cond op

paddle/operators/CMakeLists.txt

@@ -55,12 +55,14 @@ set(DEPS_OPS
     minus_op
     mul_op
     recurrent_op
+    cond_op
     scale_op)
 op_library(identity_op DEPS scale_op)
 op_library(minus_op DEPS scale_op)
 op_library(mul_op DEPS math_function)
 op_library(recurrent_op SRCS recurrent_op.cc rnn/recurrent_op_utils.cc
   DEPS framework_proto tensor operator net_op)
+op_library(cond_op SRCS cond_op.cc DEPS framework_proto tensor operator net_op)
 op_library(scale_op DEPS net_op)
 
 list(REMOVE_ITEM GENERAL_OPS ${DEPS_OPS})

paddle/operators/cond_op.cc

@@ -0,0 +1,45 @@
+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+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/operators/cond_op.h"
+#include "paddle/framework/op_registry.h"
+#include "paddle/operators/net_op.h"
+
+namespace paddle {
+namespace operators {
+
+class CondOpProtoAndCheckerMaker : public OpProtoAndCheckerMaker {
+ public:
+  CondOpProtoAndCheckerMaker(OpProto *proto, OpAttrChecker *op_checker)
+      : OpProtoAndCheckerMaker(proto, op_checker) {
+    AddInput("Cond", "The condition, which is a bool vector");
+    AddInput("Xs", "Inputs of Subnets").AsDuplicable();
+    AddOutput("Outs", "Outputs of Cond_Op after merge").AsDuplicable();
+
+    AddOutput("SubScopes", "sub scopes for true and false branches");
+    AddOutput("IndexTensors", "Index Tensors contains indices for true/false");
+
+    AddComment(R"DOC(
+Sample dependent Cond Operator:
+The equation is: Out[i] = subnet_t[i], if Cond[i] == true
+Out[i] = subnet_t[i], if Cond[i] == false
+)DOC");
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+REGISTER_OP_WITHOUT_GRADIENT(cond_op, paddle::operators::CondOp,
+                             paddle::operators::CondOpProtoAndCheckerMaker);

paddle/operators/cond_op.h

@@ -0,0 +1,232 @@
+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+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 <vector>
+#include "glog/logging.h"
+#include "paddle/framework/ddim.h"
+#include "paddle/framework/eigen.h"
+#include "paddle/framework/operator.h"
+#include "paddle/framework/tensor.h"
+#include "paddle/operators/gather.h"
+#include "paddle/operators/scatter.h"
+
+namespace paddle {
+namespace operators {
+
+using namespace paddle::framework;
+
+class CondOp : public OperatorBase {
+ public:
+  CondOp(const std::string& type, const VariableNameMap& inputs,
+         const VariableNameMap& outputs, const AttributeMap& attrs)
+      : OperatorBase(type, inputs, outputs, attrs) {
+    index_.resize(2);
+    sub_net_op_.resize(2);
+    LOG(INFO) << "Initialization Done.";
+  }
+
+  CondOp(const CondOp& o)
+      : framework::OperatorBase(
+            static_cast<const framework::OperatorBase&>(o)) {
+    // TODO(yuyang18): Implement copy ctor well.
+    PADDLE_THROW("Not implemented");
+  }
+
+  void CreateScope(const Scope& scope) const {
+    auto sub_scopes_var = scope.FindVar("SubScopes");
+    PADDLE_ENFORCE(sub_scopes_var != nullptr, "");
+    auto sub_scopes = sub_scopes_var->GetMutable<std::vector<Scope*>>();
+    auto& sub_scope = scope.NewScope();
+    sub_scopes->push_back(&sub_scope);
+  }
+
+  void CreateIndexTensor(const Scope& scope) const {
+    auto index_tensors_var = scope.FindVar("IndexTensors");
+    PADDLE_ENFORCE(index_tensors_var != nullptr, "");
+    auto& index_tensors =
+        *index_tensors_var->GetMutable<std::vector<Tensor*>>();
+    Tensor index_tensor;
+    index_tensors.push_back(&index_tensor);
+  }
+
+  /**
+   * InferShape must be called before Run.
+   */
+  void InferShape(const framework::Scope& scope) const override {
+    auto sub_scopes_var = scope.FindVar("SubScopes");
+    PADDLE_ENFORCE_NOT_NULL(sub_scopes_var);
+    auto& sub_scopes = *sub_scopes_var->GetMutable<std::vector<Scope*>>();
+    // auto& index_tensors =
+    // *scope.FindVar("IndexTensors")->GetMutable<std::vector<Tensor*>>();
+
+    for (int i = 0; i < 2; ++i) {
+      // Create two sub scopes for true and false branches
+      // sub_scopes[0] for the true branch and sub_scopes[1] for the false
+      // branch
+      CreateScope(scope);
+
+      // Create two tensors for true and false indices
+      // index_tensors[0] for the true branch and index_tensors[1] for the false
+      // branch
+      CreateIndexTensor(scope);
+
+      for (auto& input : Inputs("Xs")) {
+        // Create a new tensor in sub-scope for input-type tensor
+        Variable* v = sub_scopes[i]->NewVar(input);
+        Tensor* sub_input = v->GetMutable<Tensor>();
+        sub_input->Resize(scope.FindVar(input)->GetMutable<Tensor>()->dims());
+      }
+
+      // Inputs that do not require tailoring
+      /*for (auto& input : (*sub_net_op_[i]).Inputs()) {
+        // weights are located in the parent scope rather than sub scope
+        for (auto& var_name : input.second) {
+          if (!sub_scopes[i]->FindVar(var_name)) {
+            sub_scopes[i]->NewVar(var_name)->GetMutable<Tensor>();
+          }
+        }
+      }*/
+
+      // Outputs
+      for (auto& output : (*sub_net_op_[i]).Outputs()) {
+        for (auto& var_name : output.second) {
+          sub_scopes[i]->NewVar(var_name);
+        }
+      }
+
+      // each net calls InferShape
+      LOG(INFO) << "OK 3";
+      sub_net_op_[i]->InferShape(*sub_scopes[i]);
+      LOG(INFO) << "OK 4";
+    }
+
+    for (auto& output : Outputs("Outs")) {
+      Tensor* tensor_t_out =
+          sub_scopes[0]->FindVar(output)->GetMutable<Tensor>();
+      Tensor* tensor_f_out =
+          sub_scopes[1]->FindVar(output)->GetMutable<Tensor>();
+      Tensor* tensor_out = scope.FindVar(output)->GetMutable<Tensor>();
+      // check output size should be same
+      PADDLE_ENFORCE_EQ(tensor_t_out->dims(), tensor_f_out->dims(),
+                        "Outputs not of the same shape");
+      tensor_out->Resize(tensor_t_out->dims());
+    }
+    LOG(INFO) << "OK 5";
+  }
+
+  // Set True Block
+  void set_truenet(std::unique_ptr<OperatorBase> net) {
+    sub_net_op_[0] = std::move(net);
+  }
+
+  // Set False Block
+  void set_falsenet(std::unique_ptr<OperatorBase> net) {
+    sub_net_op_[1] = std::move(net);
+  }
+
+  void Run(const framework::Scope& scope,
+           const platform::DeviceContext& dev_ctx) const override {
+    auto sub_scopes = scope.FindVar("SubScopes")->Get<std::vector<Scope*>>();
+    auto index_tensors =
+        scope.FindVar("IndexTensors")->Get<std::vector<Tensor*>>();
+
+    std::string cond_name = Input("Cond");
+    Variable* cond_var = scope.FindVar(cond_name);
+    PADDLE_ENFORCE_NOT_NULL(cond_var)
+    const Tensor* cond = cond_var->GetMutable<Tensor>();
+
+    // Step 1: get the true/false index at runtime
+    // index_[0]: vector<int>, contains all index for cond[i] == true
+    // index_[1]: vector<int>, contains all index for cond[i] == false
+    for (int i = 0; i < 2; ++i) index_[i].clear();
+
+    const bool* cond_data = cond->data<bool>();
+    for (int i = 0; i < cond->dims()[0]; ++i) {
+      if (cond_data[i])
+        index_[0].push_back(i);
+      else
+        index_[1].push_back(i);
+    }
+    // put index_[0] and index_[1] into two tensors:
+    // index_tensor_[0] and index_tensor_[1]
+    framework::DDim dim = paddle::framework::make_ddim({0});
+    for (int i = 0; i < 2; ++i) {
+      dim[0] = index_[i].size();
+      int* tmp_ptr =
+          index_tensors[i]->mutable_data<int>(dim, platform::CPUPlace());
+      index_tensors[i]->Resize(dim);
+      memcpy(tmp_ptr, index_[i].data(), dim[0] * sizeof(int));
+    }
+
+    // Step 2: collect data by calling gather
+    for (int i = 0; i < 2; ++i) {
+      // i= 0/i for True and False branches respectively
+      for (auto& input : Inputs("Xs")) {
+        // find Tensor
+        // Tensor* tensor_parent = scope.FindVar(input)->GetMutable<Tensor>();
+        Variable* v = scope.FindVar(input);
+        Tensor* tensor_parent = v->GetMutable<Tensor>();
+        // Tensor* tensor_child =
+        // sub_scope_[i].FindVar(input)->GetMutable<Tensor>();
+        v = sub_scopes[i]->FindVar(input);
+        Tensor* tensor_child = v->GetMutable<Tensor>();
+        Gather<float>(dev_ctx.GetPlace(), tensor_parent, index_tensors[i],
+                      tensor_child);
+      }
+    }
+
+    // Step 3: run
+    for (int i = 0; i < 2; ++i) sub_net_op_[i]->Run(*sub_scopes[i], dev_ctx);
+
+    // Step 4: merge output results
+    for (int i = 0; i < 2; ++i) {
+      // i= 0/i for True and False branches respectively
+      // for (auto& output : GetAttr<std::vector<std::string>>("sub_outputs")) {
+      for (auto& output : Outputs("Outs")) {
+        // find Tensor
+        Variable* v = scope.FindVar(output);
+        Tensor* tensor_parent = v->GetMutable<Tensor>();
+        v = sub_scopes[i]->FindVar(output);
+        Tensor* tensor_child = v->GetMutable<Tensor>();
+        ScatterUpdate<float>(dev_ctx.GetPlace(), tensor_child, index_tensors[i],
+                             tensor_parent);
+      }
+    }
+  }
+
+ private:
+  // sub_net_op_[0]: subnet_t
+  // sub_net_op_[1]: subnet_f
+  std::vector<std::unique_ptr<framework::OperatorBase>> sub_net_op_;
+
+  // index_[0]: True_index;
+  // index_[1]: False_index;
+  mutable std::vector<std::vector<int>> index_;
+};
+
+/*
+class CondGradientOp final : public OperatorBase {
+public:
+        void Init() override;
+
+        virtual void InferShape(const std::shared_ptr<Scope>& scope) const
+override;
+
+        virtual void Run(const std::shared_ptr<Scope>& scope,
+                   const platform::DeviceContext& dev_ctx) const override;
+};*/
+
+}  // namespace operators
+}  // namespace paddle

paddle/pybind/pybind.cc

@@ -41,6 +41,7 @@ USE_OP(softmax);
 USE_OP(rowwise_add);
 USE_OP(fill_zeros_like);
 USE_NO_KERNEL_OP(recurrent);
+USE_NO_KERNEL_OP(cond);
 USE_OP(gaussian_random);
 USE_OP(uniform_random);
 USE_OP(lookup_table);
@@ -324,6 +325,28 @@ All parameter, weight, gradient are variables in Paddle.
            [](operators::RecurrentOp &self, const operators::NetOp &net)
                -> void { self.set_stepnet(net.Clone()); });
 
+  // cond_op
+  py::class_<operators::CondOp, OperatorBase>(m, "CondOp")
+      .def_static("create",
+                  [](py::bytes protobin) -> operators::CondOp * {
+                    OpDesc desc;
+                    PADDLE_ENFORCE(desc.ParsePartialFromString(protobin),
+                                   "Cannot parse user input to OpDesc");
+                    PADDLE_ENFORCE(desc.IsInitialized(),
+                                   "User OpDesc is not initialized, reason %s",
+                                   desc.InitializationErrorString());
+                    auto cond_op = OpRegistry::CreateOp(desc);
+                    return static_cast<operators::CondOp *>(cond_op.release());
+                  })
+      .def("set_truenet",
+           [](operators::CondOp &self, const operators::NetOp &net) -> void {
+             self.set_truenet(net.Clone());
+           })
+      .def("set_falsenet",
+           [](operators::CondOp &self, const operators::NetOp &net) -> void {
+             self.set_falsenet(net.Clone());
+           });
+
   m.def("unique_integer", UniqueIntegerGenerator);
 
   m.def("is_compile_gpu", IsCompileGPU);

python/paddle/v2/framework/op.py

@@ -215,5 +215,27 @@ class __RecurrentOp__(object):
         return core.RecurrentOp.create(proto.SerializeToString())
 
 
+class __CondOp__(object):
+    __proto__ = None
+    type = 'cond_op'
+
+    def __init__(self):
+        # cache recurrent_op's proto
+        if self.__proto__ is None:
+            for op_proto in get_all_op_protos():
+                if op_proto.type == self.type:
+                    self.__proto__ = op_proto
+
+    def __call__(self, *args, **kwargs):
+        if self.type not in args and 'type' not in kwargs:
+            kwargs['type'] = self.type
+        # create proto
+        create_method = OpDescCreationMethod(self.__proto__)
+        proto = create_method(*args, **kwargs)
+        # create condop
+        return core.CondOp.create(proto.SerializeToString())
+
+
 Operator = OperatorFactory()  # The default global factory
 RecurrentOp = __RecurrentOp__()
+CondOp = __CondOp__()

python/paddle/v2/framework/tests/test_cond_op.py

@@ -0,0 +1,114 @@
+import logging
+import paddle.v2.framework.core as core
+import unittest
+import numpy as np
+from paddle.v2.framework.op import Operator, CondOp
+
+
+class PySimpleCond(object):
+    '''
+    A simple implementation of dynamic if-else based on numpy
+    '''
+
+    def __init__(self):
+        array = [True] * 10
+        for i in range(1, 10, 2):
+            array[i] = False
+        self.cond = np.array(array)
+        self.x = np.ones(shape=(10, 1))
+
+    def forward(self):
+        self.index_t = np.where(self.cond)
+        self.index_f = np.where(self.cond == False)
+        y_t = self.x[self.index_t]
+        y_f = self.x[self.index_f]
+        y_t = y_t * 2.
+        y_f = y_f * (-2.)
+        output = np.zeros(shape=(10, 1))
+        output[self.index_t] = y_t
+        output[self.index_f] = y_f
+        return output
+
+
+class PySimpleCondTest(unittest.TestCase):
+    def setUp(self):
+        self.condnn = PySimpleCond()
+
+    def test_forward(self):
+        output = self.condnn.forward()
+        print 'output', output
+
+
+def create_tensor(scope, name, shape, np_data):
+    tensor = scope.new_var(name).get_tensor()
+    tensor.set_dims(shape)
+    tensor.set(np_data, core.CPUPlace())
+    return tensor
+
+
+class TestCondOp(unittest.TestCase):
+    '''
+    Test CondOp
+
+    equation:
+        cond = [True, False, True, False, ...]
+        y[index_t] = x[index_t] * 2.
+        y[index_f] = x[index_f] * -2.
+    outputs:
+        y
+    '''
+
+    def setUp(self):
+        self.py_cond = PySimpleCond()
+
+    def forward(self):
+        self.scope = core.Scope()
+        self.create_global_variables()
+        self.create_cond_op()
+        self.create_sub_net()
+        ctx = core.DeviceContext.create(core.CPUPlace())
+        print 'running infer shape'
+        print self.scope.find_var("SubScopes")
+        self.condop.infer_shape(self.scope)
+        print 'ok 2'
+        self.condop.run(self.scope, ctx)
+        print 'ok 3'
+        return np.array(self.scope.find_var("Outs").get_tensor())
+
+    def create_global_variables(self):
+        x_np_data = self.py_cond.x
+        create_tensor(self.scope, "x", [10, 1], x_np_data)
+        cond_np_data = self.py_cond.cond
+        create_tensor(self.scope, "cond", [10, 1], x_np_data)
+        self.scope.new_var("SubScopes")
+        self.scope.new_var("IndexTensors")
+        self.scope.new_var("Outs")
+
+    def create_cond_op(self):
+        self.condop = CondOp(
+            Cond="cond",
+            Xs=["x"],
+            Outs=['Out_final'],
+            SubScopes="SubScopes",
+            IndexTensors="IndexTensors")
+
+    def create_sub_net(self):
+        truenet = core.Net.create()
+        scale_op_t = Operator("scale", X='X', Y='Out', scale=2.)
+        truenet.append_op(scale_op_t)
+        truenet.complete_add_op(True)
+        self.condop.set_truenet(truenet)
+
+        falsenet = core.Net.create()
+        scale_op_t = Operator("scale", X='X', Y='Out', scale=-2.)
+        falsenet.append_op(scale_op_t)
+        falsenet.complete_add_op(True)
+        self.condop.set_falsenet(falsenet)
+
+    def test_forward(self):
+        print 'test cond op forward'
+        py_output = self.forward()
+
+
+if __name__ == "__main__":
+    unittest.main()