listwise_l2r_op.h

// Copyright 2004-present Facebook. All Rights Reserved.

#pragma once

#include "caffe2/core/context.h"
#include "caffe2/core/logging.h"
#include "caffe2/core/operator.h"
#include "caffe2/utils/math.h"

namespace caffe2 {

template <typename T, class Context>
class LambdaRankNdcgOp final : public Operator<Context> {
 public:
  template <class... Args>
  explicit LambdaRankNdcgOp(Args&&... args)
      : Operator<Context>(std::forward<Args>(args)...),
        use_ndcg_as_loss_(
            this->template GetSingleArgument<bool>("use_ndcg_as_loss", false)),
        use_exp_gain_(
            this->template GetSingleArgument<bool>("use_exp_gain", true)) {}
  USE_OPERATOR_CONTEXT_FUNCTIONS;
  bool RunOnDevice() override;

 private:
  INPUT_TAGS(PRED, REL, SESSION_LENS);
  OUTPUT_TAGS(LOSS, DPRED);

  void ResizeInvLogITensor(int);
  void ComputeDiscounts(int*, int);
  float LambdaRankNdcgSession(
      int start_index,
      int end_index,
      const Tensor& y,
      const Tensor& r,
      Tensor** dy);
  bool use_ndcg_as_loss_;
  bool use_exp_gain_;
  Tensor gain_;
  Tensor discount_;
  Tensor rank_idx_;
  Tensor ideal_idx_;
  Tensor lambda_;
  Tensor inv_log_i_;
};

template <typename T, class Context>
class LambdaRankNdcgGradientOp final : public Operator<Context> {
 public:
  USE_SIMPLE_CTOR_DTOR(LambdaRankNdcgGradientOp);
  USE_OPERATOR_CONTEXT_FUNCTIONS;
  bool RunOnDevice() override;

 private:
  INPUT_TAGS(Y, SESSION_LENS, DY_CACHE, DLOSS);
  OUTPUT_TAGS(DY);
};

} // namespace caffe2