grad_node_info.h

// Copyright (c) 2021 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/eager/eager_tensor.h"
#include "paddle/fluid/eager/hooks.h"
#include "paddle/phi/api/all.h"

namespace egr {
/**
 * GradNodeBase is base class of all grad node, which is what should be used by
 * eager execution, we define most of backward autograd members here, and for
 * each Operator, they should hold their onw forward Inputs as TensorWrapper.
 *
 * The GradNodeBase will be held in autograd_meta, and it is also a member of
 * Edge, which indicates the edge of backward graph.
 *
 * TODO:(yangzhanlue) GradNodeBase will also in charge of get the correct input
 * from GradOpDescMaker to GradNodeBase.
 *
 * NOTE:GradNodeBase has a method named run, this method should be overrided by
 * the
 * specific derived class, it will prepare backward inputs and double backward's
 * depends. Then, it will call C++ API of backward kernel functions to finish
 * backward computation.
 *
 * NOTE:GradNodeBase holds its own inputs and Outputs
 *
 * Edge is defined to descripe depend of backward, an Edge is what linked
 * between two
 * node, it should contain a Node and rank of this Node (this is used to
 * indicate which
 * input of grad this edge belong).
 * */
class Edge;
class AutogradMeta;

/**
 * GradSlotMeta is used to Record Forward Tensor info to backward, since paddle
 * has lots of operators
 * whose backward logic is depends on if it has some specific inputs or outputs.
 * So, we need a meta info
 * to record it's needs.
 * **/
class GradSlotMeta {
 public:
  GradSlotMeta() = default;
  bool IsStopGradient() const { return stop_gradient_; }
  void SetStopGradient(bool stop_gradient = true) {
    stop_gradient_ = stop_gradient;
  }

  void SetTensorMeta(const phi::DenseTensorMeta& meta) {
    meta_ = std::make_shared<phi::DenseTensorMeta>(meta);
  }
  bool HasTensorMeta() const { return meta_ && meta_.get(); }
  const phi::DenseTensorMeta& GetTensorMeta() const {
    if (!HasTensorMeta()) {
      PADDLE_THROW(paddle::platform::errors::Fatal(
          "meta_ of GradSlotMeta has not been initialized yet."
          "You're expected to check Edge availability with HasTensorMeta()"
          "before calling GetTensorMeta() interface."));
    }
    return *meta_.get();
  }

 private:
  bool stop_gradient_{false};
  std::shared_ptr<phi::DenseTensorMeta> meta_ = nullptr;
};

class GradNodeBase {
 public:
  GradNodeBase() { VLOG(6) << "Construct GradNodeBase"; }
  GradNodeBase(size_t bwd_in_slot_num, size_t bwd_out_slot_num);
  // TODO(jiabin): Should we have other constructor here?
  virtual ~GradNodeBase() { VLOG(6) << "Destruct GradNodeBase"; }

  /**
   * operator() designed to contian the real backward execution logic, it should
   * be
   * overrided by derived class defined for each operator. It accepts a vector
   * of
   * Tensor which contains grads input of current operator
   *
   * Note: why we need backward inputs and outputs construct as vector of vector
   * of paddle::experimental::Tensor?
   * Since all of paddle op composite in form of {"Slot name ", vector<Var>},
   * so, vector of vector
   * is better choice to fit this format.
   * **/
  virtual std::vector<std::vector<paddle::experimental::Tensor>> operator()(
      const std::vector<std::vector<paddle::experimental::Tensor>>& grads,
      bool create_graph = false) = 0;

  virtual void ClearTensorWrappers() = 0;

  virtual bool IsTensorWrappersCleared() = 0;
  /**
   * AddEdges is designed to set input tensors' backward Node as current
   * node's Edges.
   * This method should be call in forward code and for double backward depends
   * computation.
   *
   * This one is called slot by slot
   * **/
  void AddEdges(std::vector<AutogradMeta*>* metas, size_t slot_id);
  void AddEdges(AutogradMeta* meta, size_t slot_id);

  // adj_edges were moved inside OutputMeta(), so no available direct access
  // from GradNodeBase.
  // To access Edges, get GradSlotMeta by calling OutputMeta(), then use
  // slot_meta.GetEdge()

  /**
   * Get Input Meta of current Grad node**/
  const std::vector<std::vector<GradSlotMeta>>& InputMeta() const;
  /**
   * Get Output Meta of current Grad node**/
  const std::vector<std::vector<GradSlotMeta>>& OutputMeta() const;
  /**
   * Set bwd ins and outs info with forward vars
   * **/

  void SetGradInMeta(const std::vector<paddle::experimental::Tensor>& fwd_out,
                     size_t slot_rank);
  void SetGradInMeta(const paddle::experimental::Tensor& fwd_out,
                     size_t slot_rank);

  void SetGradOutMeta(const std::vector<paddle::experimental::Tensor>& fwd_in,
                      size_t slot_rank);
  void SetGradOutMeta(const paddle::experimental::Tensor& fwd_in,
                      size_t slot_rank);

  /**
   * Default setters for Grad in/out meta this should be used for same special
   * Node which will not create by user
   * **/
  void SetDefaultGradInOutMeta();
  /**
   * Register GradientHook
   * **/
  int64_t RegisterGradientHook(size_t slot_id, size_t rank,
                               std::shared_ptr<egr::TensorHook>&& hook);

  /**
  * Remove GradientHook
  * **/
  bool RemoveGradientHook(const int64_t& hook_id) {
    auto remove_cnt = gradient_hooks_.erase(hook_id);
    if (remove_cnt == 0) {
      return false;
    }
    return true;
  }

  /**
   * Apply GradientHook
   * **/
  inline bool GradientHooksRegistered() { return !gradient_hooks_.empty(); }

  std::vector<std::vector<paddle::experimental::Tensor>> ApplyGradientHooks(
      const std::vector<std::vector<paddle::experimental::Tensor>>& tensors);

  /**
    * Handle Complex - Real Type Promotion
    * **/
  void HandleComplexGradToRealGrad(
      std::vector<std::vector<paddle::experimental::Tensor>>* out_grads);
  bool NeedComplexToRealConversion() { return need_complex_to_real_; }

  virtual std::string name() { return "GradNodeBase"; }

  /**
       * GetEdges is designed to get all edges of current node**/
  const std::vector<std::vector<Edge>>& GetEdges() const;

 private:
  // TODO(zhanlve): Merge adj_edges_ into GradOutMeta
  // Edges recorded the backward related node info, which indicate all edges
  // linked
  // by this Grad Node.
  // Why we need vector<vector<Edge>>: Edges is as same rank as bwd output.
  std::vector<std::vector<Edge>> adj_edges_;

  // bwd_out_meta_ is used to record Grad output info for backward
  std::vector<std::vector<GradSlotMeta>> bwd_out_meta_;

  // bwd_in_meta_ used to record Grad input info for backward
  std::vector<std::vector<GradSlotMeta>> bwd_in_meta_;
  // Gradient Hooks
  // Customer may register a list of hooks which will be called in order during
  // backward
  // Each entry consists one pair of
  // <hook_id, <out_rank, std::shared_ptr<TensorHook>>>
  std::map<int64_t, std::tuple<
                        /* slot id */ size_t, /* rank */ size_t,
                        /* hook */ std::shared_ptr<TensorHook>>>
      gradient_hooks_;

  // We handle complex to real conversion only if any complex GradIn is involved
  bool need_complex_to_real_ = false;
  int64_t next_hook_id_{0};
};

class Edge {
 public:
  // Default constructor for Edges in order to construct it for AutogradMeta
  Edge() : in_slot_id_(0), in_rank_(0), grad_node_(nullptr) {}

  // In real use cases we should create Edge from grad node and input rank which
  // indicate which edge it is.
  // Since we have slot design in operators we will have to locate an edge with
  // slot
  // and rank.
  Edge(const std::shared_ptr<GradNodeBase>& grad_node, size_t in_slot_id,
       size_t in_rank)
      : in_slot_id_(in_slot_id), in_rank_(in_rank), grad_node_(grad_node) {}

  Edge(const std::shared_ptr<GradNodeBase>& grad_node,
       const std::pair</* slot_id */ size_t, /* rank */ size_t>& rank_info)
      : in_slot_id_(rank_info.first),
        in_rank_(rank_info.second),
        grad_node_(grad_node) {}

  GradNodeBase* GetGradNode() const { return grad_node_.get(); }

  std::shared_ptr<GradNodeBase> GetMutableGradNode() const {
    return grad_node_;
  }

  std::pair<size_t, size_t> GetEdgeRankInfo() const {
    return std::make_pair(in_slot_id_, in_rank_);
  }

  void SetEdgeRankInfo(size_t slot_id, size_t in_rank) {
    in_slot_id_ = slot_id;
    in_rank_ = in_rank;
  }

  void SetEdgeRankInfo(
      const std::pair</* slot_id */ size_t, /* rank */ size_t>& edge_rank) {
    in_slot_id_ = edge_rank.first;
    in_rank_ = edge_rank.second;
  }

  // Currently we use grad_node_ to identify if a edge is initialized.
  bool IsInitialized() const { return grad_node_.get(); }

 private:
  size_t in_slot_id_;
  size_t in_rank_;
  std::shared_ptr<GradNodeBase> grad_node_;
};

}  // namespace egr