Fusion: seqpool_cvm_concat (#18471)

* add fusion_seqpool_cvm_concat test=develop

* simplify pass, test=develop

* fix code style, test=develop

paddle/fluid/framework/ir/CMakeLists.txt

@@ -61,6 +61,7 @@ pass_library(multi_batch_merge_pass base)
 pass_library(conv_bn_fuse_pass inference)
 pass_library(seqconv_eltadd_relu_fuse_pass inference)
 pass_library(seqpool_concat_fuse_pass inference)
+pass_library(seqpool_cvm_concat_fuse_pass inference)
 pass_library(repeated_fc_relu_fuse_pass inference)
 pass_library(squared_mat_sub_fuse_pass inference)
 pass_library(is_test_pass base)
@@ -118,6 +119,7 @@ cc_test(graph_to_program_pass_test SRCS graph_to_program_pass_test.cc DEPS graph
 cc_test(test_graph_pattern_detector SRCS graph_pattern_detector_tester.cc DEPS graph_pattern_detector)
 cc_test(test_fc_fuse_pass SRCS fc_fuse_pass_tester.cc DEPS fc_fuse_pass framework_proto)
 cc_test(test_seqpool_concat_fuse_pass SRCS seqpool_concat_fuse_pass_tester.cc DEPS seqpool_concat_fuse_pass framework_proto)
+cc_test(test_seqpool_cvm_concat_fuse_pass SRCS seqpool_cvm_concat_fuse_pass_tester.cc DEPS seqpool_cvm_concat_fuse_pass framework_proto)
 cc_test(test_is_test_pass SRCS is_test_pass_tester.cc DEPS is_test_pass)
 if(NOT WIN32)
     cc_test(test_sync_batch_norm_pass SRCS sync_batch_norm_pass_tester.cc DEPS sync_batch_norm_pass)

paddle/fluid/framework/ir/seqpool_cvm_concat_fuse_pass.cc

+/* 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. */
+
+#include "paddle/fluid/framework/ir/seqpool_cvm_concat_fuse_pass.h"
+#include <string>
+#include <unordered_map>
+#include <unordered_set>
+#include <vector>
+#include "paddle/fluid/framework/lod_tensor.h"
+
+namespace paddle {
+namespace framework {
+namespace ir {
+
+namespace {
+static PDNode* BuildCVMConcatPattern(PDPattern* pattern) {
+  auto cvm_behind_x = [](Node* x) -> bool {
+    Node* adj = x->inputs[0];
+    Node* alt = x->inputs[0]->inputs[0];
+    return x && adj && adj->IsVar() && alt->IsOp() &&
+           alt->Op()->Type() == "cvm";
+  };
+  auto* concat_op_node = pattern->NewNode("concat_op")
+                             ->assert_is_op("concat")
+                             ->assert_op_attr<int>("axis", 1)
+                             ->assert_more(cvm_behind_x);
+  return concat_op_node;
+}
+
+static void GetConcatNodes(ir::Graph* graph, std::vector<Node*>* concat_nodes) {
+  GraphPatternDetector gpd;
+  auto* pattern = gpd.mutable_pattern();
+  auto concat_op_node = BuildCVMConcatPattern(pattern);
+  GraphPatternDetector::handle_t handler = [&](
+      const GraphPatternDetector::subgraph_t& subgraph, Graph* graph) {
+    Node* concat_op = subgraph.at(concat_op_node);
+    concat_nodes->push_back(concat_op);
+  };
+  gpd(graph, handler);
+}
+}  // anonymous namespace
+
+void SeqPoolCVMConcatFusePass::ApplyImpl(ir::Graph* graph) const {
+  FusePassBase::Init("seqpool_cvm_concat_fuse", graph);
+  std::vector<Node*> concat_nodes;
+  GetConcatNodes(graph, &concat_nodes);
+
+  int count = 0;
+  for (auto* concat_node : concat_nodes) {
+    GraphPatternDetector gpd;
+    auto* pattern = gpd.mutable_pattern();
+    auto concat_before_x = [=](Node* x) -> bool {
+      return x && x->outputs[0] == concat_node;
+    };
+    PDNode* seqpool_in_var_node =
+        pattern->NewNode("seqpool_in_var")
+            ->assert_is_only_input_of_op("sequence_pool");
+    PDNode* seqpool_op_node =
+        pattern->NewNode("seqpool_op")
+            ->assert_is_op("sequence_pool")
+            ->assert_op_attr<std::string>("pooltype", "SUM");
+    PDNode* seqpool_out_var_node =
+        pattern->NewNode("seqpool_out_var")
+            ->assert_is_op_nth_output("sequence_pool", "Out", 0)
+            ->assert_is_op_nth_input("cvm", "X", 0);
+    PDNode* seqpool_idx_out_var_node =
+        pattern->NewNode("seqpool_idx_out_var")
+            ->assert_is_op_nth_output("sequence_pool", "MaxIndex", 0);
+    PDNode* cvm_op_node =
+        pattern->NewNode("cvm_op")->assert_is_op("cvm")->assert_op_attr<bool>(
+            "use_cvm", true);
+    PDNode* cvm_out_var_node = pattern->NewNode("cvm_op_out_var")
+                                   ->assert_is_op_nth_output("cvm", "Y", 0)
+                                   ->assert_more(concat_before_x);
+    PDNode* cvm_cvm_in_var_node = pattern->NewNode("cvm_cvm_in_var")
+                                      ->assert_is_op_nth_input("cvm", "CVM", 0);
+
+    seqpool_op_node->LinksFrom({seqpool_in_var_node})
+        .LinksTo({seqpool_out_var_node, seqpool_idx_out_var_node});
+    seqpool_out_var_node->LinksFrom({seqpool_op_node}).LinksTo({cvm_op_node});
+    cvm_op_node->LinksTo({cvm_out_var_node})
+        .LinksFrom({cvm_cvm_in_var_node, seqpool_out_var_node});
+
+    std::unordered_map<std::string, Node*> ins_to_concat;
+    std::vector<Node*> subgraph_ins;
+    std::vector<std::string> subgraph_ins_name;
+    std::unordered_set<const Node*> marked_nodes;
+
+    Node* cvm_input_of_cvm;
+    Node* concat_out_var = concat_node->outputs[0];
+
+    GraphPatternDetector::handle_t handler = [&](
+        const GraphPatternDetector::subgraph_t& subgraph, Graph* graph) {
+      Node* seqpool_in_var = subgraph.at(seqpool_in_var_node);
+      Node* seqpool_op = subgraph.at(seqpool_op_node);
+      Node* seqpool_out_var = subgraph.at(seqpool_out_var_node);
+      Node* seqpool_idx_out_var = subgraph.at(seqpool_idx_out_var_node);
+      Node* cvm_op = subgraph.at(cvm_op_node);
+      Node* cvm_out_var = subgraph.at(cvm_out_var_node);
+      cvm_input_of_cvm = subgraph.at(cvm_cvm_in_var_node);
+      marked_nodes.insert({seqpool_op, seqpool_out_var, seqpool_idx_out_var,
+                           cvm_op, cvm_out_var, concat_node});
+      ins_to_concat[cvm_out_var->Name()] = seqpool_in_var;
+    };
+    gpd(graph, handler);
+
+    if (!ins_to_concat.empty()) {
+      for (const auto* in : concat_node->inputs) {
+        subgraph_ins.push_back(ins_to_concat.at(in->Name()));
+        subgraph_ins_name.push_back(ins_to_concat.at(in->Name())->Name());
+      }
+
+      // Create New OpDesc
+      OpDesc op_desc;
+      op_desc.SetType("fusion_seqpool_cvm_concat");
+      op_desc.SetInput("X", subgraph_ins_name);
+      op_desc.SetInput("CVM", {cvm_input_of_cvm->Name()});
+      op_desc.SetAttr("pooltype", std::string("SUM"));
+      op_desc.SetAttr("use_cvm", true);
+      op_desc.SetAttr("axis", concat_node->Op()->GetAttr("axis"));
+      op_desc.SetOutput("Out", {concat_out_var->Name()});
+      auto* op = graph->CreateOpNode(&op_desc);
+
+      for (size_t i = 0; i < subgraph_ins.size(); ++i) {
+        IR_NODE_LINK_TO(subgraph_ins[i], op);
+      }
+      IR_NODE_LINK_TO(cvm_input_of_cvm, op);
+      IR_NODE_LINK_TO(op, concat_out_var);
+
+      GraphSafeRemoveNodes(graph, marked_nodes);
+      count++;
+    }
+  }
+  AddStatis(count);
+}
+
+}  // namespace ir
+}  // namespace framework
+}  // namespace paddle
+
+REGISTER_PASS(seqpool_cvm_concat_fuse_pass,
+              paddle::framework::ir::SeqPoolCVMConcatFusePass);

paddle/fluid/framework/ir/seqpool_cvm_concat_fuse_pass.h

+/* 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. */
+
+#pragma once
+
+#include <string>
+#include "paddle/fluid/framework/ir/fuse_pass_base.h"
+#include "paddle/fluid/framework/ir/graph.h"
+#include "paddle/fluid/framework/ir/graph_pattern_detector.h"
+
+namespace paddle {
+namespace framework {
+namespace ir {
+
+/**
+ * Fuse SequencePool(with sum pooltype yet) and Concat;
+ *
+ * Before fuse:
+ *    |         |             |
+ * seq_pool, seq_pool, ... seq_pool
+ *    |         |             |
+ *   cvm       cvm           cvm
+ *    \         |      ...   /
+ *            concat
+ *              |
+ * After fuse:
+ *    \      |       /
+ * FusionSeqPoolCVMConcat
+ *           |
+ */
+class SeqPoolCVMConcatFusePass : public FusePassBase {
+ public:
+  virtual ~SeqPoolCVMConcatFusePass() {}
+
+ protected:
+  void ApplyImpl(ir::Graph* graph) const override;
+
+  const std::string name_scope_{"seqpool_cvm_concat_fuse"};
+};
+
+}  // namespace ir
+}  // namespace framework
+}  // namespace paddle

paddle/fluid/framework/ir/seqpool_cvm_concat_fuse_pass_tester.cc

+// 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.
+
+#include "paddle/fluid/framework/ir/seqpool_cvm_concat_fuse_pass.h"
+#include <gtest/gtest.h>
+#include "paddle/fluid/framework/op_proto_maker.h"
+
+namespace paddle {
+namespace framework {
+namespace ir {
+
+void SetOp(ProgramDesc* prog, const std::string& type,
+           const std::vector<std::string>& inputs,
+           const std::vector<std::string>& outputs) {
+  auto* op = prog->MutableBlock(0)->AppendOp();
+  op->SetType(type);
+  if (type == "sequence_pool") {
+    op->SetInput("X", {inputs[0]});
+    std::string pooltype = "SUM";
+    op->SetAttr("pooltype", pooltype);
+    op->SetOutput("MaxIndex", {outputs[0]});
+    op->SetOutput("Out", {outputs[1]});
+  } else if (type == "concat") {
+    op->SetInput("X", inputs);
+    op->SetAttr("axis", 1);
+    op->SetOutput("Out", {outputs[0]});
+  } else if (type == "cvm") {
+    op->SetInput("X", {inputs[0]});
+    op->SetInput("CVM", {inputs[1]});
+    op->SetOutput("Y", {outputs[0]});
+    op->SetAttr("use_cvm", true);
+  } else {
+    op->SetInput("X", inputs);
+    op->SetOutput("Out", outputs);
+  }
+  op->SetAttr(OpProtoAndCheckerMaker::OpRoleAttrName(),
+              static_cast<int>(OpRole::kForward));
+}
+
+int CountOpType(const ir::Graph* graph,
+                const std::string& op_type = "fusion_seqpool_cvm_concat") {
+  int count = 0;
+  for (auto* node : graph->Nodes()) {
+    if (node->IsOp() && node->Op()->Type() == op_type) {
+      ++count;
+    }
+  }
+  return count;
+}
+
+std::unique_ptr<ir::Graph> GetNumNodesOfBeforeAfter(
+    std::unique_ptr<ir::Graph> graph, int* before, int* after,
+    const std::string& pass_type = "seqpool_cvm_concat_fuse_pass") {
+  auto pass = PassRegistry::Instance().Get(pass_type);
+  *before = graph->Nodes().size();
+  graph.reset(pass->Apply(graph.release()));
+  *after = graph->Nodes().size();
+  return graph;
+}
+
+/*
+ * Before fuse:
+ *
+ *
+ *    a          b          c
+ *    |          |          |
+ *   op1        op2        op3
+ *   / \        / \        / \
+ *  d  e  n    f   g   n   h  i   n
+ *     |  /        |  /       |  /
+ *    op4         op5        op6
+ *     |           |          |
+       j           k          l
+ *     \           |         /
+ *               concat
+ *                 |
+ *                 m
+ *
+ * Type of op1, op2 and op3 are sequence_pool, with "SUM" pooltype attr.
+ * Type of op4, op5 and op6 are cvm, with use_cvm is true.
+ *
+ * After fuse:
+ *    a      b      c      n
+ *    \      |      |     /
+ *  fusion_seqpool_cvm_concat
+ *              |
+ *              m
+ */
+TEST(SeqPoolCVMConcatFusePass, basic) {
+  ProgramDesc prog;
+  for (auto& v :
+       std::vector<std::string>({"a", "b", "c", "d", "e", "f", "g", "h", "i",
+                                 "j", "k", "l", "m", "n"})) {
+    auto* var = prog.MutableBlock(0)->Var(v);
+    var->SetType(proto::VarType::LOD_TENSOR);
+  }
+
+  SetOp(&prog, "sequence_pool", std::vector<std::string>({"a"}),
+        std::vector<std::string>({"d", "e"}));
+  SetOp(&prog, "sequence_pool", std::vector<std::string>({"b"}),
+        std::vector<std::string>({"f", "g"}));
+  SetOp(&prog, "sequence_pool", std::vector<std::string>({"c"}),
+        std::vector<std::string>({"h", "i"}));
+  SetOp(&prog, "cvm", std::vector<std::string>({"e", "n"}),
+        std::vector<std::string>({"j"}));
+  SetOp(&prog, "cvm", std::vector<std::string>({"g", "n"}),
+        std::vector<std::string>({"k"}));
+  SetOp(&prog, "cvm", std::vector<std::string>({"i", "n"}),
+        std::vector<std::string>({"l"}));
+  SetOp(&prog, "concat", std::vector<std::string>({"j", "k", "l"}),
+        std::vector<std::string>({"m"}));
+
+  std::unique_ptr<ir::Graph> graph(new ir::Graph(prog));
+  int before, after;
+  graph = GetNumNodesOfBeforeAfter(std::move(graph), &before, &after);
+  // Remove 16 Nodes: op1, op2, op3, op4, op5, op6, d, e, f, g, h, i, j, k, l,
+  // concat_op
+  // Add 1 Node: fusion_seqpool_cvm_concat
+  EXPECT_EQ(after, before - 15);
+  EXPECT_EQ(CountOpType(graph.get()), 1);
+}
+
+/*
+ * Before fuse:
+ *    a               b
+ *    |           /       \
+ *   op1  k     op2   k   op3
+ *   / \ /      / \  /      \
+ *  c  d       e   f         g
+ *     |           |
+ *    op4         op5
+ *     |           |
+ *     h           i
+ *      \         /
+ *        concat
+ *          |
+ *          j
+ * Type of op1 and op2 are sequence_pool, with "SUM" pooltype attr.
+ * Type of op4 and op5 are cvm, with use_cvm is true.
+ *
+ * After fuse:
+ *   a          k              b
+ *    \         |           /     \
+ *   fusion_seqpool_cvm_concat    op3
+ *              |                  |
+ *              j                  g
+ */
+TEST(SeqPoolCVMConcatFusePass, advanced) {
+  ProgramDesc prog;
+  for (auto& v : std::vector<std::string>(
+           {"a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k"})) {
+    auto* var = prog.MutableBlock(0)->Var(v);
+    var->SetType(proto::VarType::LOD_TENSOR);
+  }
+
+  SetOp(&prog, "sequence_pool", std::vector<std::string>({"a"}),
+        std::vector<std::string>({"c", "d"}));
+  SetOp(&prog, "sequence_pool", std::vector<std::string>({"b"}),
+        std::vector<std::string>({"e", "f"}));
+  SetOp(&prog, "op3", std::vector<std::string>({"b"}),
+        std::vector<std::string>({"g"}));
+  SetOp(&prog, "cvm", std::vector<std::string>({"d", "k"}),
+        std::vector<std::string>({"h"}));
+  SetOp(&prog, "cvm", std::vector<std::string>({"f", "k"}),
+        std::vector<std::string>({"i"}));
+  SetOp(&prog, "concat", std::vector<std::string>({"h", "i"}),
+        std::vector<std::string>({"j"}));
+
+  std::unique_ptr<ir::Graph> graph(new ir::Graph(prog));
+  int before, after;
+  graph = GetNumNodesOfBeforeAfter(std::move(graph), &before, &after);
+  // Remove 11 Nodes: op1, op2, op4, op5, c, d, e, f, h, i, concat_op
+  // Add 1 Node: fusion_seqpool_cvm_concat
+  EXPECT_EQ(after, before - 10);
+  EXPECT_EQ(CountOpType(graph.get()), 1);
+}
+
+ProgramDesc BuildProgramDesc(int num_inputs_of_concat) {
+  ProgramDesc prog;
+  auto new_var = [&](const std::string& name) {
+    auto* var = prog.MutableBlock(0)->Var(name);
+    var->SetType(proto::VarType::LOD_TENSOR);
+  };
+  std::vector<std::string> concat_inputs;
+  new_var("cvm_in");
+  for (int i = 0; i < num_inputs_of_concat; ++i) {
+    std::string seqpool_prefix = "seqpool_op_" + std::to_string(i);
+    new_var(seqpool_prefix + "in");
+    new_var(seqpool_prefix + "out");
+    new_var(seqpool_prefix + "out_unused");
+    SetOp(&prog, "sequence_pool",
+          std::vector<std::string>({seqpool_prefix + "in"}),
+          std::vector<std::string>(
+              {seqpool_prefix + "out_unused", seqpool_prefix + "out"}));
+
+    std::string cvm_prefix = "cvm_op_" + std::to_string(i);
+    new_var(cvm_prefix + "out");
+    SetOp(&prog, "cvm",
+          std::vector<std::string>({seqpool_prefix + "out", "cvm_in"}),
+          std::vector<std::string>({cvm_prefix + "out"}));
+
+    concat_inputs.push_back(cvm_prefix + "out");
+  }
+  SetOp(&prog, "concat", concat_inputs,
+        std::vector<std::string>({"concat_out"}));
+  return prog;
+}
+
+// test more inputs of concat
+TEST(SeqPoolCVMConcatFusePass, more_inputs) {
+  for (int num : {1, 2, 10}) {
+    ProgramDesc prog = BuildProgramDesc(num);
+    std::unique_ptr<ir::Graph> graph(new ir::Graph(prog));
+    int before, after;
+    graph = GetNumNodesOfBeforeAfter(std::move(graph), &before, &after);
+    // Remove Nodes: n * (seqpool_op, seqpool_out, out_unused, cvm_op, cvm_out),
+    // and concat_op
+    // Add Node: fusion_seqpool_cvm_concat op
+    EXPECT_EQ(after, before - num * 5);
+    EXPECT_EQ(CountOpType(graph.get()), 1);
+  }
+}
+
+}  // namespace ir
+}  // namespace framework
+}  // namespace paddle
+
+USE_PASS(seqpool_cvm_concat_fuse_pass);

paddle/fluid/inference/api/paddle_pass_builder.cc

@@ -145,6 +145,7 @@ CpuPassStrategy::CpuPassStrategy() : PassStrategy({}) {
                   "attention_lstm_fuse_pass",       //
                   "seqconv_eltadd_relu_fuse_pass",  //
                   // "seqpool_concat_fuse_pass",    //
+                  "seqpool_cvm_concat_fuse_pass",  //
                   // "embedding_fc_lstm_fuse_pass", //
                   "fc_lstm_fuse_pass",             //
                   "mul_lstm_fuse_pass",            //

paddle/fluid/operators/fused/fusion_seqpool_cvm_concat_op.cc

+/* 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. */
+
+#include "paddle/fluid/operators/fused/fusion_seqpool_cvm_concat_op.h"
+#include <string>
+#include <vector>
+#include "paddle/fluid/operators/jit/kernels.h"
+
+namespace paddle {
+namespace operators {
+
+void FusionSeqPoolCVMConcatOp::InferShape(
+    framework::InferShapeContext* ctx) const {
+  PADDLE_ENFORCE_GE(
+      ctx->Inputs("X").size(), 1UL,
+      "Inputs(X) of FusionSeqPoolCVMConcatOp should not be empty.");
+  PADDLE_ENFORCE(ctx->HasOutput("Out"),
+                 "Output(Out) of FusionSeqPoolCVMConcatOp should not be null.");
+  int axis = ctx->Attrs().Get<int>("axis");
+  PADDLE_ENFORCE_EQ(
+      axis, 1, "FusionSeqPoolCVMConcatOp only supports concat axis=1 yet.");
+  bool use_cvm = ctx->Attrs().Get<bool>("use_cvm");
+  PADDLE_ENFORCE_EQ(
+      use_cvm, true,
+      "FusionSeqPoolCVMConcatOp only supports use_cvm is true yet.");
+
+  auto ins_dims = ctx->GetInputsDim("X");
+  const size_t n = ins_dims.size();
+  PADDLE_ENFORCE_GT(n, 0UL, "Input tensors count should > 0.");
+  if (n == 1) {
+    LOG(WARNING) << "Only have one input, may waste memory";
+  }
+
+  // The output height should be confirmed in Compute,
+  // since input lod is not accessible here.
+  PADDLE_ENFORCE_EQ(ins_dims[0].size(), 2,
+                    "The dims size of first input should be 2.");
+  ctx->SetOutputDim("Out", {-1, ins_dims[0][axis] * static_cast<int>(n)});
+}
+
+framework::OpKernelType FusionSeqPoolCVMConcatOp::GetExpectedKernelType(
+    const framework::ExecutionContext& ctx) const {
+  return framework::OpKernelType(
+      framework::GetDataTypeOfVar(ctx.MultiInputVar("X")[0]), ctx.GetPlace());
+}
+
+void FusionSeqPoolCVMConcatOpMaker::Make() {
+  AddInput("X", "(LoDTensor) Input tensors of this operator.").AsDuplicable();
+  AddInput("CVM",
+           "(Tensor),  a 2-D Tensor with shape [N x 2], where N is the batch "
+           "size, 2 is show and click.");
+  AddOutput("Out", "(LoDTensor) Output tensor of concat operator.");
+  AddAttr<std::string>("pooltype",
+                       "(string, default 'SUM') some of the pooling "
+                       "pooltype of SequencePoolOp.")
+      .SetDefault("SUM")
+      .InEnum({"AVERAGE", "SUM", "SQRT"});
+  AddAttr<bool>("use_cvm", "bool, use cvm or not").SetDefault(true);
+  AddAttr<int>("axis",
+               "The axis along which the input tensors will be concatenated. "
+               "Only supports concat axis=1 yet.")
+      .SetDefault(1);
+  AddComment(R"DOC(
+Fusion Sequence Pool of pooltype(sum, average and sqrt) and Concat Operator.
+)DOC");
+}
+
+template <typename T>
+class FusionSeqPoolCVMConcatKernel : public framework::OpKernel<T> {
+ public:
+  void Compute(const framework::ExecutionContext& ctx) const override {
+    auto ins = ctx.MultiInput<LoDTensor>("X");
+    auto* out = ctx.Output<LoDTensor>("Out");
+    std::string pooltype = ctx.Attr<std::string>("pooltype");
+    auto x0_lod = ins[0]->lod();
+    auto x0_dims = ins[0]->dims();
+    auto y_dims = out->dims();
+    size_t bs = x0_lod[0].size() - 1;
+    out->Resize({static_cast<int64_t>(bs), y_dims[1]});
+    framework::LoD y_lod(1);
+    y_lod[0].resize(bs + 1);
+    for (size_t i = 0; i <= bs; ++i) {
+      y_lod[0][i] = i;
+    }
+    out->set_lod(y_lod);
+    auto place = ctx.GetPlace();
+    T* y_data = out->mutable_data<T>(place);
+
+    int w = ins[0]->numel() / x0_dims[0];
+    PADDLE_ENFORCE_EQ(y_dims[1] % w, 0,
+                      "The output of dims[1] should be dividable of w");
+    jit::seq_pool_attr_t attr(w, jit::SeqPoolType::kSum);
+    if (pooltype == "AVERAGE") {
+      attr.type = jit::SeqPoolType::kAvg;
+    } else if (pooltype == "SQRT") {
+      attr.type = jit::SeqPoolType::kSqrt;
+    }
+    auto seqpool =
+        jit::KernelFuncs<jit::SeqPoolTuple<T>, platform::CPUPlace>::Cache().At(
+            attr);
+    size_t n = ins.size();
+    size_t dst_step_size = n * w;
+    for (size_t i = 0; i < n; ++i) {
+      auto x_dims = ins[i]->dims();
+      auto x_lod = ins[i]->lod()[0];
+      const T* src = ins[i]->data<T>();
+      T* dst = y_data + i * w;
+      PADDLE_ENFORCE_EQ(static_cast<int>(ins[i]->numel() / x_dims[0]), w,
+                        "Width of all inputs should be equal.");
+      PADDLE_ENFORCE_EQ(x_lod.size(), bs + 1,
+                        "Batchsize of all inputs should be equal.");
+      for (size_t j = 0; j < bs; ++j) {
+        attr.h = static_cast<int>(x_lod[j + 1] - x_lod[j]);
+        seqpool(src, dst, &attr);
+
+        // Currently only use_cvm is true.
+        dst[0] = log(dst[0] + 1);
+        dst[1] = log(dst[1] + 1) - dst[0];
+
+        dst += dst_step_size;
+        src += attr.h * attr.w;
+      }
+    }
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+REGISTER_OPERATOR(fusion_seqpool_cvm_concat, ops::FusionSeqPoolCVMConcatOp,
+                  ops::FusionSeqPoolCVMConcatOpMaker,
+                  paddle::framework::EmptyGradOpMaker);
+
+REGISTER_OP_CPU_KERNEL(fusion_seqpool_cvm_concat,
+                       ops::FusionSeqPoolCVMConcatKernel<float>,
+                       ops::FusionSeqPoolCVMConcatKernel<double>);

paddle/fluid/operators/fused/fusion_seqpool_cvm_concat_op.h

+/* 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. */
+
+#pragma once
+#include "paddle/fluid/framework/op_registry.h"
+
+namespace paddle {
+namespace operators {
+
+using LoDTensor = framework::LoDTensor;
+using Tensor = framework::Tensor;
+
+class FusionSeqPoolCVMConcatOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+  void InferShape(framework::InferShapeContext* ctx) const override;
+
+ protected:
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext& ctx) const override;
+};
+
+class FusionSeqPoolCVMConcatOpMaker : public framework::OpProtoAndCheckerMaker {
+ public:
+  void Make() override;
+};
+
+}  // namespace operators
+}  // namespace paddle

python/paddle/fluid/tests/unittests/test_fusion_seqpool_cvm_concat_op.py

+#   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.
+
+from __future__ import print_function
+
+import unittest
+import numpy as np
+from op_test import OpTest
+from test_reorder_lod_tensor import convert_to_offset
+from test_seq_pool import compute_seqpool_sum, compute_seqpool_avg, compute_seqpool_sqrt
+from test_cvm_op import cvm_compute
+
+
+class TestFusionSeqPoolCVMConcatOp(OpTest):
+    def setUp(self):
+        self.w = 11
+        self.use_cvm = True
+        self.lods = [[[2, 3, 5]], [[1, 5, 2]]]
+        self.set_conf()
+        self.set_pooltype()
+        self.op_type = 'fusion_seqpool_cvm_concat'
+        self.axis = 1
+        bs = len(self.lods[0][0])
+        inputs = []
+        outs = []
+        # The cvm variable is not actually used.
+        cvm = np.array([[0.6, 0.4]]).astype("float32")
+        i = 0
+        for lod in self.lods:
+            assert bs == len(lod[0]), 'All lod size should be equal'
+            x = np.random.uniform(0.1, 1,
+                                  [sum(lod[0]), self.w]).astype('float32')
+            offset = convert_to_offset(lod)
+            out = np.zeros((bs, self.w)).astype('float32')
+            if self.pooltype == "SUM":
+                compute_seqpool_sum(x, offset, out)
+                out = cvm_compute(out, self.w, self.use_cvm)
+            elif self.pooltype == "AVERAGE":
+                compute_seqpool_avg(x, offset, out)
+                out = cvm_compute(out, self.w, self.use_cvm)
+            elif self.pooltype == "SQRT":
+                compute_seqpool_sqrt(x, offset, out)
+                out = cvm_compute(out, self.w, self.use_cvm)
+            else:
+                raise Exception("Unsupported pool type!")
+            inputs.append(('x_{0}'.format(i), (x, lod)))
+            outs.append(out)
+            i = i + 1
+
+        self.inputs = {'X': inputs, "CVM": cvm}
+        self.outputs = {'Out': np.concatenate(outs, axis=self.axis)}
+        self.attrs = {
+            'pooltype': self.pooltype,
+            'axis': self.axis,
+        }
+
+    def set_pooltype(self):
+        self.pooltype = "SUM"
+
+    def set_conf(self):
+        pass
+
+    def test_check_output(self):
+        self.check_output()
+
+
+class TestFusionSeqPoolCVMConcatOpCase1(TestFusionSeqPoolCVMConcatOp):
+    def set_conf(self):
+        self.lods = [[[1]]]
+
+
+class TestFusionSeqPoolCVMConcatOpCase2(TestFusionSeqPoolCVMConcatOp):
+    def set_conf(self):
+        self.lods = [[[1]], [[1]], [[1]]]
+
+
+class TestFusionSeqPoolCVMConcatOpCase3(TestFusionSeqPoolCVMConcatOp):
+    def set_conf(self):
+        self.lods = [[[1, 3, 4, 6]]]
+        self.w = 10
+
+
+class TestFusionSeqPoolCVMConcatOpCase4(TestFusionSeqPoolCVMConcatOp):
+    def set_conf(self):
+        self.lods = [[[2, 13, 4]], [[1, 1, 1]], [[5, 3, 1]], [[9, 10, 3]]]
+        self.w = 3
+
+
+## test avg pool and sqrt
+def create_test_avg_sqrt_class(parent):
+    class TestSeqPoolAvgCase(parent):
+        def set_pooltype(self):
+            self.pooltype = "AVERAGE"
+
+    class TestSeqPoolSqrtCase(parent):
+        def set_pooltype(self):
+            self.pooltype = "SQRT"
+
+    cls_name_avg = "{0}_{1}".format(parent.__name__, "avg")
+    cls_name_sqrt = "{0}_{1}".format(parent.__name__, "sqrt")
+    TestSeqPoolAvgCase.__name__ = cls_name_avg
+    TestSeqPoolSqrtCase.__name__ = cls_name_sqrt
+    globals()[cls_name_avg] = TestSeqPoolAvgCase
+    globals()[cls_name_sqrt] = TestSeqPoolSqrtCase
+
+
+create_test_avg_sqrt_class(TestFusionSeqPoolCVMConcatOp)
+create_test_avg_sqrt_class(TestFusionSeqPoolCVMConcatOpCase1)
+create_test_avg_sqrt_class(TestFusionSeqPoolCVMConcatOpCase2)
+create_test_avg_sqrt_class(TestFusionSeqPoolCVMConcatOpCase3)
+create_test_avg_sqrt_class(TestFusionSeqPoolCVMConcatOpCase4)
+
+if __name__ == '__main__':
+    unittest.main()