
    }
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408

    shape::type_t get_type(int dtype)
    {
        switch(dtype)
        {
        case 1: return shape::float_type;
        case 2: return shape::uint8_type;
        case 3: return shape::int8_type;
        case 4: return shape::uint16_type;
        case 5: return shape::int16_type;
        case 6: return shape::int32_type;
        case 7: return shape::int64_type;
        case 10: return shape::half_type;
        case 11: return shape::double_type;
        case 12: return shape::uint32_type;
        case 13: return shape::uint64_type;
        default:
        {
            MIGRAPHX_THROW("Prototensor data type " + std::to_string(dtype) + " not supported");
        }
        }
    }
Paul's avatar
Paul committed
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
};

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);
}

Paul's avatar
Paul committed
1432
} // namespace MIGRAPHX_INLINE_NS
Paul's avatar
Paul committed
1433
} // namespace migraphx