operator.h 5.5 KB
Newer Older
Q
Qiao Longfei 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
/* 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

Y
Yu Yang 已提交
17 18 19
#include <paddle/framework/attr_checker.h>
#include <paddle/framework/op_desc.pb.h>
#include <paddle/framework/scope.h>
Y
Yu Yang 已提交
20
#include <paddle/framework/tensor.h>
Y
Yu Yang 已提交
21 22 23
#include <paddle/platform/device_context.h>
#include <paddle/platform/place.h>
#include <paddle/utils/Error.h>
Q
Qiao Longfei 已提交
24 25 26 27 28 29 30 31 32
#include <boost/variant.hpp>
#include <string>
#include <unordered_map>
#include <vector>

namespace paddle {
namespace framework {

class OperatorBase;
Q
Qiao Longfei 已提交
33
using OperatorPtr = std::shared_ptr<OperatorBase>;
Q
Qiao Longfei 已提交
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
/**
 * OperatorBase has the basic element that Net will call to do computation.
 * Only CreateOperator from OpRegistry will new Operator directly. User
 * should always construct a proto message OpDesc and call
 * OpRegistry::CreateOp(op_desc) to get an Operator instance.
 */
class OperatorBase {
 public:
  virtual ~OperatorBase() {}

  template <typename T>
  inline const T& GetAttr(const std::string& name) const {
    PADDLE_ENFORCE(attrs_.count(name) != 0, "%s should be in AttributeMap",
                   name);
    return boost::get<T>(attrs_.at(name));
  }

  std::string DebugString() const;

Q
Qiao Longfei 已提交
53 54 55 56
  /// Init will be called after CreateOperator, you can put some initialization
  /// logic here.
  virtual void Init() {}

Q
Qiao Longfei 已提交
57 58
  /// InferShape infer the size of Variables used by this Operator with
  /// information inside scope
Q
Qiao Longfei 已提交
59
  virtual void InferShape(const ScopePtr& scope) const = 0;
Q
Qiao Longfei 已提交
60 61

  /// Net will call this function to Run an op.
Q
Qiao Longfei 已提交
62
  virtual void Run(const ScopePtr& scope,
Y
Yu Yang 已提交
63 64 65 66
                   const platform::DeviceContext& dev_ctx) const = 0;

 protected:
  std::string Type() const { return desc_.type(); }
Q
Qiao Longfei 已提交
67 68 69 70 71 72 73 74

 public:
  OpDesc desc_;
  std::vector<std::string> inputs_;
  std::vector<std::string> outputs_;
  AttributeMap attrs_;
};

Y
Yu Yang 已提交
75 76 77 78 79 80 81 82 83 84
class OpKernel {
 public:
  /**
   * KernelContext is the only parameter of Kernel Run function.
   * Run will get input/output variables, state such as momentum and
   * device resource such as CUDA stream, cublas handle, etc. from
   * KernelContext. User should construct it before run the Operator.
   */
  class KernelContext {
   public:
Q
Qiao Longfei 已提交
85
    KernelContext(const OperatorBase* op, const ScopePtr& scope,
Y
Yu Yang 已提交
86 87 88 89 90 91 92 93 94 95 96 97
                  const platform::DeviceContext& device_context)
        : op_(*op), scope_(scope), device_context_(device_context) {}

    const Variable* Input(int index) const {
      return scope_->GetVariable(op_.inputs_[index]);
    }

    Variable* Output(int index) const {
      return scope_->GetVariable(op_.outputs_[index]);
    }

    const OperatorBase& op_;
Q
Qiao Longfei 已提交
98
    const ScopePtr& scope_;
Y
Yu Yang 已提交
99 100 101 102 103 104 105 106
    const platform::DeviceContext& device_context_;
  };

  virtual void Compute(const KernelContext& context) const = 0;

  virtual ~OpKernel() {}
};

Y
Yu Yang 已提交
107 108 109 110 111 112 113 114 115 116 117 118 119
template <typename T>
struct VarToTensor {};

template <>
struct VarToTensor<Tensor*> {
  Tensor* operator()(Variable* var) { return var->GetMutable<Tensor>(); }
};

template <>
struct VarToTensor<const Tensor*> {
  const Tensor* operator()(Variable* var) { return &var->Get<Tensor>(); }
};

Q
Qiao Longfei 已提交
120 121
class OperatorWithKernel : public OperatorBase {
 public:
Y
Yu Yang 已提交
122 123
  struct OpKernelKey {
    platform::Place place_;
Q
Qiao Longfei 已提交
124

Y
Yu Yang 已提交
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
    OpKernelKey() = default;
    OpKernelKey(const platform::DeviceContext& dev_ctx) {
      place_ = dev_ctx.GetPlace();
    }

    bool operator==(const OpKernelKey& o) const { return place_ == o.place_; }
  };

  struct OpKernelHash {
    std::hash<bool> hash_;
    size_t operator()(const OpKernelKey& key) const {
      return hash_(platform::is_gpu_place(key.place_));
    }
  };

  using OpKernelMap =
      std::unordered_map<OpKernelKey, std::unique_ptr<OpKernel>, OpKernelHash>;
Q
Qiao Longfei 已提交
142

Q
Qiao Longfei 已提交
143
  void Run(const ScopePtr& scope,
Y
Yu Yang 已提交
144 145 146
           const platform::DeviceContext& dev_ctx) const final {
    auto& opKernel = AllOpKernels().at(Type()).at(OpKernelKey(dev_ctx));
    opKernel->Compute(OpKernel::KernelContext(this, scope, dev_ctx));
Q
Qiao Longfei 已提交
147 148
  }

Y
Yu Yang 已提交
149 150 151 152
  static std::unordered_map<std::string /* op_type */, OpKernelMap>&
  AllOpKernels() {
    static std::unordered_map<std::string, OpKernelMap> g_all_op_kernels;
    return g_all_op_kernels;
Y
Yu Yang 已提交
153 154 155 156 157 158 159
  }
  void InferShape(const std::shared_ptr<Scope>& scope) const final {
    std::vector<const Tensor*> ins;
    VarNamesToTensors(scope, inputs_, &ins);
    std::vector<Tensor*> outs;
    VarNamesToTensors(scope, outputs_, &outs);
    InferShape(ins, outs);
Y
Yu Yang 已提交
160
  };
Y
Yu Yang 已提交
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181

 private:
  template <typename T>
  void VarNamesToTensors(const std::shared_ptr<Scope>& scope,
                         const std::vector<std::string>& var_names,
                         std::vector<T>* container) const {
    container->reserve(var_names.size());
    VarToTensor<T> convert;
    for (auto& name : var_names) {
      auto var = scope->GetVariable(name);
      if (var != nullptr) {
        container->push_back(convert(var));
      } else {
        container->push_back(nullptr);
      }
    }
  }

 protected:
  virtual void InferShape(const std::vector<const Tensor*>& inputs,
                          const std::vector<Tensor*>& outputs) const = 0;
Q
Qiao Longfei 已提交
182 183 184 185
};

}  // namespace framework
}  // namespace paddle