graph_pattern_detector.cc 44.6 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 18 19 20
#include <string>
#include <vector>

#include "paddle/fluid/framework/ir/graph_helper.h"
21
#include "paddle/fluid/framework/ir/graph_pattern_detector.h"
22
#include "paddle/fluid/framework/ir/graph_traits.h"
23
#include "paddle/fluid/framework/ir/graph_viz_pass.h"
C
chengduo 已提交
24
#include "paddle/fluid/framework/operator.h"
25
#include "paddle/fluid/platform/enforce.h"
Y
Yan Chunwei 已提交
26
#include "paddle/fluid/string/pretty_log.h"
Y
Yan Chunwei 已提交
27
#include "paddle/fluid/string/printf.h"
28

29 30 31 32
namespace paddle {
namespace framework {
namespace ir {

Y
Yan Chunwei 已提交
33 34 35 36
using string::PrettyLogEndl;
using string::PrettyLog;
using string::Style;

37 38
size_t PDPattern::id_ = 0UL;

C
chengduo 已提交
39
PDNode *PDPattern::NewNode(const std::string &name) {
Y
Yan Chunwei 已提交
40 41 42 43 44 45 46
  if (!name.empty()) {
    PADDLE_ENFORCE_EQ(node_map_.count(name), 0,
                      "PDNode's name should be unique, get duplicate [%s]",
                      name);
  }

  nodes_.emplace_back(new PDNode(this, name));
C
chengduo 已提交
47
  auto *cur = nodes_.back().get();
Y
Yan Chunwei 已提交
48 49 50 51
  node_map_[name] = cur;
  return cur;
}

C
chengduo 已提交
52
PDNode *PDPattern::NewNode(PDNode::teller_t &&teller, const std::string &name) {
53 54 55 56 57 58
  if (!name.empty()) {
    PADDLE_ENFORCE_EQ(node_map_.count(name), 0,
                      "PDNode's name should be unique, get duplicate [%s]",
                      name);
  }

59
  nodes_.emplace_back(new PDNode(std::move(teller), this, name));
C
chengduo 已提交
60
  auto *cur = nodes_.back().get();
61
  node_map_[name] = cur;
62 63 64
  return cur;
}

C
chengduo 已提交
65
PDNode *PDPattern::RetrieveNode(const std::string &id) const {
66 67 68 69 70 71 72 73
  auto it = node_map_.find(id);
  if (it == node_map_.end()) {
    return nullptr;
  }

  return it->second;
}

C
chengduo 已提交
74
void PDPattern::AddEdge(PDNode *a, PDNode *b) {
75 76 77 78 79 80
  PADDLE_ENFORCE(a);
  PADDLE_ENFORCE(b);
  PADDLE_ENFORCE(a != b, "can't connect to the same nodes.");
  edges_.emplace_back(a, b);
}

C
chengduo 已提交
81
void GraphPatternDetector::operator()(Graph *graph,
82
                                      GraphPatternDetector::handle_t handler) {
83 84 85 86
  if (!MarkPDNodesInGraph(*graph)) {
    return;
  }

87 88 89
  auto subgraphs = DetectPatterns();
  UniquePatterns(&subgraphs);
  RemoveOverlappedMatch(&subgraphs);
Y
Yan Chunwei 已提交
90
  ValidateByNodeRole(&subgraphs);
91

Y
Yan Chunwei 已提交
92
  if (subgraphs.empty()) return;
Y
Yan Chunwei 已提交
93
  PrettyLogEndl(Style::detail(), "---  detect %d subgraphs", subgraphs.size());
94
  int id = 0;
C
chengduo 已提交
95
  for (auto &g : subgraphs) {
M
minqiyang 已提交
96
    VLOG(3) << "optimizing #" << id++ << " subgraph";
97 98 99 100
    handler(g, graph);
  }
}

C
chengduo 已提交
101
bool GraphPatternDetector::MarkPDNodesInGraph(const ir::Graph &graph) {
M
minqiyang 已提交
102
  VLOG(3) << "mark pdnodes in graph";
103 104
  if (graph.Nodes().empty()) return false;

C
chengduo 已提交
105 106
  for (auto &node : GraphTraits::DFS(graph)) {
    for (const auto &pdnode : pattern_.nodes()) {
107
      if (pdnode->Tell(&node)) {
108
        VLOG(4) << "Node " << node.Name() << " marked as " << pdnode->name();
109 110 111 112
        pdnodes2nodes_[pdnode.get()].insert(&node);
      }
    }
  }
Y
Yan Chunwei 已提交
113
  // Check to early stop if some PDNode can't find matched Node.
C
chengduo 已提交
114
  for (auto &pdnode : pattern_.nodes()) {
Y
Yan Chunwei 已提交
115
    if (!pdnodes2nodes_.count(pdnode.get())) {
M
minqiyang 已提交
116
      VLOG(4) << pdnode->name() << " can't find matched Node, early stop";
Y
Yan Chunwei 已提交
117
      // return false;
Y
Yan Chunwei 已提交
118 119
    }
  }
C
chengduo 已提交
120 121 122
  for (auto &item : pdnodes2nodes_) {
    for (auto &n : item.second) {
      GetMarkedNodes(const_cast<Graph *>(&graph)).insert(n);
Y
Yan Chunwei 已提交
123 124
    }
  }
M
minqiyang 已提交
125
  VLOG(3) << pdnodes2nodes_.size() << " nodes marked";
126

127 128 129
  return !pdnodes2nodes_.empty();
}

Y
Yan Chunwei 已提交
130 131 132
// The intermediate Nodes can only link to the nodes inside the pattern, or this
// subgraph will be droped.
void GraphPatternDetector::ValidateByNodeRole(
C
chengduo 已提交
133
    std::vector<GraphPatternDetector::subgraph_t> *subgraphs) {
Y
Yan Chunwei 已提交
134 135 136 137 138
  std::vector<GraphPatternDetector::subgraph_t> result;

  subgraphs->erase(
      std::remove_if(
          subgraphs->begin(), subgraphs->end(),
C
chengduo 已提交
139
          [](const GraphPatternDetector::subgraph_t &subgraph) -> bool {
Y
Yan Chunwei 已提交
140
            // Collect the inputs and outputs.
C
chengduo 已提交
141 142
            std::unordered_set<Node *> ios;
            for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
143 144 145 146
              if (!item.first->IsIntermediate()) {
                ios.insert(item.second);
              }
            }
C
chengduo 已提交
147
            for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
148
              if (item.first->IsIntermediate()) {
C
chengduo 已提交
149
                for (auto *x : item.second->inputs) {
Y
Yan Chunwei 已提交
150 151 152 153
                  if (!ios.count(x)) {
                    return true;
                  }
                }
C
chengduo 已提交
154
                for (auto *x : item.second->outputs) {
Y
Yan Chunwei 已提交
155 156 157 158 159 160 161 162 163 164 165
                  if (!ios.count(x)) {
                    return true;
                  }
                }
              }
            }
            return false;
          }),
      subgraphs->end());
}

166
struct HitGroup {
C
chengduo 已提交
167
  std::unordered_map<PDNode *, Node *> roles;
168

C
chengduo 已提交
169
  bool Match(Node *node, PDNode *pat) {
170
    if (nodes_.count(node)) {
T
Tao Luo 已提交
171 172 173 174 175
      if (roles.count(pat) && roles[pat] == node) return true;
      return false;
    } else {
      if (roles.count(pat) && roles[pat] != node) return false;
      return true;
176
    }
177 178
  }

C
chengduo 已提交
179
  void Register(Node *node, PDNode *pat) {
180 181 182 183 184
    roles[pat] = node;
    nodes_.insert(node);
  }

 private:
C
chengduo 已提交
185
  std::unordered_set<Node *> nodes_;
186 187 188
};

// Tell whether Node a links to b.
C
chengduo 已提交
189 190
bool IsNodesLink(Node *a, Node *b) {
  for (auto *node : a->outputs) {
191 192 193 194 195 196 197
    if (b == node) {
      return true;
    }
  }
  return false;
}

198 199
std::vector<GraphPatternDetector::subgraph_t>
GraphPatternDetector::DetectPatterns() {
200
  // Init empty subgraphs.
201
  std::vector<GraphPatternDetector::subgraph_t> result;
202
  std::vector<HitGroup> init_groups;
203
  std::array<std::vector<HitGroup>, 2> bi_records;
C
chengduo 已提交
204
  auto *first_pnode = pattern_.edges().empty() ? pattern().nodes().front().get()
205
                                               : pattern_.edges().front().first;
206
  if (!pdnodes2nodes_.count(first_pnode)) return result;
C
chengduo 已提交
207
  for (auto *node : pdnodes2nodes_[first_pnode]) {
208 209 210 211 212 213 214 215 216 217
    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
218
  for (const auto &edge : pattern_.edges()) {
M
minqiyang 已提交
219
    VLOG(4) << "check " << edge.first->name() << " -> " << edge.second->name();
Y
Yan Chunwei 已提交
220
    // TODO(Superjomn) Fix bug here, the groups might be duplicate here.
221 222
    // 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
223 224
    auto &pre_groups = bi_records[step % 2];
    auto &cur_groups = bi_records[1 - (step++ % 2)];
225
    cur_groups.clear();
226
    if (pre_groups.empty()) break;
227
    // source -> target
C
chengduo 已提交
228 229
    for (Node *source : pdnodes2nodes_[edge.first]) {
      for (Node *target : pdnodes2nodes_[edge.second]) {
M
minqiyang 已提交
230
        VLOG(8) << "check " << source->id() << " -- " << target->id();
231
        // TODO(Superjomn) add some prune strategies.
C
chengduo 已提交
232
        for (const auto &group : pre_groups) {
T
Tao Luo 已提交
233 234 235 236 237 238
          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);
239 240 241 242 243 244 245 246
              new_group.Register(target, edge.second);
              cur_groups.push_back(new_group);
              // TODO(Superjomn) need to unique
            }
          }
        }
      }
    }
M
minqiyang 已提交
247
    VLOG(3) << "step " << step << " get records: " << cur_groups.size();
C
chengduo 已提交
248 249
    for (auto &group : cur_groups) {
      for (auto &item : group.roles) {
M
minqiyang 已提交
250
        VLOG(4) << "node " << item.second->id() << " as " << item.first->name();
Y
Yan Chunwei 已提交
251
      }
M
minqiyang 已提交
252
      VLOG(4) << "=========================================================";
Y
Yan Chunwei 已提交
253
    }
254 255
  }

C
chengduo 已提交
256
  for (auto &group : bi_records[step % 2]) {
257
    GraphPatternDetector::subgraph_t subgraph;
C
chengduo 已提交
258
    for (auto &role : group.roles) {
259 260 261 262 263 264 265
      subgraph.emplace(role.first, role.second);
    }
    result.emplace_back(subgraph);
  }
  return result;
}

Y
Yan Chunwei 已提交
266 267
struct GraphItemLessThan {
  bool operator()(const std::pair<PDNode *, Node *> &a,
Y
Yan Chunwei 已提交
268
                  const std::pair<PDNode *, Node *> &b) {
Y
Yan Chunwei 已提交
269 270 271 272 273
    if (a.first != b.first) {
      return a.first < b.first;
    } else {
      return a.second < b.second;
    }
Y
Yan Chunwei 已提交
274
  }
Y
Yan Chunwei 已提交
275
};
Y
Yan Chunwei 已提交
276

277 278
// TODO(Superjomn) enhance the function as it marks unique unique as duplicates
// see https://github.com/PaddlePaddle/Paddle/issues/13550
279
void GraphPatternDetector::UniquePatterns(
C
chengduo 已提交
280
    std::vector<GraphPatternDetector::subgraph_t> *subgraphs) {
281
  if (subgraphs->empty()) return;
282
  std::vector<GraphPatternDetector::subgraph_t> result;
283 284

  std::unordered_set<size_t> set;
Y
Yan Chunwei 已提交
285
  std::hash<std::string> hasher;
C
chengduo 已提交
286
  for (auto &g : *subgraphs) {
Y
Yan Chunwei 已提交
287 288
    // 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
289
    std::sort(sorted_keys.begin(), sorted_keys.end(), GraphItemLessThan());
Y
Yan Chunwei 已提交
290 291 292
    std::stringstream ss;
    for (auto &item : sorted_keys) {
      ss << item.first << ":" << item.second;
293
    }
Y
Yan Chunwei 已提交
294
    auto key = hasher(ss.str());
295 296 297 298 299 300 301 302
    if (!set.count(key)) {
      result.emplace_back(g);
      set.insert(key);
    }
  }
  *subgraphs = result;
}

303
void GraphPatternDetector::RemoveOverlappedMatch(
C
chengduo 已提交
304
    std::vector<subgraph_t> *subgraphs) {
305
  std::vector<subgraph_t> result;
C
chengduo 已提交
306
  std::unordered_set<Node *> node_set;
307

C
chengduo 已提交
308
  for (const auto &subgraph : *subgraphs) {
309
    bool valid = true;
C
chengduo 已提交
310
    for (auto &item : subgraph) {
Y
Yan Chunwei 已提交
311
      if (item.first->IsIntermediate() && node_set.count(item.second)) {
312 313 314 315 316
        valid = false;
        break;
      }
    }
    if (valid) {
C
chengduo 已提交
317
      for (auto &item : subgraph) {
318 319 320 321 322 323 324 325
        node_set.insert(item.second);
      }
      result.push_back(subgraph);
    }
  }
  *subgraphs = result;
}

326 327 328 329 330
std::string PDPattern::DotString() const {
  using inference::analysis::Dot;
  Dot dot;
  int id = 0;
  // Create Nodes
C
chengduo 已提交
331 332
  std::unordered_map<PDNode *, std::string> node2dot;
  for (const auto &node : nodes()) {
333 334 335 336 337
    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
338
  for (const auto &edge : edges()) {
339 340 341 342
    if (!node2dot.count(edge.first) || !node2dot.count(edge.second)) {
      LOG(ERROR) << "no node " << edge.first << " " << edge.second;
      continue;
    }
C
chengduo 已提交
343 344
    auto &src = node2dot.at(edge.first);
    auto &trg = node2dot.at(edge.second);
345 346 347 348 349
    dot.AddEdge(src, trg, {});
  }
  return dot.Build();
}

C
chengduo 已提交
350
PDNode &PDNode::LinksTo(const std::vector<PDNode *> &others) {
351
  // extend outlinks.
C
chengduo 已提交
352
  for (PDNode *x : others) {
353 354 355 356 357
    pattern_->AddEdge(this, x);
  }
  return *this;
}

C
chengduo 已提交
358
PDNode &PDNode::LinksFrom(const std::vector<PDNode *> &others) {
359
  // extend outlinks.
C
chengduo 已提交
360
  for (PDNode *x : others) {
361 362 363 364 365
    pattern_->AddEdge(x, this);
  }
  return *this;
}

C
chengduo 已提交
366 367
PDNode *PDNode::assert_is_op() {
  asserts_.emplace_back([](Node *x) { return x && x->IsOp(); });
Y
Yan Chunwei 已提交
368 369
  return this;
}
C
chengduo 已提交
370 371 372

PDNode *PDNode::assert_is_op(const std::string &op_type) {
  asserts_.emplace_back([op_type](Node *x) {
Y
Yan Chunwei 已提交
373 374 375 376
    return x && x->IsOp() && x->Op()->Type() == op_type;
  });
  return this;
}
C
chengduo 已提交
377 378 379 380 381 382 383 384

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

PDNode *PDNode::assert_is_not_ctrl_var() {
  asserts_.emplace_back([](Node *x) { return x && !x->IsCtrlVar(); });
Y
Yan Chunwei 已提交
385 386
  return this;
}
C
chengduo 已提交
387 388

PDNode *PDNode::assert_var_not_persistable() {
Y
Yan Chunwei 已提交
389
  assert_is_var();
C
chengduo 已提交
390
  asserts_.emplace_back([](Node *x) { return !x->Var()->Persistable(); });
Y
Yan Chunwei 已提交
391 392
  return this;
}
C
chengduo 已提交
393 394

PDNode *PDNode::assert_is_persistable_var() {
Y
Yan Chunwei 已提交
395
  assert_is_var();
C
chengduo 已提交
396
  asserts_.emplace_back([=](Node *x) { return x->Var()->Persistable(); });
Y
Yan Chunwei 已提交
397 398
  return this;
}
C
chengduo 已提交
399 400 401

PDNode *PDNode::assert_is_op_nth_input(const std::string &op_type,
                                       const std::string &argument, int nth) {
Y
Yan Chunwei 已提交
402 403
  assert_is_var();
  assert_is_op_input(op_type);
C
chengduo 已提交
404 405
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
406 407 408
      if (op->IsOp() && op->Op()->Type() == op_type &&
          IsNthInput(x, op, argument, nth))
        return true;
Y
Yan Chunwei 已提交
409 410 411 412 413
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
414 415 416

PDNode *PDNode::assert_is_op_nth_output(const std::string &op_type,
                                        const std::string &argument, int nth) {
Y
Yan Chunwei 已提交
417
  assert_is_var();
C
chengduo 已提交
418 419
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
420 421 422
      if (op->IsOp() && op->Op()->Type() == op_type &&
          IsNthOutput(x, op, argument, nth))
        return true;
Y
Yan Chunwei 已提交
423 424 425 426 427
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
428 429

PDNode *PDNode::assert_is_only_input_of_op(const std::string &op_type) {
Y
Yan Chunwei 已提交
430
  assert_is_var();
C
chengduo 已提交
431 432
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
Y
Yan Chunwei 已提交
433 434 435 436 437 438 439 440 441
      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
442 443

PDNode *PDNode::assert_is_only_output_of_op(const std::string &op_type) {
Y
Yan Chunwei 已提交
444
  assert_is_var();
C
chengduo 已提交
445 446
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
Y
Yan Chunwei 已提交
447 448 449 450 451 452 453 454 455
      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
456 457

PDNode *PDNode::assert_is_op_output(const std::string &op_type) {
Y
Yan Chunwei 已提交
458
  assert_is_var();
C
chengduo 已提交
459 460
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->inputs) {
Y
Yan Chunwei 已提交
461 462 463 464 465 466 467 468
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
469 470 471

PDNode *PDNode::assert_is_op_output(const std::string &op_type,
                                    const std::string &argument) {
472 473 474 475
  assert_is_var();
  assert_is_op_nth_output(op_type, argument, 0);
  return this;
}
C
chengduo 已提交
476
PDNode *PDNode::assert_is_op_input(const std::string &op_type) {
Y
Yan Chunwei 已提交
477
  assert_is_var();
C
chengduo 已提交
478 479
  asserts_.emplace_back([=](Node *x) {
    for (auto *op : x->outputs) {
Y
Yan Chunwei 已提交
480 481 482 483 484 485 486 487
      if (op && op->IsOp() && op->Op() && op->Op()->Type() == op_type) {
        return true;
      }
    }
    return false;
  });
  return this;
}
C
chengduo 已提交
488 489 490

PDNode *PDNode::assert_is_op_input(const std::string &op_type,
                                   const std::string &argument) {
491 492 493 494
  assert_is_var();
  assert_is_op_nth_input(op_type, argument, 0);
  return this;
}
C
chengduo 已提交
495 496

PDNode *PDNode::assert_op_has_n_inputs(const std::string &op_type, size_t n) {
Y
Yan Chunwei 已提交
497
  assert_is_op(op_type);
C
chengduo 已提交
498
  asserts_.emplace_back([=](Node *x) { return x->inputs.size() == n; });
Y
Yan Chunwei 已提交
499 500
  return this;
}
C
chengduo 已提交
501 502

PDNode *PDNode::assert_op_has_n_outputs(const std::string &op_type, size_t n) {
Y
Yan Chunwei 已提交
503
  assert_is_op(op_type);
C
chengduo 已提交
504
  asserts_.emplace_back([=](Node *x) { return x->outputs.size() == n; });
Y
Yan Chunwei 已提交
505 506
  return this;
}
C
chengduo 已提交
507 508

PDNode *PDNode::assert_more(PDNode::teller_t &&teller) {
Y
Yan Chunwei 已提交
509 510 511 512
  asserts_.emplace_back(std::move(teller));
  return this;
}

C
chengduo 已提交
513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
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) {
596 597 598 599 600 601
    if (out->IsOp() && out->Op()->Type() == op_type) {
      return true;
    }
  }
  return false;
}
C
chengduo 已提交
602 603

bool IsNthInput(Node *var, Node *op, const std::string &argument, size_t nth) {
604 605 606 607 608
  PADDLE_ENFORCE(var->IsVar());
  PADDLE_ENFORCE(op->IsOp());
  if (op->Op()->Input(argument).size() <= nth) return false;
  return var->Name() == op->Op()->Input(argument)[nth];
}
C
chengduo 已提交
609 610

bool IsNthOutput(Node *var, Node *op, const std::string &argument, size_t nth) {
611 612 613 614 615
  PADDLE_ENFORCE(var->IsVar());
  PADDLE_ENFORCE(op->IsOp());
  if (op->Op()->Output(argument).size() <= nth) return false;
  return var->Name() == op->Op()->Output(argument)[nth];
}
C
chengduo 已提交
616 617 618 619 620

void GraphSafeRemoveNodes(Graph *graph,
                          const std::unordered_set<const Node *> &nodes) {
  for (auto *node : nodes) {
    graph->RemoveNode(const_cast<Node *>(node));
621 622
  }

C
chengduo 已提交
623
  for (auto *node : graph->Nodes()) {
624 625
    for (auto it = node->inputs.begin(); it != node->inputs.end();) {
      if (nodes.count(*it)) {
C
chengduo 已提交
626
        it = const_cast<Node *>(node)->inputs.erase(it);
627
      } else {
628
        it++;
629
      }
630 631 632
    }
    for (auto it = node->outputs.begin(); it != node->outputs.end();) {
      if (nodes.count(*it)) {
C
chengduo 已提交
633
        it = const_cast<Node *>(node)->outputs.erase(it);
634
      } else {
635
        it++;
636
      }
637 638 639
    }
  }
}
C
chengduo 已提交
640 641 642

bool VarLinksFromOp(Node *node, const std::string &op_type) {
  for (auto *out : node->inputs) {
643 644 645 646 647 648 649
    if (out->IsOp() && out->Op()->Type() == op_type) {
      return true;
    }
  }
  return false;
}

S
Sylwester Fraczek 已提交
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
PDNode *patterns::ConvBN::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 *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()
                              ->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 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;
}

C
chengduo 已提交
756 757
PDNode *patterns::ConvReLU::operator()(
    paddle::framework::ir::PDNode *conv_input) {
758 759
  // Create Operators
  conv_input->assert_is_op_input("conv2d", "Input");
C
chengduo 已提交
760 761
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op("conv2d");
  auto *relu_op = pattern->NewNode(relu_repr())->assert_is_op("relu");
762 763
  // Create variables
  // Filter
C
chengduo 已提交
764
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
765 766 767 768
                              ->AsInput()
                              ->assert_is_persistable_var()
                              ->assert_is_op_input("conv2d", "Filter");
  // intermediate variable, will be removed in the IR after fuse.
C
chengduo 已提交
769
  auto *conv_out_var = pattern->NewNode(conv_out_repr())
770 771 772 773
                           ->AsIntermediate()
                           ->assert_is_only_output_of_op("conv2d")
                           ->assert_is_op_input("relu");
  // output
C
chengduo 已提交
774
  auto *relu_out_var = pattern->NewNode(relu_out_repr())
775 776 777
                           ->AsOutput()
                           ->assert_is_op_output("relu");

778
  conv_op->LinksFrom({conv_input, conv_weight_var}).LinksTo({conv_out_var});
779 780 781 782
  relu_op->LinksFrom({conv_out_var}).LinksTo({relu_out_var});
  return relu_out_var;
}

T
tensor-tang 已提交
783 784 785 786
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
787 788 789 790
  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
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

  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
828
PDNode *patterns::FC::operator()(paddle::framework::ir::PDNode *x,
Y
Yan Chunwei 已提交
829 830 831
                                 bool with_bias) {
  // Create shared nodes.
  x->assert_is_op_input("mul", "X");
C
chengduo 已提交
832
  auto *mul = pattern->NewNode(mul_repr())->assert_is_op("mul");
Y
Yan Chunwei 已提交
833

C
chengduo 已提交
834
  auto *mul_w_var = pattern->NewNode(w_repr())
Y
Yan Chunwei 已提交
835 836 837 838
                        ->AsInput()
                        ->assert_is_persistable_var()
                        ->assert_is_op_input("mul", "Y");

C
chengduo 已提交
839
  auto *mul_out_var =
Y
Yan Chunwei 已提交
840 841 842 843 844 845 846 847 848 849
      pattern->NewNode(mul_out_repr())->assert_is_op_output("mul");

  if (!with_bias) {  // not with bias
    // Add links.
    mul->LinksFrom({x, mul_w_var}).LinksTo({mul_out_var});
    return mul_out_var;

  } else {  // with bias
    mul_out_var->AsIntermediate()->assert_is_op_input("elementwise_add");
    // Create operators.
C
chengduo 已提交
850
    auto *elementwise_add = pattern->NewNode(elementwise_add_repr())
Y
Yan Chunwei 已提交
851 852
                                ->assert_is_op("elementwise_add");
    // Create variables.
C
chengduo 已提交
853
    auto *bias = pattern->NewNode(bias_repr())
Y
Yan Chunwei 已提交
854 855 856
                     ->assert_is_op_input("elementwise_add")
                     ->AsInput();

C
chengduo 已提交
857
    auto *fc_out = pattern->NewNode(Out_repr())
Y
Yan Chunwei 已提交
858 859 860 861 862 863
                       ->AsOutput()
                       ->assert_is_op_output("elementwise_add");

    mul->LinksFrom({mul_w_var, x}).LinksTo({mul_out_var});
    elementwise_add->LinksFrom({mul_out_var, bias}).LinksTo({fc_out});
    return fc_out;
864 865
  }
}
T
tensor-tang 已提交
866

867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
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
885
PDNode *patterns::LSTM::operator()(PDNode *x) {
886
  x->assert_is_op_input("lstm", "Input");
C
chengduo 已提交
887
  auto *lstm_op = pattern->NewNode(lstm_repr())->assert_is_op("lstm");
Y
Yan Chunwei 已提交
888
#define NEW_NODE(arg__, io__) \
C
chengduo 已提交
889
  auto *arg__ =               \
Y
Yan Chunwei 已提交
890
      pattern->NewNode(arg__##_repr())->assert_is_op_##io__("lstm", #arg__);
891 892 893 894 895

  // 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
896 897
  NEW_NODE(Weight, input);
  NEW_NODE(Bias, input);
898

Y
Yan Chunwei 已提交
899 900 901 902 903
  NEW_NODE(Hidden, output);
  NEW_NODE(Cell, output);
  NEW_NODE(BatchGate, output);
  NEW_NODE(BatchCellPreAct, output);
#undef NEW_NODE
904 905 906 907 908

  lstm_op->LinksFrom({x, Weight, Bias});
  lstm_op->LinksTo({Hidden, Cell, BatchGate, BatchCellPreAct});
  return Hidden;
}
T
tensor-tang 已提交
909

C
chengduo 已提交
910
PDNode *patterns::GRU::operator()(PDNode *x) {
T
tensor-tang 已提交
911
  x->assert_is_op_input("gru", "Input");
C
chengduo 已提交
912
  auto *gru_op = pattern->NewNode(gru_repr())->assert_is_op("gru");
Y
Yan Chunwei 已提交
913
#define NEW_NODE(arg__, io__) \
C
chengduo 已提交
914
  auto *arg__ =               \
Y
Yan Chunwei 已提交
915
      pattern->NewNode(arg__##_repr())->assert_is_op_##io__("gru", #arg__);
T
tensor-tang 已提交
916

Y
Yan Chunwei 已提交
917
  NEW_NODE(Weight, input);
T
tensor-tang 已提交
918 919
  // TODO(Superjomn): upgrade the fuse framework to support optional.
  // H0 and bias are optional
Y
Yan Chunwei 已提交
920
  NEW_NODE(Bias, input);  // also optional
T
tensor-tang 已提交
921 922
  // NEW_NODE(H0, input);

Y
Yan Chunwei 已提交
923
  NEW_NODE(Hidden, output);
T
tensor-tang 已提交
924
  // below are intermediate
Y
Yan Chunwei 已提交
925 926 927 928
  NEW_NODE(BatchGate, output);
  NEW_NODE(BatchResetHiddenPrev, output);
  NEW_NODE(BatchHidden, output);
#undef NEW_NODE
T
tensor-tang 已提交
929

T
tensor-tang 已提交
930 931 932 933
  BatchGate->AsIntermediate();
  BatchResetHiddenPrev->AsIntermediate();
  BatchHidden->AsIntermediate();

T
tensor-tang 已提交
934 935 936 937 938
  gru_op->LinksFrom({x, Weight, Bias});
  gru_op->LinksTo({Hidden, BatchGate, BatchResetHiddenPrev, BatchHidden});
  return Hidden;
}

C
chengduo 已提交
939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
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;
}

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;
}

M
Michal Gallus 已提交
1033
PDNode *patterns::ConvBias::operator()(
1034
    paddle::framework::ir::PDNode *conv_input, bool is_conv3d) {
Y
Yihua Xu 已提交
1035
  std::string type = is_conv3d ? "conv3d" : "conv2d";
M
Michal Gallus 已提交
1036
  // Create Operators
Y
Yihua Xu 已提交
1037 1038
  conv_input->assert_is_op_input(type, "Input");
  auto *conv_op = pattern->NewNode(conv_repr())->assert_is_op(type);
M
Michal Gallus 已提交
1039 1040 1041 1042
  auto *eltiwse_op =
      pattern->NewNode(eltwise_repr())->assert_is_op("elementwise_add");
  // Create variables
  // Filter
Y
Yihua Xu 已提交
1043 1044 1045 1046
  auto *conv_weight_var = pattern->NewNode(conv_weight_repr())
                              ->AsInput()
                              ->assert_is_persistable_var()
                              ->assert_is_op_input(type, "Filter");
M
Michal Gallus 已提交
1047
  // intermediate variable, will be removed in the IR after fuse.
Y
Yihua Xu 已提交
1048 1049 1050 1051
  auto *conv_out_var = pattern->NewNode(conv_out_repr())
                           ->AsIntermediate()
                           ->assert_is_only_output_of_op(type)
                           ->assert_is_op_input("elementwise_add");
M
Michal Gallus 已提交
1052 1053 1054
  // Bias stored in elementwise_add
  auto *eltwise_bias_var = pattern->NewNode(eltwise_bias_repr())
                               ->AsInput()
M
Michal Gallus 已提交
1055
                               ->assert_is_persistable_var()
M
Michal Gallus 已提交
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
                               ->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;
}

1067 1068 1069 1070
PDNode *patterns::Conv::operator()() {
  auto conv_op = pattern->NewNode(conv_op_repr())->assert_is_op("conv2d");

  auto input_var = pattern->NewNode(conv_input_repr())
1071
                       ->AsInput()
1072 1073 1074
                       ->assert_is_op_input("conv2d", "Input");

  auto filter_var = pattern->NewNode(conv_filter_repr())
1075
                        ->AsInput()
1076 1077 1078
                        ->assert_is_op_input("conv2d", "Filter");

  auto output_var = pattern->NewNode(conv_output_repr())
1079
                        ->AsOutput()
1080 1081
                        ->assert_is_op_output("conv2d", "Output");

1082
  conv_op->LinksFrom({input_var, filter_var});
1083 1084 1085 1086 1087
  conv_op->LinksTo({output_var});

  return output_var;
}

1088
PDNode *patterns::ElementwiseAdd::operator()(PDNode *x_var, PDNode *y_var) {
1089 1090 1091
  auto elementwise_add_op = pattern->NewNode(elementwise_add_op_repr())
                                ->assert_is_op("elementwise_add");

1092 1093
  x_var->AsInput()->assert_is_op_input("elementwise_add", "X");
  y_var->AsInput()->assert_is_op_input("elementwise_add", "Y");
1094 1095 1096 1097
  auto out_var = pattern->NewNode(elementwise_add_out_repr())
                     ->AsOutput()
                     ->assert_is_op_output("elementwise_add", "Out");

1098
  elementwise_add_op->LinksFrom({x_var, y_var});
1099 1100 1101 1102
  elementwise_add_op->LinksTo({out_var});

  return out_var;
}
1103

H
hjchen2 已提交
1104
std::unordered_set<std::string> conv_act_set({"identity", "relu"});
1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 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

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())
                                  ->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())
                                  ->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
1170
                                 ->assert_is_op_input("elementwise_add", "Y")
1171 1172 1173 1174 1175
                                 ->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
1176
                                    ->assert_is_op_input("elementwise_add", "X")
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
                                    ->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
1204 1205
  elementwise_add_op_1->LinksFrom({elementwise_add_out, elementwise_add_in_y_1})
      .LinksTo({elementwise_add_out_1});
1206 1207 1208 1209
  act_op->LinksFrom({elementwise_add_out_1}).LinksTo({act_out});
  return act_out;
}

N
nhzlx 已提交
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
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())
                                  ->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
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 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
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()
                           ->assert_is_op_input("affine_channel", "Scale");
  // AC Bias
  auto *ac_bias_var = pattern->NewNode(ac_bias_repr())
                          ->AsInput()
                          ->assert_is_persistable_var()
                          ->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;
}

1309 1310 1311
}  // namespace ir
}  // namespace framework
}  // namespace paddle