graph_pattern_detector.cc 93.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14
// Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

15
#include <algorithm>
Q
Qiao Longfei 已提交
16
#include <array>
17
#include <memory>
18
#include <string>
19 20
#include <unordered_map>
#include <unordered_set>
21 22 23
#include <vector>

#include "paddle/fluid/framework/ir/graph_helper.h"
24
#include "paddle/fluid/framework/ir/graph_pattern_detector.h"
25
#include "paddle/fluid/framework/ir/graph_traits.h"
26
#include "paddle/fluid/framework/ir/graph_viz_pass.h"
C
chengduo 已提交
27
#include "paddle/fluid/framework/operator.h"
28
#include "paddle/fluid/platform/enforce.h"
Y
Yan Chunwei 已提交
29
#include "paddle/fluid/string/pretty_log.h"
Y
Yan Chunwei 已提交
30
#include "paddle/fluid/string/printf.h"
31

32 33 34 35
namespace paddle {
namespace framework {
namespace ir {

Y
Yan Chunwei 已提交
36 37 38
using string::PrettyLog;
using string::Style;

39 40
size_t PDPattern::id_ = 0UL;

C
chengduo 已提交
41
PDNode *PDPattern::NewNode(const std::string &name) {
Y
Yan Chunwei 已提交
42
  if (!name.empty()) {
43 44 45 46
    PADDLE_ENFORCE_EQ(
        node_map_.count(name), 0UL,
        platform::errors::PreconditionNotMet(
            "PDNode's name should be unique, get duplicate [%s]", name));
Y
Yan Chunwei 已提交
47 48 49
  }

  nodes_.emplace_back(new PDNode(this, name));
C
chengduo 已提交
50
  auto *cur = nodes_.back().get();
Y
Yan Chunwei 已提交
51 52 53 54
  node_map_[name] = cur;
  return cur;
}

C
chengduo 已提交
55
PDNode *PDPattern::NewNode(PDNode::teller_t &&teller, const std::string &name) {
56
  if (!name.empty()) {
57 58 59 60
    PADDLE_ENFORCE_EQ(
        node_map_.count(name), 0UL,
        platform::errors::PreconditionNotMet(
            "PDNode's name should be unique, get duplicate [%s]", name));
61 62
  }

63
  nodes_.emplace_back(new PDNode(std::move(teller), this, name));
C
chengduo 已提交
64
  auto *cur = nodes_.back().get();
65
  node_map_[name] = cur;
66 67 68
  return cur;
}

C
chengduo 已提交
69
PDNode *PDPattern::RetrieveNode(const std::string &id) const {
70 71 72 73 74 75 76 77
  auto it = node_map_.find(id);
  if (it == node_map_.end()) {
    return nullptr;
  }

  return it->second;
}

C
chengduo 已提交
78
void PDPattern::AddEdge(PDNode *a, PDNode *b) {
79 80 81 82
  PADDLE_ENFORCE_NOT_NULL(
      a, platform::errors::NotFound("PDNode %s is not found.", a->name()));
  PADDLE_ENFORCE_NOT_NULL(
      b, platform::errors::NotFound("PDNode %s is not found.", b->name()));
83 84
  PADDLE_ENFORCE_NE(a, b, platform::errors::PermissionDenied(
                              "Cannot connect the same node in the graph."));
85 86 87
  edges_.emplace_back(a, b);
}

C
chengduo 已提交
88
void GraphPatternDetector::operator()(Graph *graph,
89
                                      GraphPatternDetector::handle_t handler) {
90 91 92 93
  if (!MarkPDNodesInGraph(*graph)) {
    return;
  }

94 95
  auto subgraphs = DetectPatterns();
  UniquePatterns(&subgraphs);
Z
Zhang Ting 已提交
96
  SortSubgraphs(&subgraphs);
97
  RemoveOverlappedMatch(&subgraphs);
Y
Yan Chunwei 已提交
98
  ValidateByNodeRole(&subgraphs);
99

Y
Yan Chunwei 已提交
100
  if (subgraphs.empty()) return;
101
  LOG(INFO) << "---  detected " << subgraphs.size() << " subgraphs";
102
  int id = 0;
C
chengduo 已提交
103
  for (auto &g : subgraphs) {
M
minqiyang 已提交
104
    VLOG(3) << "optimizing #" << id++ << " subgraph";
105 106 107 108
    handler(g, graph);
  }
}

C
chengduo 已提交
109
bool GraphPatternDetector::MarkPDNodesInGraph(const ir::Graph &graph) {
M
minqiyang 已提交
110
  VLOG(3) << "mark pdnodes in graph";
111 112
  if (graph.Nodes().empty()) return false;

C
chengduo 已提交
113 114
  for (auto &node : GraphTraits::DFS(graph)) {
    for (const auto &pdnode : pattern_.nodes()) {
115
      if (pdnode->Tell(&node)) {
116
        VLOG(4) << "Node " << node.Name() << " marked as " << pdnode->name();
117 118 119 120
        pdnodes2nodes_[pdnode.get()].insert(&node);
      }
    }
  }
Y
Yan Chunwei 已提交
121
  // Check to early stop if some PDNode can't find matched Node.
C
chengduo 已提交
122
  for (auto &pdnode : pattern_.nodes()) {
Y
Yan Chunwei 已提交
123
    if (!pdnodes2nodes_.count(pdnode.get())) {
M
minqiyang 已提交
124
      VLOG(4) << pdnode->name() << " can't find matched Node, early stop";
Y
Yan Chunwei 已提交
125
      // return false;
Y
Yan Chunwei 已提交
126 127
    }
  }
M
minqiyang 已提交
128
  VLOG(3) << pdnodes2nodes_.size() << " nodes marked";
129

130 131 132
  return !pdnodes2nodes_.empty();
}

Y
Yan Chunwei 已提交
133
// The intermediate Nodes can only link to the nodes inside the pattern, or this
T
tianshuo78520a 已提交
134
// subgraph will be dropped.
Y
Yan Chunwei 已提交
135
void GraphPatternDetector::ValidateByNodeRole(
C
chengduo 已提交
136
    std::vector<GraphPatternDetector::subgraph_t> *subgraphs) {
Y
Yan Chunwei 已提交
137 138 139 140 141
  std::vector<GraphPatternDetector::subgraph_t> result;

  subgraphs->erase(
      std::remove_if(
          subgraphs->begin(), subgraphs->end(),
C
chengduo 已提交
142
          [](const GraphPatternDetector::subgraph_t &subgraph) -> bool {
Y
Yan Chunwei 已提交
143
            // Collect the inputs and outputs.
144
            std::set<Node *> ios;
C
chengduo 已提交
145
            for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
146 147 148 149
              if (!item.first->IsIntermediate()) {
                ios.insert(item.second);
              }
            }
C
chengduo 已提交
150
            for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
151
              if (item.first->IsIntermediate()) {
C
chengduo 已提交
152
                for (auto *x : item.second->inputs) {
Y
Yan Chunwei 已提交
153 154 155 156
                  if (!ios.count(x)) {
                    return true;
                  }
                }
C
chengduo 已提交
157
                for (auto *x : item.second->outputs) {
Y
Yan Chunwei 已提交
158 159 160 161 162 163 164 165 166 167 168
                  if (!ios.count(x)) {
                    return true;
                  }
                }
              }
            }
            return false;
          }),
      subgraphs->end());
}

169
struct HitGroup {
170
  std::map<PDNode *, Node *> roles;
171

C
chengduo 已提交
172
  bool Match(Node *node, PDNode *pat) {
173
    if (nodes_.count(node)) {
T
Tao Luo 已提交
174 175 176 177 178
      if (roles.count(pat) && roles[pat] == node) return true;
      return false;
    } else {
      if (roles.count(pat) && roles[pat] != node) return false;
      return true;
179
    }
180 181
  }

C
chengduo 已提交
182
  void Register(Node *node, PDNode *pat) {
183 184 185 186 187
    roles[pat] = node;
    nodes_.insert(node);
  }

 private:
188
  std::set<Node *> nodes_;
189 190 191
};

// Tell whether Node a links to b.
C
chengduo 已提交
192 193
bool IsNodesLink(Node *a, Node *b) {
  for (auto *node : a->outputs) {
194 195 196 197 198 199 200
    if (b == node) {
      return true;
    }
  }
  return false;
}

201 202
std::vector<GraphPatternDetector::subgraph_t>
GraphPatternDetector::DetectPatterns() {
203
  // Init empty subgraphs.
204
  std::vector<GraphPatternDetector::subgraph_t> result;
205
  std::vector<HitGroup> init_groups;
206
  std::array<std::vector<HitGroup>, 2> bi_records;
C
chengduo 已提交
207
  auto *first_pnode = pattern_.edges().empty() ? pattern().nodes().front().get()
208
                                               : pattern_.edges().front().first;
209
  if (!pdnodes2nodes_.count(first_pnode)) return result;
C
chengduo 已提交
210
  for (auto *node : pdnodes2nodes_[first_pnode]) {
211 212 213 214 215 216 217 218 219 220
    HitGroup group;
    group.roles[first_pnode] = node;
    init_groups.emplace_back(group);
  }

  int step = 0;
  bi_records[0] = std::move(init_groups);

  // Extend a PDNode to subgraphs by deducing the connection relations defined
  // in edges of PDNodes.
C
chengduo 已提交
221
  for (const auto &edge : pattern_.edges()) {
M
minqiyang 已提交
222
    VLOG(4) << "check " << edge.first->name() << " -> " << edge.second->name();
Y
Yan Chunwei 已提交
223
    // TODO(Superjomn) Fix bug here, the groups might be duplicate here.
224 225
    // Each role has two PDNodes, which indicates two roles.
    // Detect two Nodes that can match these two roles and they are connected.
C
chengduo 已提交
226 227
    auto &pre_groups = bi_records[step % 2];
    auto &cur_groups = bi_records[1 - (step++ % 2)];
228
    cur_groups.clear();
229
    if (pre_groups.empty()) break;
230
    // source -> target
C
chengduo 已提交
231 232
    for (Node *source : pdnodes2nodes_[edge.first]) {
      for (Node *target : pdnodes2nodes_[edge.second]) {
M
minqiyang 已提交
233
        VLOG(8) << "check " << source->id() << " -- " << target->id();
234
        // TODO(Superjomn) add some prune strategies.
C
chengduo 已提交
235
        for (const auto &group : pre_groups) {
T
Tao Luo 已提交
236 237 238 239 240 241
          if (IsNodesLink(source, target)) {
            HitGroup new_group = group;
            bool flag = new_group.Match(source, edge.first) &&
                        new_group.Match(target, edge.second);
            if (flag) {
              new_group.Register(source, edge.first);
242 243 244 245 246 247 248 249
              new_group.Register(target, edge.second);
              cur_groups.push_back(new_group);
              // TODO(Superjomn) need to unique
            }
          }
        }
      }
    }
M
minqiyang 已提交
250
    VLOG(3) << "step " << step << " get records: " << cur_groups.size();
C
chengduo 已提交
251 252
    for (auto &group : cur_groups) {
      for (auto &item : group.roles) {
M
minqiyang 已提交
253
        VLOG(4) << "node " << item.second->id() << " as " << item.first->name();
Y
Yan Chunwei 已提交
254
      }
M
minqiyang 已提交
255
      VLOG(4) << "=========================================================";
Y
Yan Chunwei 已提交
256
    }
257 258
  }

C
chengduo 已提交
259
  for (auto &group : bi_records[step % 2]) {
260
    GraphPatternDetector::subgraph_t subgraph;
C
chengduo 已提交
261
    for (auto &role : group.roles) {
262 263 264 265 266 267 268
      subgraph.emplace(role.first, role.second);
    }
    result.emplace_back(subgraph);
  }
  return result;
}

Y
Yan Chunwei 已提交
269 270
struct GraphItemLessThan {
  bool operator()(const std::pair<PDNode *, Node *> &a,
Y
Yan Chunwei 已提交
271
                  const std::pair<PDNode *, Node *> &b) {
Y
Yan Chunwei 已提交
272 273 274 275 276
    if (a.first != b.first) {
      return a.first < b.first;
    } else {
      return a.second < b.second;
    }
Y
Yan Chunwei 已提交
277
  }
Y
Yan Chunwei 已提交
278
};
Y
Yan Chunwei 已提交
279

280 281
// TODO(Superjomn) enhance the function as it marks unique unique as duplicates
// see https://github.com/PaddlePaddle/Paddle/issues/13550
282
void GraphPatternDetector::UniquePatterns(
C
chengduo 已提交
283
    std::vector<GraphPatternDetector::subgraph_t> *subgraphs) {
284
  if (subgraphs->empty()) return;
285
  std::vector<GraphPatternDetector::subgraph_t> result;
286

287
  std::set<size_t> set;
Y
Yan Chunwei 已提交
288
  std::hash<std::string> hasher;
C
chengduo 已提交
289
  for (auto &g : *subgraphs) {
Y
Yan Chunwei 已提交
290 291
    // Sort the items in the sub-graph, and transform to a string key.
    std::vector<std::pair<PDNode *, Node *>> sorted_keys(g.begin(), g.end());
Y
Yan Chunwei 已提交
292
    std::sort(sorted_keys.begin(), sorted_keys.end(), GraphItemLessThan());
Y
Yan Chunwei 已提交
293 294 295
    std::stringstream ss;
    for (auto &item : sorted_keys) {
      ss << item.first << ":" << item.second;
296
    }
Y
Yan Chunwei 已提交
297
    auto key = hasher(ss.str());
298 299 300 301 302 303 304 305
    if (!set.count(key)) {
      result.emplace_back(g);
      set.insert(key);
    }
  }
  *subgraphs = result;
}

Z
Zhang Ting 已提交
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345
void GraphPatternDetector::SortSubgraphs(
    std::vector<GraphPatternDetector::subgraph_t> *subgraphs) {
  if (subgraphs->empty()) return;
  bool has_bn_add_act = false;
  for (auto &subgraph : *subgraphs) {
    for (auto &item : subgraph) {
      if (item.first->name().find("bn_add_act") != std::string::npos) {
        has_bn_add_act = true;
        break;
      }
    }
  }
  if (!has_bn_add_act) {
    return;
  }

  std::sort(
      subgraphs->begin(), subgraphs->end(),
      [](const GraphPatternDetector::subgraph_t &a,
         const GraphPatternDetector::subgraph_t &b) {
        for (auto &item : a) {
          if (item.first->name().find("bn_add_act") != std::string::npos &&
              item.first->name().find("bn_reserve_space") !=
                  std::string::npos) {
            auto it_b = b.find(item.first);
            if (it_b != b.end()) {
              if (item.second->Name() != it_b->second->Name()) {
                return item.second->Name() < it_b->second->Name();
              } else {
                return false;
              }
            } else {
              return false;
            }
          }
        }
        return false;
      });
}

346
void GraphPatternDetector::RemoveOverlappedMatch(
C
chengduo 已提交
347
    std::vector<subgraph_t> *subgraphs) {
348
  std::vector<subgraph_t> result;
349
  std::set<Node *> node_set;
350

C
chengduo 已提交
351
  for (const auto &subgraph : *subgraphs) {
352
    bool valid = true;
C
chengduo 已提交
353
    for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
354
      if (item.first->IsIntermediate() && node_set.count(item.second)) {
355 356 357 358 359
        valid = false;
        break;
      }
    }
    if (valid) {
C
chengduo 已提交
360
      for (auto &item : subgraph) {
361 362 363 364 365 366 367 368
        node_set.insert(item.second);
      }
      result.push_back(subgraph);
    }
  }
  *subgraphs = result;
}

369 370 371 372 373
std::string PDPattern::DotString() const {
  using inference::analysis::Dot;
  Dot dot;
  int id = 0;
  // Create Nodes
C
chengduo 已提交
374 375
  std::unordered_map<PDNode *, std::string> node2dot;
  for (const auto &node : nodes()) {
376 377 378 379 380
    std::string node_id = "Node" + std::to_string(id++);
    dot.AddNode(node_id, {}, node->name());
    node2dot[node.get()] = node_id;
  }
  // Create Edges
C
chengduo 已提交
381
  for (const auto &edge : edges()) {
382 383 384 385
    if (!node2dot.count(edge.first) || !node2dot.count(edge.second)) {
      LOG(ERROR) << "no node " << edge.first << " " << edge.second;
      continue;
    }
C
chengduo 已提交
386 387
    auto &src = node2dot.at(edge.first);
    auto &trg = node2dot.at(edge.second);
388 389 390 391 392
    dot.AddEdge(src, trg, {});
  }
  return dot.Build();
}

C
chengduo 已提交
393
PDNode &PDNode::LinksTo(const std::vector<PDNode *> &others) {
394
  // extend outlinks.
C
chengduo 已提交
395
  for (PDNode *x : others) {
396 397 398 399 400
    pattern_->AddEdge(this, x);
  }
  return *this;
}

C
chengduo 已提交
401
PDNode &PDNode::LinksFrom(const std::vector<PDNode *> &others) {
402
  // extend outlinks.
C
chengduo 已提交
403
  for (PDNode *x : others) {
404 405 406 407 408
    pattern_->AddEdge(x, this);
  }
  return *this;
}

C
chengduo 已提交
409 410
PDNode *PDNode::assert_is_op() {
  asserts_.emplace_back([](Node *x) { return x && x->IsOp(); });
Y
Yan Chunwei 已提交
411 412
  return this;
}
C
chengduo 已提交
413 414 415

PDNode *PDNode::assert_is_op(const std::string &op_type) {
  asserts_.emplace_back([op_type](Node *x) {
Y
Yan Chunwei 已提交
416 417 418 419
    return x && x->IsOp() && x->Op()->Type() == op_type;
  });
  return this;
}
C
chengduo 已提交
420 421 422 423 424 425

PDNode *PDNode::assert_is_var() {
  asserts_.emplace_back([](Node *x) { return x && x->IsVar(); });
  return this;
}

Z
Zhen Wang 已提交
426 427 428 429 430 431 432
PDNode *PDNode::assert_var_dtype(proto::VarType::Type dtype) {
  assert_is_var();
  asserts_.emplace_back(
      [dtype](Node *x) { return x->Var()->GetDataType() == dtype; });
  return this;
}

C
chengduo 已提交
433 434
PDNode *PDNode::assert_is_not_ctrl_var() {
  asserts_.emplace_back([](Node *x) { return x && !x->IsCtrlVar(); });
Y
Yan Chunwei 已提交
435 436
  return this;
}
C
chengduo 已提交
437 438

PDNode *PDNode::assert_var_not_persistable() {
Y
Yan Chunwei 已提交
439
  assert_is_var();
C
chengduo 已提交
440
  asserts_.emplace_back([](Node *x) { return !x->Var()->Persistable(); });
Y
Yan Chunwei 已提交
441 442
  return this;
}
C
chengduo 已提交
443 444

PDNode *PDNode::assert_is_persistable_var() {
Y
Yan Chunwei 已提交
445
  assert_is_var();
C
chengduo 已提交
446
  asserts_.emplace_back([=](Node *x) { return x->Var()->Persistable(); });
Y
Yan Chunwei 已提交
447 448
  return this;
}
C
chengduo 已提交
449 450 451

PDNode *PDNode::assert_is_op_nth_input(const std::string &op_type,
                                       const std::string &argument, int nth) {
Y
Yan Chunwei 已提交
452 453
  assert_is_var();
  assert_is_op_input(op_type);
C
chengduo 已提交
454 455
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
456 457 458
      if (op->IsOp() && op->Op()->Type() == op_type &&
          IsNthInput(x, op, argument, nth))
        return true;
Y
Yan Chunwei 已提交
459 460 461 462 463
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
464 465 466

PDNode *PDNode::assert_is_op_nth_output(const std::string &op_type,
                                        const std::string &argument, int nth) {
Y
Yan Chunwei 已提交
467
  assert_is_var();
C
chengduo 已提交
468 469
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
470 471 472
      if (op->IsOp() && op->Op()->Type() == op_type &&
          IsNthOutput(x, op, argument, nth))
        return true;
Y
Yan Chunwei 已提交
473 474 475 476 477
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
478 479

PDNode *PDNode::assert_is_only_input_of_op(const std::string &op_type) {
Y
Yan Chunwei 已提交
480
  assert_is_var();
C
chengduo 已提交
481 482
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
Y
Yan Chunwei 已提交
483 484 485 486 487 488 489 490 491
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type &&
          op->inputs.size() == 1) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
492 493

PDNode *PDNode::assert_is_only_output_of_op(const std::string &op_type) {
Y
Yan Chunwei 已提交
494
  assert_is_var();
C
chengduo 已提交
495 496
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
Y
Yan Chunwei 已提交
497 498 499 500 501 502 503 504 505
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type &&
          op->outputs.size() == 1) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
506 507

PDNode *PDNode::assert_is_op_output(const std::string &op_type) {
Y
Yan Chunwei 已提交
508
  assert_is_var();
C
chengduo 已提交
509 510
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
Y
Yan Chunwei 已提交
511 512 513 514 515 516 517 518
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
519 520 521

PDNode *PDNode::assert_is_op_output(const std::string &op_type,
                                    const std::string &argument) {
522 523 524 525
  assert_is_var();
  assert_is_op_nth_output(op_type, argument, 0);
  return this;
}
Z
Zhen Wang 已提交
526

C
chengduo 已提交
527
PDNode *PDNode::assert_is_op_input(const std::string &op_type) {
Y
Yan Chunwei 已提交
528
  assert_is_var();
C
chengduo 已提交
529 530
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
Y
Yan Chunwei 已提交
531 532 533 534 535 536 537 538
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
539

Z
Zhen Wang 已提交
540 541 542 543 544 545 546 547 548 549
PDNode *PDNode::assert_is_not_op_input(const std::string &argument) {
  assert_is_op();
  asserts_.emplace_back([=](Node *x) {
    auto &ins = x->Op()->Inputs();
    auto iter = ins.find(argument);
    return iter == ins.end() || iter->second.empty();
  });
  return this;
}

C
chengduo 已提交
550 551
PDNode *PDNode::assert_is_op_input(const std::string &op_type,
                                   const std::string &argument) {
552 553 554 555
  assert_is_var();
  assert_is_op_nth_input(op_type, argument, 0);
  return this;
}
C
chengduo 已提交
556 557

PDNode *PDNode::assert_op_has_n_inputs(const std::string &op_type, size_t n) {
Y
Yan Chunwei 已提交
558
  assert_is_op(op_type);
C
chengduo 已提交
559
  asserts_.emplace_back([=](Node *x) { return x->inputs.size() == n; });
Y
Yan Chunwei 已提交
560 561
  return this;
}
C
chengduo 已提交
562 563

PDNode *PDNode::assert_op_has_n_outputs(const std::string &op_type, size_t n) {
Y
Yan Chunwei 已提交
564
  assert_is_op(op_type);
C
chengduo 已提交
565
  asserts_.emplace_back([=](Node *x) { return x->outputs.size() == n; });
Y
Yan Chunwei 已提交
566 567
  return this;
}
C
chengduo 已提交
568

569 570 571 572 573 574 575 576 577 578
PDNode *PDNode::assert_has_n_inputs(size_t n) {
  asserts_.emplace_back([=](Node *x) { return x->inputs.size() == n; });
  return this;
}

PDNode *PDNode::assert_has_n_outputs(size_t n) {
  asserts_.emplace_back([=](Node *x) { return x->outputs.size() == n; });
  return this;
}

C
chengduo 已提交
579
PDNode *PDNode::assert_more(PDNode::teller_t &&teller) {
Y
Yan Chunwei 已提交
580 581 582 583
  asserts_.emplace_back(std::move(teller));
  return this;
}

C
chengduo 已提交
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
PDNode *PDNode::assert_is_ops(const std::unordered_set<std::string> &op_types) {
  asserts_.emplace_back([op_types](Node *x) {
    return x && x->IsOp() && op_types.count(x->Op()->Type());
  });
  return this;
}

PDNode *PDNode::assert_is_ops_nth_input(
    const std::unordered_set<std::string> &op_types,
    const std::string &argument, int nth) {
  assert_is_var();
  assert_is_ops_input(op_types);
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
      if (op->IsOp() && op_types.count(op->Op()->Type()) &&
          IsNthInput(x, op, argument, nth))
        return true;
    }
    return false;
  });
  return this;
}

PDNode *PDNode::assert_is_ops_nth_output(
    const std::unordered_set<std::string> &op_types,
    const std::string &argument, int nth) {
  assert_is_var();
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
      if (op->IsOp() && op_types.count(op->Op()->Type()) &&
          IsNthOutput(x, op, argument, nth))
        return true;
    }
    return false;
  });
  return this;
}
PDNode *PDNode::assert_is_ops_output(
    const std::unordered_set<std::string> &op_types) {
  assert_is_var();
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
      if (op && op->IsOp() && op->Op() && op_types.count(op->Op()->Type())) {
        return true;
      }
    }
    return false;
  });
  return this;
}

PDNode *PDNode::assert_is_ops_output(
    const std::unordered_set<std::string> &op_types,
    const std::string &argument) {
  assert_is_var();
  assert_is_ops_nth_output(op_types, argument, 0);
  return this;
}

PDNode *PDNode::assert_is_ops_input(
    const std::unordered_set<std::string> &op_types) {
  assert_is_var();
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
      if (op && op->IsOp() && op->Op() && op_types.count(op->Op()->Type())) {
        return true;
      }
    }
    return false;
  });
  return this;
}

PDNode *PDNode::assert_is_ops_input(
    const std::unordered_set<std::string> &op_types,
    const std::string &argument) {
  assert_is_var();
  assert_is_ops_nth_input(op_types, argument, 0);
  return this;
}

bool VarLinksToOp(Node *node, const std::string &op_type) {
  for (auto *out : node->outputs) {
667 668 669 670 671 672
    if (out->IsOp() && out->Op()->Type() == op_type) {
      return true;
    }
  }
  return false;
}
C
chengduo 已提交
673 674

bool IsNthInput(Node *var, Node *op, const std::string &argument, size_t nth) {
675 676 677 678 679 680 681 682
  PADDLE_ENFORCE_EQ(
      var->IsVar(), true,
      platform::errors::InvalidArgument(
          "First parameter of function IsNthInput must be Node::Var"));
  PADDLE_ENFORCE_EQ(
      op->IsOp(), true,
      platform::errors::InvalidArgument(
          "Second parameter of function IsNthInput must be Node::Op"));
683 684
  if (!HasInput(op, argument) || op->Op()->Input(argument).size() <= nth)
    return false;
685 686
  return var->Name() == op->Op()->Input(argument)[nth];
}
C
chengduo 已提交
687

688
bool HasInput(Node *op, const std::string &argument) {
689 690 691 692
  PADDLE_ENFORCE_EQ(
      op->IsOp(), true,
      platform::errors::InvalidArgument(
          "First parameter of function HasInput must be Node::Op"));
693 694 695 696 697 698
  auto const &names = op->Op()->InputNames();
  if (std::find(names.begin(), names.end(), argument) == names.end())
    return false;
  return true;
}

699 700 701 702 703 704 705 706 707 708 709
bool HasOutput(Node *op, const std::string &argument) {
  PADDLE_ENFORCE_EQ(
      op->IsOp(), true,
      platform::errors::InvalidArgument(
          "First parameter of function HasOuput must be Node::Op"));
  auto const &names = op->Op()->OutputNames();
  if (std::find(names.begin(), names.end(), argument) == names.end())
    return false;
  return true;
}

C
chengduo 已提交
710
bool IsNthOutput(Node *var, Node *op, const std::string &argument, size_t nth) {
711 712 713 714 715 716 717 718
  PADDLE_ENFORCE_EQ(
      var->IsVar(), true,
      platform::errors::InvalidArgument(
          "First parameter of function IsNthOutput must be Node::Var"));
  PADDLE_ENFORCE_EQ(
      op->IsOp(), true,
      platform::errors::InvalidArgument(
          "Second parameter of function IsNthOutput must be Node::Op"));
719 720
  if (!HasOutput(op, argument) || op->Op()->Output(argument).size() <= nth)
    return false;
721 722
  return var->Name() == op->Op()->Output(argument)[nth];
}
C
chengduo 已提交
723 724 725 726 727

void GraphSafeRemoveNodes(Graph *graph,
                          const std::unordered_set<const Node *> &nodes) {
  for (auto *node : nodes) {
    graph->RemoveNode(const_cast<Node *>(node));
728 729
  }

C
chengduo 已提交
730
  for (auto *node : graph->Nodes()) {
731 732
    for (auto it = node->inputs.begin(); it != node->inputs.end();) {
      if (nodes.count(*it)) {
C
chengduo 已提交
733
        it = const_cast<Node *>(node)->inputs.erase(it);
734
      } else {
735
        it++;
736
      }
737 738 739
    }
    for (auto it = node->outputs.begin(); it != node->outputs.end();) {
      if (nodes.count(*it)) {
C
chengduo 已提交
740
        it = const_cast<Node *>(node)->outputs.erase(it);
741
      } else {
742
        it++;
743
      }
744 745 746
    }
  }
}
C
chengduo 已提交
747 748 749

bool VarLinksFromOp(Node *node, const std::string &op_type) {
  for (auto *out : node->inputs) {
750 751 752 753 754 755 756
    if (out->IsOp() && out->Op()->Type() == op_type) {
      return true;
    }
  }
  return false;
}

S
Sylwester Fraczek 已提交
757
PDNode *patterns::ConvBN::operator()(paddle::framework::ir::PDNode *conv_input,
758
                                     const std::string &conv_type,
S
Sylwester Fraczek 已提交
759 760
                                     bool with_eltwise_add) {
  // Create Operators
761 762
  conv_input->assert_is_op_input(conv_type, "Input");
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op(conv_type);
S
Sylwester Fraczek 已提交
763 764 765 766 767 768 769 770 771 772 773 774 775

  PDNode *eltwise_op = nullptr;
  if (with_eltwise_add) {
    eltwise_op =
        pattern->NewNode(eltwise_repr())->assert_is_op("elementwise_add");
  }
  auto *batch_norm_op =
      pattern->NewNode(batch_norm_repr())->assert_is_op("batch_norm");
  // Create variables
  // Conv Filter
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
776
                              ->assert_is_op_input(conv_type, "Filter");
S
Sylwester Fraczek 已提交
777 778 779

  auto *conv_out_var = pattern->NewNode(conv_out_repr())
                           ->AsIntermediate()
780
                           ->assert_is_only_output_of_op(conv_type);
S
Sylwester Fraczek 已提交
781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863

  PDNode *eltwise_y_in_var = nullptr;
  PDNode *eltwise_out_var = nullptr;
  if (with_eltwise_add) {
    // Conv output as Bias input
    conv_out_var->assert_is_op_input("elementwise_add", "X");
    // Bias
    eltwise_y_in_var = pattern->NewNode(eltwise_y_in_repr())
                           ->assert_is_op_input("elementwise_add", "Y")
                           ->AsInput();
    eltwise_out_var = pattern->NewNode(eltwise_out_repr())
                          ->AsIntermediate()
                          ->assert_is_only_output_of_op("elementwise_add");
  } else {
    // Conv output as BN input
    conv_out_var->assert_is_op_input("batch_norm", "X");
  }

  // BN Scale
  auto *bn_scale_var = pattern->NewNode(bn_scale_repr())
                           ->AsInput()
                           ->assert_is_persistable_var()
                           ->assert_is_op_input("batch_norm", "Scale");
  // BN Bias
  auto *bn_bias_var = pattern->NewNode(bn_bias_repr())
                          ->AsInput()
                          ->assert_is_persistable_var()
                          ->assert_is_op_input("batch_norm", "Bias");
  // BN Mean
  auto *bn_mean_var = pattern->NewNode(bn_mean_repr())
                          ->AsInput()
                          ->assert_is_persistable_var()
                          ->assert_is_op_input("batch_norm", "Mean");
  // BN Variance
  auto *bn_variance_var = pattern->NewNode(bn_variance_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
                              ->assert_is_op_input("batch_norm", "Variance");

  // BN output
  auto *bn_out_var = pattern->NewNode(bn_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("batch_norm");

  auto *bn_mean_out_var = pattern->NewNode(bn_mean_out_repr())
                              ->AsOutput()
                              ->assert_is_op_output("batch_norm", "MeanOut");

  auto *bn_variance_out_var =
      pattern->NewNode(bn_variance_out_repr())
          ->AsOutput()
          ->assert_is_op_output("batch_norm", "VarianceOut");

  auto *bn_saved_mean_var =
      pattern->NewNode(bn_saved_mean_repr())
          ->AsOutput()
          ->assert_is_op_output("batch_norm", "SavedMean");

  auto *bn_saved_variance_var =
      pattern->NewNode(bn_saved_variance_repr())
          ->AsOutput()
          ->assert_is_op_output("batch_norm", "SavedVariance");

  conv_op->LinksFrom({conv_input, conv_weight_var}).LinksTo({conv_out_var});

  if (with_eltwise_add) {
    eltwise_op->LinksFrom({conv_out_var, eltwise_y_in_var})
        .LinksTo({eltwise_out_var});
    batch_norm_op
        ->LinksFrom({eltwise_out_var, bn_scale_var, bn_bias_var, bn_mean_var,
                     bn_variance_var})
        .LinksTo({bn_out_var, bn_mean_out_var, bn_variance_out_var,
                  bn_saved_mean_var, bn_saved_variance_var});
  } else {
    batch_norm_op
        ->LinksFrom({conv_out_var, bn_scale_var, bn_bias_var, bn_mean_var,
                     bn_variance_var})
        .LinksTo({bn_out_var, bn_mean_out_var, bn_variance_out_var,
                  bn_saved_mean_var, bn_saved_variance_var});
  }
  return bn_out_var;
}

864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892
PDNode *patterns::ConvActivation::operator()(
    paddle::framework::ir::PDNode *conv_input, std::string conv_type,
    std::string activation_type) {
  // Create Operators
  conv_input->assert_is_op_input(conv_type, "Input");
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op(conv_type);
  auto *activation_op =
      pattern->NewNode(activation_repr())->assert_is_op(activation_type);
  // Create variables
  // Filter
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
                              ->assert_is_op_input(conv_type, "Filter");
  // intermediate variable, will be removed in the IR after fuse.
  auto *conv_out_var = pattern->NewNode(conv_out_repr())
                           ->AsIntermediate()
                           ->assert_is_only_output_of_op(conv_type)
                           ->assert_is_op_input(activation_type);
  // output
  auto *activation_out_var = pattern->NewNode(activation_out_repr())
                                 ->AsOutput()
                                 ->assert_is_op_output(activation_type);

  conv_op->LinksFrom({conv_input, conv_weight_var}).LinksTo({conv_out_var});
  activation_op->LinksFrom({conv_out_var}).LinksTo({activation_out_var});
  return activation_out_var;
}

T
tensor-tang 已提交
893 894 895 896
PDNode *patterns::SeqConvEltAddRelu::operator()(
    paddle::framework::ir::PDNode *seqconv_input) {
  // Create Operators
  seqconv_input->assert_is_op_input("sequence_conv", "X");
T
tensor-tang 已提交
897 898 899 900
  auto *seqconv_op = pattern->NewNode(seqconv_repr())
                         ->assert_is_op("sequence_conv")
                         ->assert_op_attr<bool>("paddingTrainable", false)
                         ->assert_op_attr<int>("contextStride", 1);
T
tensor-tang 已提交
901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937

  auto *eltadd_op =
      pattern->NewNode(eltadd_repr())->assert_is_op("elementwise_add");
  auto *relu_op = pattern->NewNode(relu_repr())->assert_is_op("relu");
  // Create variables
  // Filter
  auto *seqconv_weight_var =
      pattern->NewNode(seqconv_weight_repr())
          ->AsInput()
          ->assert_is_persistable_var()
          ->assert_is_op_input("sequence_conv", "Filter");
  // Bias
  auto *eltadd_bias_var = pattern->NewNode(eltadd_bias_repr())
                              ->AsInput()
                              ->assert_is_op_input("elementwise_add");
  // intermediate variable, will be removed in the IR after fuse.
  auto *seqconv_out_var = pattern->NewNode(seqconv_out_repr())
                              ->AsIntermediate()
                              ->assert_is_only_output_of_op("sequence_conv")
                              ->assert_is_op_input("elementwise_add");
  auto *eltadd_out_var = pattern->NewNode(eltadd_out_repr())
                             ->AsIntermediate()
                             ->assert_is_only_output_of_op("elementwise_add")
                             ->assert_is_only_input_of_op("relu");
  // output
  auto *relu_out_var = pattern->NewNode(relu_out_repr())
                           ->AsOutput()
                           ->assert_is_op_output("relu");

  seqconv_op->LinksFrom({seqconv_input, seqconv_weight_var})
      .LinksTo({seqconv_out_var});
  eltadd_op->LinksFrom({seqconv_out_var, eltadd_bias_var})
      .LinksTo({eltadd_out_var});
  relu_op->LinksFrom({eltadd_out_var}).LinksTo({relu_out_var});
  return relu_out_var;
}

C
chengduo 已提交
938
PDNode *patterns::FC::operator()(paddle::framework::ir::PDNode *x,
939
                                 bool with_bias, bool with_relu) {
Y
Yan Chunwei 已提交
940 941
  // Create shared nodes.
  x->assert_is_op_input("mul", "X");
C
chengduo 已提交
942
  auto *mul = pattern->NewNode(mul_repr())->assert_is_op("mul");
Y
Yan Chunwei 已提交
943

C
chengduo 已提交
944
  auto *mul_w_var = pattern->NewNode(w_repr())
Y
Yan Chunwei 已提交
945 946 947 948
                        ->AsInput()
                        ->assert_is_persistable_var()
                        ->assert_is_op_input("mul", "Y");

C
chengduo 已提交
949
  auto *mul_out_var =
Y
Yan Chunwei 已提交
950 951
      pattern->NewNode(mul_out_repr())->assert_is_op_output("mul");

952 953
  // Add links.
  mul->LinksFrom({x, mul_w_var}).LinksTo({mul_out_var});
Y
Yan Chunwei 已提交
954 955 956 957 958
  if (!with_bias) {  // not with bias
    return mul_out_var;
  } else {  // with bias
    mul_out_var->AsIntermediate()->assert_is_op_input("elementwise_add");
    // Create operators.
C
chengduo 已提交
959
    auto *elementwise_add = pattern->NewNode(elementwise_add_repr())
Y
Yan Chunwei 已提交
960 961
                                ->assert_is_op("elementwise_add");
    // Create variables.
C
chengduo 已提交
962
    auto *bias = pattern->NewNode(bias_repr())
Y
Yan Chunwei 已提交
963
                     ->assert_is_op_input("elementwise_add")
964
                     ->assert_is_persistable_var()
Y
Yan Chunwei 已提交
965 966
                     ->AsInput();

967 968 969 970
    auto *elementwise_add_out_var =
        pattern->NewNode(elementwise_add_out_repr())
            ->AsOutput()
            ->assert_is_op_output("elementwise_add");
Y
Yan Chunwei 已提交
971

972 973 974 975 976 977 978 979 980 981 982 983 984 985 986
    elementwise_add->LinksFrom({mul_out_var, bias})
        .LinksTo({elementwise_add_out_var});
    if (!with_relu) {
      return elementwise_add_out_var;
    } else {
      elementwise_add_out_var->AsIntermediate()->assert_is_op_input("relu");
      // Create operators.
      auto *relu = pattern->NewNode(relu_repr())->assert_is_op("relu");
      auto *relu_out_var = pattern->NewNode(relu_out_repr())
                               ->AsOutput()
                               ->assert_is_op_output("relu");

      relu->LinksFrom({elementwise_add_out_var}).LinksTo({relu_out_var});
      return relu_out_var;
    }
987 988
  }
}
T
tensor-tang 已提交
989

990 991 992 993 994 995 996
PDNode *patterns::FCMKLDNN::operator()(paddle::framework::ir::PDNode *x,
                                       bool with_bias) {
  // Create shared nodes.
  x->assert_is_op_input("fc", "Input");

  auto *fc_op = pattern->NewNode(fc_repr())->assert_is_op("fc");
  // Create variables
M
Michał Gallus 已提交
997 998 999 1000
  // Input
  auto *input_var = pattern->NewNode(input_repr())
                        ->AsInput()
                        ->assert_is_op_input("fc", "Input");
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
  // Filter
  auto *fc_weight_var = pattern->NewNode(weights_repr())
                            ->AsInput()
                            ->assert_is_op_input("fc", "W");
  // Bias
  auto *fc_bias_var = pattern->NewNode(bias_repr())
                          ->AsInput()
                          ->assert_is_op_input("fc", "Bias");
  // Output
  auto *fc_out_var = pattern->NewNode(output_repr())
                         ->AsOutput()
                         ->assert_is_op_output("fc", "Out")
                         ->assert_is_only_output_of_op("fc");

M
Michał Gallus 已提交
1015 1016
  fc_op->LinksFrom({input_var, fc_weight_var, fc_bias_var})
      .LinksTo({fc_out_var});
1017 1018 1019
  return fc_out_var;
}

1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
PDNode *patterns::Embedding::operator()(PDNode *x) {
  x->assert_is_op_input("lookup_table", "Ids");
  auto *lookup_table_op =
      pattern->NewNode(lookup_table_repr())->assert_is_op("lookup_table");
#define NEW_NODE(arg__, io__)                    \
  auto *arg__ = pattern->NewNode(arg__##_repr()) \
                    ->assert_is_op_##io__("lookup_table", #arg__);

  NEW_NODE(W, input);

  NEW_NODE(Out, output);
#undef NEW_NODE

  lookup_table_op->LinksFrom({x, W});
  lookup_table_op->LinksTo({Out});
  return Out;
}

C
chengduo 已提交
1038
PDNode *patterns::LSTM::operator()(PDNode *x) {
1039
  x->assert_is_op_input("lstm", "Input");
C
chengduo 已提交
1040
  auto *lstm_op = pattern->NewNode(lstm_repr())->assert_is_op("lstm");
Y
Yan Chunwei 已提交
1041
#define NEW_NODE(arg__, io__) \
C
chengduo 已提交
1042
  auto *arg__ =               \
Y
Yan Chunwei 已提交
1043
      pattern->NewNode(arg__##_repr())->assert_is_op_##io__("lstm", #arg__);
1044 1045 1046 1047 1048

  // Currently, the H0 and C0 are optional
  // TODO(Superjomn) upgrade the fuse framework to support optional.
  // NEW_NODE(H0, input);
  // NEW_NODE(C0, input);
Y
Yan Chunwei 已提交
1049 1050
  NEW_NODE(Weight, input);
  NEW_NODE(Bias, input);
1051

Y
Yan Chunwei 已提交
1052 1053 1054 1055 1056
  NEW_NODE(Hidden, output);
  NEW_NODE(Cell, output);
  NEW_NODE(BatchGate, output);
  NEW_NODE(BatchCellPreAct, output);
#undef NEW_NODE
1057 1058 1059 1060 1061

  lstm_op->LinksFrom({x, Weight, Bias});
  lstm_op->LinksTo({Hidden, Cell, BatchGate, BatchCellPreAct});
  return Hidden;
}
T
tensor-tang 已提交
1062

C
chengduo 已提交
1063
PDNode *patterns::GRU::operator()(PDNode *x) {
T
tensor-tang 已提交
1064
  x->assert_is_op_input("gru", "Input");
C
chengduo 已提交
1065
  auto *gru_op = pattern->NewNode(gru_repr())->assert_is_op("gru");
Y
Yan Chunwei 已提交
1066
#define NEW_NODE(arg__, io__) \
C
chengduo 已提交
1067
  auto *arg__ =               \
Y
Yan Chunwei 已提交
1068
      pattern->NewNode(arg__##_repr())->assert_is_op_##io__("gru", #arg__);
T
tensor-tang 已提交
1069

Y
Yan Chunwei 已提交
1070
  NEW_NODE(Weight, input);
T
tensor-tang 已提交
1071 1072
  // TODO(Superjomn): upgrade the fuse framework to support optional.
  // H0 and bias are optional
Y
Yan Chunwei 已提交
1073
  NEW_NODE(Bias, input);  // also optional
T
tensor-tang 已提交
1074 1075
  // NEW_NODE(H0, input);

Y
Yan Chunwei 已提交
1076
  NEW_NODE(Hidden, output);
T
tensor-tang 已提交
1077
  // below are intermediate
Y
Yan Chunwei 已提交
1078 1079 1080 1081
  NEW_NODE(BatchGate, output);
  NEW_NODE(BatchResetHiddenPrev, output);
  NEW_NODE(BatchHidden, output);
#undef NEW_NODE
T
tensor-tang 已提交
1082

T
tensor-tang 已提交
1083 1084 1085 1086
  BatchGate->AsIntermediate();
  BatchResetHiddenPrev->AsIntermediate();
  BatchHidden->AsIntermediate();

T
tensor-tang 已提交
1087 1088 1089 1090 1091
  gru_op->LinksFrom({x, Weight, Bias});
  gru_op->LinksTo({Hidden, BatchGate, BatchResetHiddenPrev, BatchHidden});
  return Hidden;
}

C
chengduo 已提交
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
PDNode *patterns::ActElewiseAdd::operator()(
    paddle::framework::ir::PDNode *in_var,
    std::unordered_set<std::string> act_types) {
  in_var->assert_is_ops_input(act_types, "X");

  auto *act = pattern->NewNode(act_repr())->assert_is_ops(act_types);
  auto *act_out_var = pattern->NewNode(act_out_repr())
                          ->assert_is_not_ctrl_var()
                          ->assert_is_ops_output(act_types);
  act_out_var->AsIntermediate()->assert_is_op_input("elementwise_add");

  auto *ele_x_var = pattern->NewNode(ele_x_repr())
                        ->assert_is_not_ctrl_var()
                        ->assert_is_op_input("elementwise_add")
                        ->AsInput();
  auto *elementwise_add =
      pattern->NewNode(ele_add_repr())->assert_is_op("elementwise_add");

  auto *elewise_add_out = pattern->NewNode(elewise_add_out_repr())
                              ->AsOutput()
                              ->assert_is_op_output("elementwise_add", "Out");

  act->LinksFrom({in_var}).LinksTo({act_out_var});
  elementwise_add->LinksFrom({act_out_var, ele_x_var})
      .LinksTo({elewise_add_out});

  return elewise_add_out;
}

Z
Zhen Wang 已提交
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
PDNode *patterns::BatchNormAct::operator()(
    paddle::framework::ir::PDNode *bn_x_var,
    std::unordered_set<std::string> act_types) {
  auto *bn_scale_var = pattern->NewNode(bn_scale_repr())
                           ->assert_is_op_input("batch_norm", "Scale");
  auto *bn_bias_var = pattern->NewNode(bn_bias_repr())
                          ->assert_is_op_input("batch_norm", "Bias");
  auto *bn_variance_var = pattern->NewNode(bn_variance_repr())
                              ->assert_is_op_input("batch_norm", "Variance");
  auto *bn_mean_var = pattern->NewNode(bn_mean_repr())
                          ->assert_is_op_input("batch_norm", "Mean");

  auto *bn = pattern->NewNode(batch_norm_repr())
                 ->assert_is_op("batch_norm")
                 ->assert_is_not_op_input("MomentumTensor")
                 ->assert_op_attr<bool>("is_test", false)
                 ->assert_op_attr<bool>("use_global_stats", false)
                 ->assert_op_attr<std::string>("data_layout", "NHWC");

  auto *bn_mean_out_var = pattern->NewNode(bn_mean_out_repr())
                              ->assert_is_op_output("batch_norm", "MeanOut");
  auto *bn_variance_out_var =
      pattern->NewNode(bn_variance_out_repr())
          ->assert_is_op_output("batch_norm", "VarianceOut");
  auto *bn_saved_variance_var =
      pattern->NewNode(bn_saved_variance_repr())
          ->assert_is_op_output("batch_norm", "SavedVariance");
  auto *bn_saved_mean_var =
      pattern->NewNode(bn_saved_mean_repr())
          ->assert_is_op_output("batch_norm", "SavedMean");
  auto *bn_reserve_space =
      pattern->NewNode(bn_reserve_space_repr())
          ->assert_is_op_output("batch_norm", "ReserveSpace");
  auto *bn_out_var = pattern->NewNode(bn_out_repr())
                         ->assert_is_op_output("batch_norm", "Y")
                         ->assert_has_n_outputs(1);

  bn_out_var->AsIntermediate()->assert_is_ops_input(act_types);

  auto *act = pattern->NewNode(act_repr())->assert_is_ops(act_types);

  auto *act_out_var =
      pattern->NewNode(act_out_repr())->assert_is_ops_output(act_types, "Out");

  bn->LinksFrom(
        {bn_x_var, bn_scale_var, bn_bias_var, bn_variance_var, bn_mean_var})
      .LinksTo({bn_mean_out_var, bn_variance_out_var, bn_saved_variance_var,
                bn_saved_mean_var, bn_reserve_space, bn_out_var});
  act->LinksFrom({bn_out_var}).LinksTo({act_out_var});

  return act_out_var;
}

PDNode *patterns::BatchNormActGrad::operator()(
    paddle::framework::ir::PDNode *d_act_out_var,
    std::unordered_set<std::string> act_grad_types) {
  auto *act_grad =
      pattern->NewNode(act_grad_repr())->assert_is_ops(act_grad_types);
  auto *bn_grad = pattern->NewNode(batch_norm_grad_repr())
                      ->assert_is_op("batch_norm_grad")
                      ->assert_op_attr<bool>("use_global_stats", false)
                      ->assert_op_attr<std::string>("data_layout", "NHWC");

  auto *act_out_var = pattern->NewNode(act_out_repr())
                          ->assert_is_ops_input(act_grad_types, "Out");
  auto *d_intermediate_var =
      pattern->NewNode(d_itermediate_out_repr())
          ->assert_is_ops_output(act_grad_types, GradVarName("X"))
          ->assert_has_n_outputs(1);
  auto *bn_x_var = pattern->NewNode(bn_x_repr())
                       ->assert_is_op_input("batch_norm_grad", "X")
                       ->assert_var_dtype(proto::VarType::FP16);
  auto *bn_scale_var = pattern->NewNode(bn_scale_repr())
                           ->assert_is_op_input("batch_norm_grad", "Scale");
  auto *bn_bias_var = pattern->NewNode(bn_bias_repr())
                          ->assert_is_op_input("batch_norm_grad", "Bias");
  auto *bn_saved_mean_var =
      pattern->NewNode(bn_saved_mean_repr())
          ->assert_is_op_input("batch_norm_grad", "SavedMean");
  auto *bn_saved_variance_var =
      pattern->NewNode(bn_saved_variance_repr())
          ->assert_is_op_input("batch_norm_grad", "SavedVariance");
  // ReserveSpace as the output is equal to:
  // data_layout == 'NHWC' && FLAGS_cudnn_batchnorm_spatial_persistent == true
  auto *bn_reserve_space =
      pattern->NewNode(bn_reserve_space_repr())
          ->assert_is_op_input("batch_norm_grad", "ReserveSpace");
  auto *d_bn_x_var =
      pattern->NewNode(d_bn_x_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("X"));
  auto *d_bn_scale_var =
      pattern->NewNode(d_bn_scale_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("Scale"));
  auto *d_bn_bias_var =
      pattern->NewNode(d_bn_bias_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("Bias"));

  act_grad->LinksFrom({d_act_out_var, act_out_var})
      .LinksTo({d_intermediate_var});

  bn_grad
      ->LinksFrom({bn_x_var, d_intermediate_var, bn_scale_var, bn_bias_var,
                   bn_saved_mean_var, bn_saved_variance_var, bn_reserve_space})
      .LinksTo({d_bn_x_var, d_bn_scale_var, d_bn_bias_var});

  return bn_grad;
}

1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
PDNode *patterns::BatchNormActOneDNN::operator()(const std::string &act_type) {
  auto *bn_x = pattern->NewNode(bn_in_repr())
                   ->AsInput()
                   ->assert_is_op_input("batch_norm", "X");
  auto *bn = pattern->NewNode(batch_norm_repr())->assert_is_op("batch_norm");
  auto *bn_out = pattern->NewNode(bn_out_repr())
                     ->assert_is_op_output("batch_norm", "Y")
                     ->assert_is_op_input(act_type);
  auto *act =
      pattern->NewNode(act_repr())->assert_is_op(act_type)->AsIntermediate();
  auto *act_out = pattern->NewNode(act_out_repr())
                      ->assert_is_op_output(act_type, "Out")
                      ->AsOutput();

  bn->LinksFrom({bn_x}).LinksTo({bn_out});
  act->LinksFrom({bn_out}).LinksTo({act_out});

  return act_out;
}

Z
Zhang Ting 已提交
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
PDNode *patterns::BatchNormAddAct::operator()(
    paddle::framework::ir::PDNode *bn_x_var,
    std::unordered_set<std::string> act_types) {
  bn_x_var->assert_is_op_input("batch_norm", "X")
      ->assert_var_dtype(proto::VarType::FP16);
  auto *bn_scale_var = pattern->NewNode(bn_scale_repr())
                           ->assert_is_op_input("batch_norm", "Scale");
  auto *bn_bias_var = pattern->NewNode(bn_bias_repr())
                          ->assert_is_op_input("batch_norm", "Bias");

  auto *bn = pattern->NewNode(batch_norm_repr())
                 ->assert_is_op("batch_norm")
                 ->assert_is_not_op_input("MomentumTensor")
                 ->assert_op_attr<bool>("is_test", false)
                 ->assert_op_attr<bool>("use_global_stats", false)
                 ->assert_op_attr<std::string>("data_layout", "NHWC");

  auto *bn_mean_out_var = pattern->NewNode(bn_mean_out_repr())
                              ->assert_is_op_output("batch_norm", "MeanOut");
  auto *bn_variance_out_var =
      pattern->NewNode(bn_variance_out_repr())
          ->assert_is_op_output("batch_norm", "VarianceOut");
  auto *bn_saved_variance_var =
      pattern->NewNode(bn_saved_variance_repr())
          ->assert_is_op_output("batch_norm", "SavedVariance");
  auto *bn_saved_mean_var =
      pattern->NewNode(bn_saved_mean_repr())
          ->assert_is_op_output("batch_norm", "SavedMean");
  auto *bn_reserve_space =
      pattern->NewNode(bn_reserve_space_repr())
          ->assert_is_op_output("batch_norm", "ReserveSpace");
  auto *bn_out_var = pattern->NewNode(bn_out_repr())
                         ->assert_is_op_output("batch_norm", "Y")
                         ->assert_var_dtype(proto::VarType::FP16);

  bn_out_var->assert_is_op_input("elementwise_add");

  auto *elewise_add =
      pattern->NewNode(elewise_add_repr())->assert_is_op("elementwise_add");

  auto *elewise_add_in_var = pattern->NewNode(elewise_add_in_repr())
                                 ->assert_is_not_ctrl_var()
                                 ->assert_is_op_input("elementwise_add")
                                 ->assert_var_dtype(proto::VarType::FP16);

  auto *elewise_add_out_var =
      pattern->NewNode(elewise_add_out_repr())
          ->assert_is_op_output("elementwise_add", "Out")
          ->assert_has_n_outputs(1);

  elewise_add_out_var->AsIntermediate()->assert_is_ops_input(act_types);

  auto *act = pattern->NewNode(act_repr())->assert_is_ops(act_types);

  auto *act_out_var =
      pattern->NewNode(act_out_repr())->assert_is_ops_output(act_types, "Out");

  bn->LinksFrom({bn_x_var, bn_scale_var, bn_bias_var})
      .LinksTo({bn_mean_out_var, bn_variance_out_var, bn_saved_variance_var,
                bn_saved_mean_var, bn_reserve_space, bn_out_var});
  elewise_add->LinksFrom({elewise_add_in_var, bn_out_var})
      .LinksTo({elewise_add_out_var});
  act->LinksFrom({elewise_add_out_var}).LinksTo({act_out_var});

  return act_out_var;
}

PDNode *patterns::BatchNormAddActGrad::operator()(
    paddle::framework::ir::PDNode *d_act_out_var,
    std::unordered_set<std::string> act_grad_types) {
  auto *act_grad =
      pattern->NewNode(act_grad_repr())->assert_is_ops(act_grad_types);
  auto *elewise_add_grad = pattern->NewNode(elewise_add_grad_repr())
                               ->assert_is_op("elementwise_add_grad");
  auto *bn_grad = pattern->NewNode(batch_norm_grad_repr())
                      ->assert_is_op("batch_norm_grad")
                      ->assert_op_attr<bool>("use_global_stats", false)
                      ->assert_op_attr<std::string>("data_layout", "NHWC");

  auto *act_out_var = pattern->NewNode(act_out_repr())
                          ->assert_is_ops_input(act_grad_types, "Out");
  auto *d_act_x_var =
      pattern->NewNode(d_act_x_repr())
          ->assert_is_ops_output(act_grad_types, GradVarName("X"))
          ->assert_has_n_outputs(1);  // d_act_x

  d_act_x_var->AsIntermediate()->assert_is_op_input("elementwise_add_grad");

  auto *d_elewise_add_in_var =
      pattern->NewNode(d_elewise_add_in_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("elementwise_add_grad")
          ->assert_var_dtype(proto::VarType::FP16);  // d_add_in_1
  auto *d_bn_out_var =
      pattern->NewNode(d_bn_out_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("elementwise_add_grad")
          ->assert_var_dtype(proto::VarType::FP16);  // d_add_in_2

  d_bn_out_var->assert_is_op_input("batch_norm_grad", GradVarName("Y"));

  auto *bn_x_var = pattern->NewNode(bn_x_repr())
                       ->assert_is_op_input("batch_norm_grad", "X")
                       ->assert_var_dtype(proto::VarType::FP16);
  auto *bn_scale_var = pattern->NewNode(bn_scale_repr())
                           ->assert_is_op_input("batch_norm_grad", "Scale");
  auto *bn_bias_var = pattern->NewNode(bn_bias_repr())
                          ->assert_is_op_input("batch_norm_grad", "Bias");
  auto *bn_saved_mean_var =
      pattern->NewNode(bn_saved_mean_repr())
          ->assert_is_op_input("batch_norm_grad", "SavedMean");
  auto *bn_saved_variance_var =
      pattern->NewNode(bn_saved_variance_repr())
          ->assert_is_op_input("batch_norm_grad", "SavedVariance");

  auto *bn_reserve_space =
      pattern->NewNode(bn_reserve_space_repr())
          ->assert_is_op_input("batch_norm_grad", "ReserveSpace");
  auto *d_bn_x_var =
      pattern->NewNode(d_bn_x_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("X"))
          ->assert_var_dtype(proto::VarType::FP16);
  auto *d_bn_scale_var =
      pattern->NewNode(d_bn_scale_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("Scale"));
  auto *d_bn_bias_var =
      pattern->NewNode(d_bn_bias_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("batch_norm_grad", GradVarName("Bias"));

  act_grad->LinksFrom({d_act_out_var, act_out_var}).LinksTo({d_act_x_var});

  elewise_add_grad->LinksFrom({d_act_x_var})
      .LinksTo({d_elewise_add_in_var, d_bn_out_var});

  bn_grad
      ->LinksFrom({bn_x_var, d_bn_out_var, bn_scale_var, bn_bias_var,
                   bn_saved_mean_var, bn_saved_variance_var, bn_reserve_space})
      .LinksTo({d_bn_x_var, d_bn_scale_var, d_bn_bias_var});

  return bn_grad;
}

C
chengduo 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
PDNode *patterns::ElewiseAddAct::operator()(
    paddle::framework::ir::PDNode *ele_x_var,
    std::unordered_set<std::string> act_types) {
  auto *ele_y_var = pattern->NewNode(ele_y_repr())
                        ->assert_is_op_input("elementwise_add", "Y");

  auto *ele_add =
      pattern->NewNode(ele_add_repr())->assert_is_op("elementwise_add");

  auto *ele_out_var = pattern->NewNode(elewise_add_out_repr())
                          ->assert_is_op_output("elementwise_add", "Out");

  ele_out_var->AsIntermediate()->assert_is_ops_input(act_types);

  auto *act = pattern->NewNode(act_repr())->assert_is_ops(act_types);

  auto *act_out_var =
      pattern->NewNode(act_out_repr())->assert_is_ops_output(act_types, "Out");

  ele_add->LinksFrom({ele_x_var, ele_y_var}).LinksTo({ele_out_var});
  act->LinksFrom({ele_out_var}).LinksTo({act_out_var});

  return act_out_var;
}

PDNode *patterns::ElewiseAddActInplaceGrad::operator()(
    paddle::framework::ir::PDNode *d_act_out_var,
    std::unordered_set<std::string> act_types) {
  // act_grad: in["Out", "Out@GRAD"], out["X@GRAD"]
  // ele_add_grad: in["Y", "Out@GRAD"], out["X@GRAD", "Y@GRAD"]
  auto *act_grad = pattern->NewNode(act_grad_repr())->assert_is_ops(act_types);

  auto *act_out_var =
      pattern->NewNode(act_out_repr())->assert_is_ops_input(act_types, "Out");

  auto *d_intermediate_var =
      pattern->NewNode(d_itermediate_out_repr())
          ->assert_is_ops_output(act_types, GradVarName("X"));

  act_grad->LinksFrom({d_act_out_var, act_out_var})
      .LinksTo({d_intermediate_var});

  auto *ele_y_var = pattern->NewNode(ele_y_repr())
                        ->assert_is_not_ctrl_var()
                        ->assert_is_op_input("elementwise_add_grad", "Y");

  auto *ele_add_grad = pattern->NewNode(ele_add_grad_repr())
                           ->assert_is_op("elementwise_add_grad");

  auto *d_ele_x_var =
      pattern->NewNode(d_ele_x_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("elementwise_add_grad", GradVarName("X"));

  auto *d_ele_y_var =
      pattern->NewNode(d_ele_y_repr())
          ->assert_is_not_ctrl_var()
          ->assert_is_op_output("elementwise_add_grad", GradVarName("Y"));

  ele_add_grad->LinksFrom({d_intermediate_var, ele_y_var})
      .LinksTo({d_ele_x_var, d_ele_y_var});

  return ele_add_grad;
}

1462
// conv_type: conv2d, conv3d, conv2d_transpose
M
Michal Gallus 已提交
1463
PDNode *patterns::ConvBias::operator()(
1464
    paddle::framework::ir::PDNode *conv_input, std::string conv_type) {
M
Michal Gallus 已提交
1465
  // Create Operators
1466 1467
  conv_input->assert_is_op_input(conv_type, "Input");
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op(conv_type);
M
Michal Gallus 已提交
1468 1469 1470 1471
  auto *eltiwse_op =
      pattern->NewNode(eltwise_repr())->assert_is_op("elementwise_add");
  // Create variables
  // Filter
Y
Yihua Xu 已提交
1472 1473 1474
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
1475
                              ->assert_is_op_input(conv_type, "Filter");
M
Michal Gallus 已提交
1476
  // intermediate variable, will be removed in the IR after fuse.
Y
Yihua Xu 已提交
1477 1478
  auto *conv_out_var = pattern->NewNode(conv_out_repr())
                           ->AsIntermediate()
1479
                           ->assert_is_only_output_of_op(conv_type)
Y
Yihua Xu 已提交
1480
                           ->assert_is_op_input("elementwise_add");
M
Michal Gallus 已提交
1481 1482 1483
  // Bias stored in elementwise_add
  auto *eltwise_bias_var = pattern->NewNode(eltwise_bias_repr())
                               ->AsInput()
M
Michal Gallus 已提交
1484
                               ->assert_is_persistable_var()
M
Michal Gallus 已提交
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495
                               ->assert_is_op_input("elementwise_add", "Y");
  // output
  auto *eltwise_out_var = pattern->NewNode(eltwise_out_repr())
                              ->AsOutput()
                              ->assert_is_op_output("elementwise_add");
  conv_op->LinksFrom({conv_input, conv_weight_var}).LinksTo({conv_out_var});
  eltiwse_op->LinksFrom({conv_out_var, eltwise_bias_var})
      .LinksTo({eltwise_out_var});
  return eltwise_out_var;
}

1496 1497 1498 1499
PDNode *patterns::Conv::operator()() {
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");

  auto input_var = pattern->NewNode(conv_input_repr())
1500
                       ->AsInput()
1501 1502 1503
                       ->assert_is_op_input("conv2d", "Input");

  auto filter_var = pattern->NewNode(conv_filter_repr())
1504
                        ->AsInput()
1505 1506 1507
                        ->assert_is_op_input("conv2d", "Filter");

  auto output_var = pattern->NewNode(conv_output_repr())
1508
                        ->AsOutput()
1509 1510
                        ->assert_is_op_output("conv2d", "Output");

1511 1512 1513 1514
  conv_op->LinksFrom({input_var, filter_var}).LinksTo({output_var});
  return output_var;
}

1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
PDNode *patterns::Transpose::operator()() {
  auto prev_op = pattern->NewNode(prev_op_repr())->assert_is_op();

  auto transpose_op =
      pattern->NewNode(transpose_op_repr())->assert_is_op("transpose2");

  auto transpose_in = pattern->NewNode(transpose_in_repr())
                          ->AsInput()
                          ->assert_is_op_input("transpose2");
  auto transpose_out = pattern->NewNode(transpose_out_repr())
                           ->AsOutput()
                           ->assert_is_op_output("transpose2", "Out");

  auto next_op = pattern->NewNode(next_op_repr())->assert_is_op();

  prev_op->LinksTo({transpose_in});
  transpose_op->LinksFrom({transpose_in}).LinksTo({transpose_out});
  next_op->LinksFrom({transpose_out});
  return transpose_out;
}

1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
PDNode *patterns::Reshape::operator()() {
  auto prev_op = pattern->NewNode(prev_op_repr())->assert_is_op();

  auto reshape_op =
      pattern->NewNode(reshape_op_repr())->assert_is_op("reshape2");

  auto reshape_in = pattern->NewNode(reshape_in_repr())
                        ->AsInput()
                        ->assert_is_op_input("reshape2", "X");
  auto reshape_out = pattern->NewNode(reshape_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("reshape2", "Out");

  auto next_op = pattern->NewNode(next_op_repr())->assert_is_op();

  prev_op->LinksTo({reshape_in});
  reshape_op->LinksFrom({reshape_in}).LinksTo({reshape_out});
  next_op->LinksFrom({reshape_out});
  return reshape_out;
}

1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
PDNode *patterns::Matmul::operator()() {
  auto prev_op_x = pattern->NewNode(prev_op_x_repr())->assert_is_op();
  auto prev_op_y = pattern->NewNode(prev_op_y_repr())->assert_is_op();

  auto matmul_op = pattern->NewNode(matmul_op_repr())->assert_is_op("matmul");
  auto matmul_in_x = pattern->NewNode(matmul_in_x_repr())
                         ->AsInput()
                         ->assert_is_op_input("matmul", "X");
  auto matmul_in_y = pattern->NewNode(matmul_in_y_repr())
                         ->AsInput()
                         ->assert_is_op_input("matmul", "Y");
  auto matmul_out = pattern->NewNode(matmul_out_repr())
                        ->AsOutput()
                        ->assert_is_op_output("matmul", "Out");

  prev_op_x->LinksTo({matmul_in_x});
  prev_op_y->LinksTo({matmul_in_y});
  matmul_op->LinksFrom({matmul_in_x, matmul_in_y}).LinksTo({matmul_out});
  return matmul_out;
}

1578 1579 1580
PDNode *patterns::ConvResidual::operator()(bool with_residual_data) {
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");

1581 1582 1583 1584 1585 1586 1587 1588 1589
  if (!with_residual_data) {
    conv_op->assert_more([&](Node *x) {
      auto node_names = x->Op()->InputNames();
      if (!HasInput(x, "ResidualData") ||
          x->Op()->Input("ResidualData").size() == 0)
        return true;
      return false;
    });
  }
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625

  auto input_var = pattern->NewNode(conv_input_repr())
                       ->AsInput()
                       ->assert_is_op_input("conv2d", "Input");

  auto filter_var = pattern->NewNode(conv_filter_repr())
                        ->AsInput()
                        ->assert_is_op_input("conv2d", "Filter");

  auto output_var = pattern->NewNode(conv_output_repr())
                        ->AsOutput()
                        ->assert_is_op_output("conv2d", "Output");

  std::vector<PDNode *> links_from{input_var, filter_var};

  if (with_residual_data) {
    auto res_conn_var = pattern->NewNode(conv_residual_data_repr())
                            ->AsInput()
                            ->assert_is_op_input("conv2d", "ResidualData");
    links_from.push_back(res_conn_var);
  }

  conv_op->LinksFrom(links_from).LinksTo({output_var});
  return output_var;
}

PDNode *patterns::Pool::operator()() {
  auto pool_op = pattern->NewNode(pool_op_repr())->assert_is_op("pool2d");

  auto input_var = pattern->NewNode(pool_input_repr())
                       ->AsInput()
                       ->assert_is_op_input("pool2d", "X");

  auto output_var = pattern->NewNode(pool_output_repr())
                        ->AsOutput()
                        ->assert_is_op_output("pool2d", "Out");
1626

1627
  pool_op->LinksFrom({input_var}).LinksTo({output_var});
1628 1629 1630
  return output_var;
}

1631
PDNode *patterns::ElementwiseAdd::operator()(PDNode *x_var, PDNode *y_var) {
1632 1633 1634
  auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
                                ->assert_is_op("elementwise_add");

1635 1636
  x_var->AsInput()->assert_is_op_input("elementwise_add", "X");
  y_var->AsInput()->assert_is_op_input("elementwise_add", "Y");
1637 1638 1639 1640
  auto out_var = pattern->NewNode(elementwise_add_out_repr())
                     ->AsOutput()
                     ->assert_is_op_output("elementwise_add", "Out");

1641
  elementwise_add_op->LinksFrom({x_var, y_var});
1642 1643 1644 1645
  elementwise_add_op->LinksTo({out_var});

  return out_var;
}
1646

1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
PDNode *patterns::Concat::operator()() {
  auto concat_op = pattern->NewNode(concat_op_repr())->assert_is_op("concat");

  auto output_var = pattern->NewNode(concat_out_repr())
                        ->AsOutput()
                        ->assert_is_op_output("concat", "Out");

  concat_op->LinksTo({output_var});
  return output_var;
}

1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697
PDNode *patterns::ConcatReLU::operator()() {
  auto concat_op = pattern->NewNode(concat_op_repr())->assert_is_op("concat");
  auto relu_op = pattern->NewNode(relu_op_repr())->assert_is_op("relu");

  auto concat_out =
      pattern->NewNode(concat_out_repr())->assert_is_op_output("concat", "Out");

  auto relu_out = pattern->NewNode(relu_out_repr())
                      ->AsOutput()
                      ->assert_is_op_output("relu", "Out");

  concat_op->LinksTo({concat_out});
  relu_op->LinksFrom({concat_out}).LinksTo({relu_out});

  return relu_out;
}

PDNode *patterns::ConvConcatReLU::operator()() {
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");
  auto concat_op = pattern->NewNode(concat_op_repr())->assert_is_op("concat");
  auto relu_op = pattern->NewNode(relu_op_repr())->assert_is_op("relu");

  auto conv_out = pattern->NewNode(conv_out_repr())
                      ->assert_is_op_output("conv2d", "Output");

  auto concat_out = pattern->NewNode(concat_out_repr())
                        ->assert_is_op_output("concat", "Out")
                        ->assert_is_op_input("relu", "X");

  auto relu_out = pattern->NewNode(relu_out_repr())
                      ->AsOutput()
                      ->assert_is_op_output("relu", "Out");

  conv_op->LinksTo({conv_out});
  concat_op->LinksFrom({conv_out}).LinksTo({concat_out});
  relu_op->LinksFrom({concat_out}).LinksTo({relu_out});

  return relu_out;
}

J
joanna.wozna.intel 已提交
1698 1699 1700 1701 1702 1703 1704 1705
PDNode *patterns::OpRequant::operator()() {
  auto any_op = pattern->NewNode(any_op_repr())
                    ->assert_is_op()
                    ->assert_more([&](Node *node) {
                      return node->Op()->HasAttr("Scale_out") ? true : false;
                    });
  auto requant_in = pattern->NewNode(requant_in_repr())
                        ->assert_is_op_input("requantize", "Input");
1706 1707 1708 1709 1710 1711
  auto requant_op =
      pattern->NewNode(requant_op_repr())->assert_is_op("requantize");
  auto requant_out = pattern->NewNode(requant_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("requantize", "Output");

J
joanna.wozna.intel 已提交
1712 1713
  any_op->LinksTo({requant_in});
  requant_op->LinksFrom({requant_in}).LinksTo({requant_out});
1714 1715 1716
  return requant_out;
}

1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
PDNode *patterns::RequantOp::operator()() {
  auto requant_in = pattern->NewNode(requant_in_repr())
                        ->assert_is_op_input("requantize", "Input");
  auto requant_op =
      pattern->NewNode(requant_op_repr())->assert_is_op("requantize");
  auto requant_out = pattern->NewNode(requant_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("requantize", "Output");
  auto any_op = pattern->NewNode(any_op_repr())
                    ->assert_is_op()
                    ->assert_more([&](Node *node) {
                      return (node->Op()->HasAttr("Scale_in") ||
                              node->Op()->HasAttr("Scale_x") ||
                              node->Op()->HasAttr("Scale_y"));
                    });

  requant_op->LinksFrom({requant_in}).LinksTo({requant_out});
  any_op->LinksFrom({requant_out});
  return any_op;
}

1738 1739 1740 1741 1742 1743 1744 1745 1746 1747
PDNode *patterns::OpDequant::operator()() {
  auto any_op = pattern->NewNode(any_op_repr())
                    ->assert_is_op()
                    ->assert_more([&](Node *node) {
                      return (node->Op()->Type() == "matmul" ||
                              node->Op()->Type() == "conv2d" ||
                              node->Op()->Type() == "fc");
                    });
  auto dequant_in = pattern->NewNode(dequant_in_repr())
                        ->assert_is_op_input("dequantize", "Input");
1748 1749 1750 1751 1752 1753
  auto dequant_op =
      pattern->NewNode(dequant_op_repr())->assert_is_op("dequantize");
  auto dequant_out = pattern->NewNode(dequant_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("dequantize", "Output");

1754 1755
  any_op->LinksTo({dequant_in});
  dequant_op->LinksFrom({dequant_in}).LinksTo({dequant_out});
1756 1757 1758
  return dequant_out;
}

1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777
PDNode *patterns::DequantScale::operator()() {
  // Create Operators
  auto dequant_op =
      pattern->NewNode(dequant_op_repr())->assert_is_op("dequantize");
  auto scale_op = pattern->NewNode(scale_op_repr())->assert_is_op("scale");

  auto dequant_out = pattern->NewNode(dequant_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("dequantize", "Output");
  auto scale_out = pattern->NewNode(scale_out_repr())
                       ->AsOutput()
                       ->assert_is_op_output("scale", "Out");

  dequant_op->LinksTo({dequant_out});
  scale_op->LinksFrom({dequant_out}).LinksTo({scale_out});

  return scale_out;
}

1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792
PDNode *patterns::ScaleMatmul::operator()() {
  auto scale_in = pattern->NewNode(scale_in_repr())
                      ->AsInput()
                      ->assert_is_op_input("scale", "X");
  auto scale_op = pattern->NewNode(scale_op_repr())->assert_is_op("scale");
  auto scale_out = pattern->NewNode(scale_out_repr())
                       ->AsOutput()
                       ->assert_is_op_output("scale", "Out");
  auto matmul_op = pattern->NewNode(matmul_op_repr())->assert_is_op("matmul");

  scale_op->LinksFrom({scale_in}).LinksTo({scale_out});
  matmul_op->LinksFrom({scale_out});
  return matmul_op;
}

1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817
PDNode *patterns::PriorBox::operator()() {
  auto prior_box_op =
      pattern->NewNode(prior_box_op_repr())->assert_is_op("prior_box");

  auto input_var = pattern->NewNode(prior_box_input_repr())
                       ->AsInput()
                       ->assert_is_op_input("prior_box", "Input");

  auto image_var = pattern->NewNode(prior_box_image_repr())
                       ->AsInput()
                       ->assert_is_op_input("prior_box", "Image");

  auto boxes_var = pattern->NewNode(prior_box_boxes_repr())
                       ->AsOutput()
                       ->assert_is_op_output("prior_box", "Boxes");

  auto variances_var = pattern->NewNode(prior_box_variances_repr())
                           ->AsOutput()
                           ->assert_is_op_output("prior_box", "Variances");

  prior_box_op->LinksFrom({input_var, image_var})
      .LinksTo({boxes_var, variances_var});
  return boxes_var;
}

H
hjchen2 已提交
1818
std::unordered_set<std::string> conv_act_set({"identity", "relu"});
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832

PDNode *patterns::ConvElementwiseaddAct::operator()(PDNode *conv_in) {
  conv_in->AsInput();
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");
  auto conv_out = pattern->NewNode(conv_out_repr())
                      ->assert_is_op_output("conv2d")
                      ->assert_is_op_input("elementwise_add", "X")
                      ->AsIntermediate();
  auto conv_filter = pattern->NewNode(conv_filter_repr())
                         ->assert_is_op_input("conv2d", "Filter")
                         ->AsInput();
  auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
                                ->assert_is_op("elementwise_add");
  auto elementwise_add_in_y = pattern->NewNode(elementwise_add_in_y_repr())
1833
                                  ->assert_is_persistable_var()
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
                                  ->assert_is_op_input("elementwise_add", "Y")
                                  ->AsInput();
  auto elementwise_add_out = pattern->NewNode(elementwise_add_out_repr())
                                 ->assert_is_op_output("elementwise_add")
                                 ->AsIntermediate();

  auto act_op = pattern->NewNode(act_op_repr())
                    ->assert_is_op()
                    ->assert_more([&](Node *node) {
                      auto op_type = node->Name();
                      return conv_act_set.count(op_type);
                    });

  auto act_out = pattern->NewNode(act_out_repr())
                     ->assert_is_var()
                     // is activation op's output.
                     ->assert_more([&](Node *node) {
                       for (auto *in_op : node->inputs) {
                         if (conv_act_set.count(in_op->Name())) {
                           return true;
                         }
                       }
                       return false;
                     })
                     ->AsOutput();

  conv_op->LinksFrom({conv_in, conv_filter});
  conv_out->LinksFrom({conv_op});
  elementwise_add_op->LinksFrom({conv_out, elementwise_add_in_y})
      .LinksTo({elementwise_add_out});
  act_op->LinksFrom({elementwise_add_out}).LinksTo({act_out});

  return act_out;
}

PDNode *patterns::ConvElementwiseadd2Act::operator()(PDNode *conv_in) {
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");
  auto conv_filter = pattern->NewNode(conv_filter_repr())
                         ->assert_is_op_input("conv2d", "Filter")
                         ->AsInput();
  auto conv_out = pattern->NewNode(conv_out_repr())
                      ->assert_is_op_output("conv2d")
                      ->assert_is_op_input("elementwise_add", "X")
                      ->AsIntermediate();
  auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
                                ->assert_is_op("elementwise_add");
  auto elementwise_add_in_y = pattern->NewNode(elementwise_add_in_y_repr())
1881
                                  ->assert_is_persistable_var()
1882 1883 1884 1885
                                  ->assert_is_op_input("elementwise_add", "Y")
                                  ->AsInput();
  auto elementwise_add_out = pattern->NewNode(elementwise_add_out_repr())
                                 ->assert_is_op_output("elementwise_add")
H
hjchen2 已提交
1886
                                 ->assert_is_op_input("elementwise_add", "Y")
1887 1888 1889 1890 1891
                                 ->AsIntermediate();

  auto elementwise_add_op_1 = pattern->NewNode(elementwise_add_op_1_repr())
                                  ->assert_is_op("elementwise_add");
  auto elementwise_add_in_y_1 = pattern->NewNode(elementwise_add_in_y_1_repr())
H
hjchen2 已提交
1892
                                    ->assert_is_op_input("elementwise_add", "X")
1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919
                                    ->AsInput();
  auto elementwise_add_out_1 = pattern->NewNode(elementwise_add_out_1_repr())
                                   ->assert_is_op_output("elementwise_add")
                                   ->AsIntermediate();

  auto act_op = pattern->NewNode(act_op_repr())
                    ->assert_is_op()
                    ->assert_more([&](Node *node) {
                      auto op_type = node->Name();
                      return conv_act_set.count(op_type);
                    });
  auto act_out = pattern->NewNode(act_out_repr())
                     ->assert_is_var()
                     // is activation op's output.
                     ->assert_more([&](Node *node) {
                       for (auto *in_op : node->inputs) {
                         if (conv_act_set.count(in_op->Name())) {
                           return true;
                         }
                       }
                       return false;
                     })
                     ->AsOutput();

  conv_op->LinksFrom({conv_in, conv_filter}).LinksTo({conv_out});
  elementwise_add_op->LinksFrom({conv_out, elementwise_add_in_y})
      .LinksTo({elementwise_add_out});
H
hjchen2 已提交
1920 1921
  elementwise_add_op_1->LinksFrom({elementwise_add_out, elementwise_add_in_y_1})
      .LinksTo({elementwise_add_out_1});
1922 1923 1924 1925
  act_op->LinksFrom({elementwise_add_out_1}).LinksTo({act_out});
  return act_out;
}

N
nhzlx 已提交
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
PDNode *patterns::ConvElementwiseadd::operator()(PDNode *conv_in) {
  conv_in->AsInput();
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");
  auto conv_out = pattern->NewNode(conv_out_repr())
                      ->assert_is_op_output("conv2d")
                      ->assert_is_op_input("elementwise_add", "X")
                      ->AsIntermediate();
  auto conv_filter = pattern->NewNode(conv_filter_repr())
                         ->assert_is_op_input("conv2d", "Filter")
                         ->AsInput();
  auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
                                ->assert_is_op("elementwise_add");
  auto elementwise_add_in_y = pattern->NewNode(elementwise_add_in_y_repr())
1939
                                  ->assert_is_persistable_var()
N
nhzlx 已提交
1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
                                  ->assert_is_op_input("elementwise_add", "Y")
                                  ->AsInput();
  auto elementwise_add_out = pattern->NewNode(elementwise_add_out_repr())
                                 ->assert_is_op_output("elementwise_add")
                                 ->AsOutput();

  conv_op->LinksFrom({conv_in, conv_filter});
  conv_out->LinksFrom({conv_op});
  elementwise_add_op->LinksFrom({conv_out, elementwise_add_in_y})
      .LinksTo({elementwise_add_out});

  return elementwise_add_out;
}

N
nhzlx 已提交
1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
PDNode *patterns::ConvAffineChannel::operator()(
    paddle::framework::ir::PDNode *conv_input, bool with_eltwise_add) {
  // Create Operators
  conv_input->assert_is_op_input("conv2d", "Input");
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op("conv2d");

  PDNode *eltwise_op = nullptr;
  if (with_eltwise_add) {
    eltwise_op =
        pattern->NewNode(eltwise_repr())->assert_is_op("elementwise_add");
  }

  auto *affine_channel_op =
      pattern->NewNode(affine_channel_repr())->assert_is_op("affine_channel");
  // Create variables
  // Conv Filter
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
                              ->assert_is_op_input("conv2d", "Filter");

  auto *conv_out_var = pattern->NewNode(conv_out_repr())
                           ->AsIntermediate()
                           ->assert_is_only_output_of_op("conv2d");

  PDNode *eltwise_y_in_var = nullptr;
  PDNode *eltwise_out_var = nullptr;
  if (with_eltwise_add) {
    // Conv output as Bias input
    conv_out_var->assert_is_op_input("elementwise_add", "X");
    // Bias
    eltwise_y_in_var = pattern->NewNode(eltwise_y_in_repr())
                           ->assert_is_op_input("elementwise_add", "Y")
                           ->AsInput();
    eltwise_out_var = pattern->NewNode(eltwise_out_repr())
                          ->AsIntermediate()
                          ->assert_is_only_output_of_op("elementwise_add");
  } else {
    // Conv output as AffineChannel input
    conv_out_var->assert_is_op_input("affine_channel", "X");
  }

  // AC Scale
  auto *ac_scale_var = pattern->NewNode(ac_scale_repr())
                           ->AsInput()
                           ->assert_is_persistable_var()
2000
                           ->assert_has_n_outputs(1)
N
nhzlx 已提交
2001 2002 2003 2004 2005
                           ->assert_is_op_input("affine_channel", "Scale");
  // AC Bias
  auto *ac_bias_var = pattern->NewNode(ac_bias_repr())
                          ->AsInput()
                          ->assert_is_persistable_var()
2006
                          ->assert_has_n_outputs(1)
N
nhzlx 已提交
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027
                          ->assert_is_op_input("affine_channel", "Bias");

  // AC output
  auto *ac_out_var = pattern->NewNode(ac_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("affine_channel");

  conv_op->LinksFrom({conv_input, conv_weight_var}).LinksTo({conv_out_var});

  if (with_eltwise_add) {
    eltwise_op->LinksFrom({conv_out_var, eltwise_y_in_var})
        .LinksTo({eltwise_out_var});
    affine_channel_op->LinksFrom({eltwise_out_var, ac_scale_var, ac_bias_var})
        .LinksTo({ac_out_var});
  } else {
    affine_channel_op->LinksFrom({conv_out_var, ac_scale_var, ac_bias_var})
        .LinksTo({ac_out_var});
  }
  return ac_out_var;
}

2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072
PDNode *patterns::DequantQuantAny::operator()() {
  auto *dequant_in = pattern->NewNode(dequant_in_repr())
                         ->AsInput()
                         ->assert_is_op_input("dequantize", "Input");

  auto *dequant_op =
      pattern->NewNode(dequant_op_repr())->assert_is_op("dequantize");

  auto *dequant_out = pattern->NewNode(dequant_out_repr())
                          ->AsOutput()
                          ->assert_is_op_output("dequantize", "Output");

  auto *quant_op = pattern->NewNode(quant_op_repr())
                       ->assert_is_op("quantize")
                       ->AsIntermediate();

  auto *quant_out = pattern->NewNode(quant_out_repr())
                        ->AsOutput()
                        ->assert_is_op_output("quantize");

  auto *next_op = pattern->NewNode(next_op_repr())->assert_is_op();

  dequant_op->LinksFrom({dequant_in}).LinksTo({dequant_out});
  quant_op->LinksFrom({dequant_out}).LinksTo({quant_out});
  next_op->LinksFrom({quant_out});

  return quant_out;
}

PDNode *patterns::DequantAny::operator()() {
  auto *dequant_op =
      pattern->NewNode(dequant_op_repr())->assert_is_op("dequantize");

  auto *dequant_out = pattern->NewNode(dequant_out_repr())
                          ->AsOutput()
                          ->assert_is_op_output("dequantize", "Output");

  auto *next_op = pattern->NewNode(next_op_repr())->assert_is_op();

  dequant_op->LinksTo({dequant_out});
  next_op->LinksFrom({dequant_out});

  return dequant_out;
}

2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
PDNode *patterns::MultipleQuantize::operator()() {
  auto *prev_out = pattern->NewNode(prev_out_repr())->AsOutput();

  // find nodes that are inputs to quantize operators
  prev_out->assert_more([&](Node *node) {
    int counter = std::count_if(
        node->outputs.begin(), node->outputs.end(), [&](Node const *iter) {
          return iter && iter->IsOp() && iter->Op()->Type() == "quantize";
        });
    return (counter > 1);
  });

  return prev_out;
}

2088 2089 2090
PDNode *patterns::QuantizePlacement::operator()(
    const std::unordered_set<std::string> &quantize_enabled_op_types) {
  std::unordered_set<std::string> supported_op_types =
2091 2092 2093
      std::unordered_set<std::string>(
          {"concat", "conv2d", "elementwise_add", "fc", "matmul", "pool2d",
           "prior_box", "relu", "reshape2", "transpose2", "fusion_gru"});
2094 2095 2096 2097 2098 2099 2100
  if (!quantize_enabled_op_types.empty()) {
    supported_op_types = quantize_enabled_op_types;
  }
  auto *op = pattern->NewNode(op_repr())->assert_is_ops(supported_op_types);
  return op;
}

2101 2102
PDNode *patterns::Bfloat16Placement::operator()(
    const std::unordered_set<std::string> &bfloat16_enabled_op_types) {
J
Jacek Czaja 已提交
2103
  std::unordered_set<std::string> supported_op_types =
2104 2105 2106 2107
      std::unordered_set<std::string>(
          {"concat", "conv2d", "elementwise_add", "elementwise_mul", "fc",
           "fusion_gru", "gelu", "layer_norm", "matmul", "reshape2", "softmax",
           "sum", "transpose2"});
2108 2109 2110 2111
  if (!bfloat16_enabled_op_types.empty()) {
    supported_op_types = bfloat16_enabled_op_types;
  }
  auto *op = pattern->NewNode(op_repr())->assert_is_ops(supported_op_types);
2112 2113 2114 2115
  op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<bool>("use_mkldnn") ||
           node->Op()->Type() == "reshape2";
  });
2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
  return op;
}

PDNode *patterns::OrphanedBfloat16::operator()() {
  auto *prev_op = pattern->NewNode(prev_op_repr())->assert_is_op();
  prev_op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") ==
           "float32";
  });
  auto *prev_out = pattern->NewNode(prev_out_repr())->AsOutput();

  auto *op = pattern->NewNode(op_repr())->assert_is_op();
  op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") ==
           "bfloat16";
  });
  auto *op_out = pattern->NewNode(op_out_repr())->AsOutput();

  auto *next_op = pattern->NewNode(next_op_repr())->assert_is_op();
  next_op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") ==
           "float32";
  });

  prev_op->LinksTo({prev_out});
  op->LinksFrom({prev_out}).LinksTo({op_out});
  next_op->LinksFrom({op_out});
  return next_op;
}

PDNode *patterns::LastBfloat16Ops::operator()() {
  auto *op = pattern->NewNode(op_repr())->assert_is_op();
  op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") ==
           "bfloat16";
  });
  auto *op_out = pattern->NewNode(op_out_repr())->AsOutput();

  auto *next_op = pattern->NewNode(next_op_repr())->assert_is_op();
  next_op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") !=
           "bfloat16";
  });

  op->LinksTo({op_out});
  next_op->LinksFrom({op_out});
  return next_op;
}

PDNode *patterns::FirstBfloat16Ops::operator()() {
  auto *prev_op = pattern->NewNode(prev_op_repr())->assert_is_op();
  prev_op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") !=
           "bfloat16";
  });
  auto *op_in = pattern->NewNode(op_in_repr())->AsOutput();

  auto *op = pattern->NewNode(op_repr())->assert_is_op();
  op->assert_more([&](Node *node) {
    return node->Op()->GetAttrIfExists<std::string>("mkldnn_data_type") ==
           "bfloat16";
  });

  prev_op->LinksTo({op_in});
  op->LinksFrom({op_in});
  return op;
}

2184
PDNode *patterns::MKLDNNInPlace::operator()() {
2185
  const std::unordered_set<std::string> &supported_op_types = {
2186 2187 2188 2189 2190 2191 2192 2193 2194 2195
      "abs",
      "elementwise_mul",
      "elementwise_add",
      "gelu",
      "leaky_relu",
      "relu",
      "softmax",
      "sqrt",
      "swish",
      "tanh"};
2196 2197 2198

  auto possible_inplace_op = pattern->NewNode(inplace_to_be_op_repr())
                                 ->assert_is_ops(supported_op_types);
2199 2200

  auto input = pattern->NewNode(inplace_to_be_op_in_repr())
2201
                   ->assert_is_ops_input(supported_op_types)
2202 2203
                   ->AsInput();
  auto output = pattern->NewNode(inplace_to_be_op_out_repr())
2204
                    ->assert_is_ops_output(supported_op_types)
2205
                    ->AsOutput();
2206 2207

  auto next_op = pattern->NewNode(next_op_repr())->assert_is_op();
2208
  auto next_output = pattern->NewNode(next_op_out_repr())->AsOutput();
2209 2210 2211 2212

  // Check if op is MKL-DNN enabled
  possible_inplace_op->assert_op_attr("use_mkldnn", true);

2213
  // linked structure
2214 2215 2216
  possible_inplace_op->LinksTo({output});
  possible_inplace_op->LinksFrom({input});
  next_op->LinksFrom({output});
2217
  next_op->LinksTo({next_output});
2218 2219 2220 2221

  return possible_inplace_op;
}

2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
// a -> transpose_op(1) -> transpose_out_a -> flatten_op(1) -> flatten_out_a
// b -> transpose_op(2) -> transpose_out_b -> flatten_op(2) -> flatten_out_b
// ...
// z -> transpose_op(n) -> transpose_out_z -> flatten_op(n) -> flatten_out_z
// flatten_out_a -> concat_op  flatten_out_b -> concat_op ... flatten_out_z ->
// concat_op
PDNode *patterns::TransposeFlattenConcat::operator()(
    std::vector<PDNode *> conv_in, int times) {
  // The times represents the repeat times of the
  // {trans, trans_out, flatten, flatten_out}
  const int kNumFields = 4;
  const int kTransOutOffset = 1;
  const int kFlattenOffset = 2;
  const int kFlattenOutOffset = 3;

  std::vector<PDNode *> nodes;

  for (int i = 0; i < times; i++) {
    nodes.push_back(
        pattern->NewNode(GetNodeName("transpose" + std::to_string(i)))
            ->assert_is_op("transpose2"));
    nodes.push_back(
        pattern->NewNode(GetNodeName("transpose_out" + std::to_string(i)))
            ->assert_is_op_output("transpose2")
            ->assert_is_op_input("flatten2", "X")
            ->AsIntermediate());
    nodes.push_back(pattern->NewNode(GetNodeName("flatten" + std::to_string(i)))
                        ->assert_is_op("flatten2"));

    nodes.push_back(
        pattern->NewNode(GetNodeName("flatten_out" + std::to_string(i)))
            ->assert_is_op_output("flatten2")
            ->assert_is_op_nth_input("concat", "X", i)
            ->AsIntermediate());
  }

  auto concat_op = pattern->NewNode(GetNodeName("concat"))
                       ->assert_is_op("concat")
                       ->assert_op_has_n_inputs("concat", times);
  auto concat_out = pattern->NewNode(GetNodeName("concat_out"))
                        ->assert_is_op_output("concat")
                        ->AsOutput();

  std::vector<PDNode *> flatten_outs;
  for (int i = 0; i < times; i++) {
    conv_in[i]->AsInput();
    // trans
    nodes[i * kNumFields]->LinksFrom({conv_in[i]});
    // trans_out
    nodes[i * kNumFields + kTransOutOffset]->LinksFrom({nodes[i * kNumFields]});
    // flatten
    nodes[i * kNumFields + kFlattenOffset]->LinksFrom(
        {nodes[i * kNumFields + kTransOutOffset]});
    // flatten_out
    nodes[i * kNumFields + kFlattenOutOffset]->LinksFrom(
        {nodes[i * kNumFields + kFlattenOffset]});
    flatten_outs.push_back(nodes[i * kNumFields + kFlattenOutOffset]);
  }

  concat_op->LinksFrom(flatten_outs).LinksTo({concat_out});
  return concat_out;
}

2285 2286 2287
void patterns::DeleteQuantOpFuse::operator()(PDNode *input_act_node,
                                             const std::string &quant_type) {
  auto *input_scale_node = pattern->NewNode(GetNodeName("input_scale_node"))
2288 2289
                               ->assert_is_op_input(quant_type, "InScale")
                               ->AsInput();
2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321
  auto *quant_node =
      pattern->NewNode(GetNodeName("quant_node"))->assert_is_op(quant_type);
  auto *output_scale_node = pattern->NewNode(GetNodeName("output_scale_node"))
                                ->assert_is_op_output(quant_type, "OutScale")
                                ->AsOutput();
  auto *output_act_node = pattern->NewNode(GetNodeName("output_act_node"))
                              ->assert_is_op_output(quant_type, "Out")
                              ->AsOutput();
  quant_node->LinksFrom({input_scale_node, input_act_node});
  output_scale_node->LinksFrom({quant_node});
  output_act_node->LinksFrom({quant_node});
}

void patterns::DequantOpFuse::operator()(PDNode *quantized_op_input,
                                         const std::string &quantized_op_type,
                                         const std::string &dequant_type,
                                         const std::string &weight_name) {
  auto *quantized_op_weight =
      pattern->NewNode(GetNodeName("quantized_op_weight"))
          ->assert_is_op_input(quantized_op_type, weight_name)
          ->AsInput();
  auto *quantized_op = pattern->NewNode(GetNodeName("quantized_op"))
                           ->assert_is_op(quantized_op_type);
  auto *quantized_op_out = pattern->NewNode(GetNodeName("quantized_op_out"))
                               ->assert_is_op_output(quantized_op_type)
                               ->assert_is_op_input(dequant_type, "X");
  auto *dequant_op =
      pattern->NewNode(GetNodeName("dequant_op"))->assert_is_op(dequant_type);
  auto *dequant_op_out = pattern->NewNode(GetNodeName("dequant_op_out"))
                             ->assert_is_op_output(dequant_type, "Out")
                             ->AsOutput();
  PDNode *dequant_channel_scale = nullptr;
2322
  if (dequant_type == "fake_channel_wise_dequantize_max_abs") {
2323 2324 2325 2326
    dequant_channel_scale =
        pattern->NewNode(GetNodeName("dequant_channel_scale"))
            ->assert_is_op_nth_input(dequant_type, "Scales", 0)
            ->AsInput();
N
nhzlx 已提交
2327
  }
2328 2329
  quantized_op->LinksFrom({quantized_op_input, quantized_op_weight});
  quantized_op_out->LinksFrom({quantized_op});
N
nhzlx 已提交
2330

2331 2332 2333 2334
  if (dequant_type == "fake_channel_wise_dequantize_max_abs") {
    dequant_op->LinksFrom({quantized_op_out, dequant_channel_scale});
  } else {
    dequant_op->LinksFrom({quantized_op_out});
N
nhzlx 已提交
2335
  }
2336
  dequant_op_out->LinksFrom({dequant_op});
N
nhzlx 已提交
2337 2338
}

2339 2340 2341
void patterns::ShuffleChannelPattern::operator()(PDNode *reshape1_in) {
  auto reshape1_op =
      pattern->NewNode(reshape1_op_repr())->assert_is_op("reshape2");
2342
  reshape1_op->assert_more([&](Node *x) {
2343 2344
    return BOOST_GET_CONST(std::vector<int>, x->Op()->GetAttr("shape"))
               .size() == 5;
2345
  });
2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373

  auto reshape1_out = pattern->NewNode(reshape1_out_repr())
                          ->assert_is_op_output("reshape2", "Out")
                          ->assert_is_op_input("transpose2")
                          ->AsIntermediate();

  auto transpose_op =
      pattern->NewNode(transpose_op_repr())->assert_is_op("transpose2");

  auto transpose_out = pattern->NewNode(transpose_out_repr())
                           ->assert_is_op_output("transpose2", "Out")
                           ->assert_is_op_input("reshape2")
                           ->AsIntermediate();

  auto reshape2_op =
      pattern->NewNode(reshape2_op_repr())->assert_is_op("reshape2");
  auto reshape2_out = pattern->NewNode(reshape2_out_repr())
                          ->assert_is_op_output("reshape2", "Out")
                          ->AsOutput();

  reshape1_op->LinksFrom({reshape1_in});
  reshape1_out->LinksFrom({reshape1_op});
  transpose_op->LinksFrom({reshape1_out});
  transpose_out->LinksFrom({transpose_op});
  reshape2_op->LinksFrom({transpose_out});
  reshape2_out->LinksFrom({reshape2_op});
}

2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
void patterns::DeleteQuantDequantOpPattern::operator()() {
  auto any_op_out =
      pattern->NewNode(any_op_out_repr())
          ->assert_is_op_input(
              "fake_quantize_dequantize_moving_average_abs_max", "X")
          ->AsInput();

  auto quant_dequant_op_inscale =
      pattern->NewNode(quant_dequant_op_inscale_repr())
          ->assert_is_op_input(
              "fake_quantize_dequantize_moving_average_abs_max", "InScale")
          ->AsInput();
  auto quant_dequant_op =
      pattern->NewNode(quant_dequant_op_repr())
          ->assert_is_op("fake_quantize_dequantize_moving_average_abs_max");

  auto quant_dequant_out =
      pattern->NewNode(quant_dequant_op_out_repr())
          ->assert_is_op_output(
              "fake_quantize_dequantize_moving_average_abs_max", "Out")
          ->AsIntermediate();

  auto quant_dequant_op_outscale =
      pattern->NewNode(quant_dequant_op_outscale_repr())
          ->assert_is_op_output(
              "fake_quantize_dequantize_moving_average_abs_max", "OutScale")
          ->AsOutput();
  auto any_op2 = pattern->NewNode(any_op2_repr())->assert_is_op()->AsOutput();

  quant_dequant_op->LinksFrom({any_op_out, quant_dequant_op_inscale});
  quant_dequant_op_outscale->LinksFrom({quant_dequant_op});
  quant_dequant_out->LinksFrom({quant_dequant_op});
  any_op2->LinksFrom({quant_dequant_out});
}

2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459
PDNode *patterns::ReshapeTransposeMatmulPattern::operator()(
    bool with_reshape_xshape, bool with_transpose_xshape) {
  auto reshape_op =
      pattern->NewNode(reshape_op_repr())->assert_is_op("reshape2");
  auto transpose_op =
      pattern->NewNode(transpose_op_repr())->assert_is_op("transpose2");
  auto matmul_op = pattern->NewNode(matmul_op_repr())->assert_is_op("matmul");

  auto reshape_in = pattern->NewNode(reshape_in_repr())
                        ->AsInput()
                        ->assert_is_op_input("reshape2", "X");

  auto reshape_out = pattern->NewNode(reshape_out_repr())
                         ->AsIntermediate()
                         ->assert_is_op_input("transpose2", "X")
                         ->assert_is_op_output("reshape2", "Out");
  if (!with_reshape_xshape)
    reshape_out->assert_is_only_output_of_op("reshape2");

  auto reshape_xshape = with_reshape_xshape
                            ? pattern->NewNode(reshape_xshape_repr())
                                  ->AsIntermediate()
                                  ->assert_is_op_output("reshape2", "XShape")
                            : nullptr;

  auto transpose_out = pattern->NewNode(transpose_out_repr())
                           ->AsIntermediate()
                           ->assert_is_op_input("matmul")
                           ->assert_is_op_output("transpose2", "Out");
  if (!with_transpose_xshape)
    transpose_out->assert_is_only_output_of_op("transpose2");

  auto transpose_xshape =
      with_transpose_xshape
          ? pattern->NewNode(transpose_xshape_repr())
                ->AsIntermediate()
                ->assert_is_op_output("transpose2", "XShape")
          : nullptr;

  auto matmul_out = pattern->NewNode(matmul_out_repr())
                        ->AsOutput()
                        ->assert_is_op_output("matmul", "Out");

  reshape_op->LinksFrom({reshape_in}).LinksTo({reshape_out});
  if (with_reshape_xshape) reshape_op->LinksTo({reshape_xshape});
  transpose_op->LinksFrom({reshape_out}).LinksTo({transpose_out});
  if (with_transpose_xshape) transpose_op->LinksTo({transpose_xshape});
  matmul_op->LinksFrom({transpose_out}).LinksTo({matmul_out});
  return matmul_out;
}

2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496
PDNode *patterns::MatmulTransposeReshapePattern::operator()() {
  auto reshape_op =
      pattern->NewNode(reshape_op_repr())->assert_is_op("reshape2");
  auto transpose_op =
      pattern->NewNode(transpose_op_repr())->assert_is_op("transpose2");
  auto matmul_op = pattern->NewNode(matmul_op_repr())->assert_is_op("matmul");

  auto matmul_out = pattern->NewNode(matmul_out_repr())
                        ->AsInput()
                        ->assert_is_op_output("matmul", "Out")
                        ->assert_is_op_input("transpose2", "X");

  auto transpose_out = pattern->NewNode(transpose_out_repr())
                           ->AsIntermediate()
                           ->assert_is_op_output("transpose2", "Out")
                           ->assert_is_op_input("reshape2", "X");

  auto transpose_out_xshape = pattern->NewNode(transpose_out_xshape_repr())
                                  ->AsIntermediate()
                                  ->assert_is_op_output("transpose2", "XShape");

  auto reshape_out = pattern->NewNode(reshape_out_repr())
                         ->AsOutput()
                         ->assert_is_op_output("reshape2");

  auto reshape_out_xshape = pattern->NewNode(reshape_out_xshape_repr())
                                ->AsIntermediate()
                                ->assert_is_op_output("reshape2", "XShape");

  matmul_op->LinksTo({matmul_out});
  transpose_op->LinksTo({transpose_out_xshape});
  reshape_op->LinksTo({reshape_out_xshape});
  transpose_op->LinksFrom({matmul_out}).LinksTo({transpose_out});
  reshape_op->LinksFrom({transpose_out}).LinksTo({reshape_out});
  return reshape_out;
}

2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513
PDNode *patterns::FusionGru::operator()() {
  auto op = pattern->NewNode(op_repr())->assert_is_op("fusion_gru");
  auto x = pattern->NewNode(x_repr())->AsInput()->assert_is_op_input(
      "fusion_gru", "X");
  auto weight_h = pattern->NewNode(weight_h_repr())
                      ->AsInput()
                      ->assert_is_op_input("fusion_gru", "WeightH");
  auto weight_x = pattern->NewNode(weight_x_repr())
                      ->AsInput()
                      ->assert_is_op_input("fusion_gru", "WeightX");
  auto out = pattern->NewNode(out_repr())
                 ->AsOutput()
                 ->assert_is_op_output("fusion_gru", "Hidden");
  op->LinksFrom({x, weight_h, weight_x}).LinksTo({out});
  return out;
}

2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566
PDNode *patterns::MultiGruSeq::operator()() {
  auto x = pattern->NewNode(x_repr())->AsInput()->assert_is_op_input(
      "multi_gru", "X");
  auto gru1 = pattern->NewNode(gru1_repr())->assert_is_op("multi_gru");
  auto wx11 = pattern->NewNode(wx11_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightX", 0);
  auto wx12 = pattern->NewNode(wx12_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightX", 1);
  auto wh11 = pattern->NewNode(wh11_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightH", 0);
  auto wh12 = pattern->NewNode(wh12_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightH", 1);
  auto b11 = pattern->NewNode(b11_repr())
                 ->AsInput()
                 ->assert_is_op_nth_input("multi_gru", "Bias", 0);
  auto b12 = pattern->NewNode(b12_repr())
                 ->AsInput()
                 ->assert_is_op_nth_input("multi_gru", "Bias", 1);
  auto h1 = pattern->NewNode(h1_repr())
                ->AsOutput()
                ->assert_is_op_output("multi_gru", "Hidden")
                ->assert_is_op_input("multi_gru", "X")
                ->AsIntermediate();
  auto gru2 = pattern->NewNode(gru2_repr())->assert_is_op("multi_gru");
  auto wx21 = pattern->NewNode(wx21_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightX", 0);
  auto wx22 = pattern->NewNode(wx22_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightX", 1);
  auto wh21 = pattern->NewNode(wh21_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightH", 0);
  auto wh22 = pattern->NewNode(wh22_repr())
                  ->AsInput()
                  ->assert_is_op_nth_input("multi_gru", "WeightH", 1);
  auto b21 = pattern->NewNode(b21_repr())
                 ->AsInput()
                 ->assert_is_op_nth_input("multi_gru", "Bias", 0);
  auto b22 = pattern->NewNode(b22_repr())
                 ->AsInput()
                 ->assert_is_op_nth_input("multi_gru", "Bias", 1);
  auto h2 = pattern->NewNode(h2_repr())->AsOutput()->assert_is_op_output(
      "multi_gru", "Hidden");
  gru1->LinksFrom({x, wx11, wx12, wh11, wh12, b11, b12}).LinksTo({h1});
  gru2->LinksFrom({h1, wx21, wx22, wh21, wh22, b21, b22}).LinksTo({h2});
  return h2;
}

2567 2568 2569
}  // namespace ir
}  // namespace framework
}  // namespace paddle