pulsatrix/api/bce__with__logits__loss_8hpp_source.html

#pragma once


#include "pulsatrix/device_backend.hpp"

#include "pulsatrix/tensor.hpp"


namespace pulsatrix {


class BCEWithLogitsLoss {

public:

    explicit BCEWithLogitsLoss(DeviceBackend* backend);


    [[nodiscard]] float forward(const Tensor& logits, const Tensor& target);


    [[nodiscard]] Tensor backward() const;


private:

    DeviceBackend* backend_;

    Tensor last_logits_;

    Tensor last_target_;

    bool has_forwarded_ = false;

};


}  // namespace pulsatrix

pulsatrix::BCEWithLogitsLoss
loss = mean( max(x,0) - x*y + log(1 + exp(-|x|)) ), over all N*k elements of a (N,...
Definition bce_with_logits_loss.hpp:50

pulsatrix::BCEWithLogitsLoss::BCEWithLogitsLoss
BCEWithLogitsLoss(DeviceBackend *backend)
Constructs a BCE-with-logits loss.

pulsatrix::BCEWithLogitsLoss::backward
Tensor backward() const
Gradient w.r.t. the logits: grad[i] = (sigmoid(x[i]) - y[i]) / numel.

pulsatrix::BCEWithLogitsLoss::forward
float forward(const Tensor &logits, const Tensor &target)
Computes the loss value and caches logits/target for backward().

pulsatrix::DeviceBackend
Vendor-agnostic compute/memory backend. CPUBackend, CUDABackend (Phase 1.5), and HIPBackend (Phase 1....
Definition device_backend.hpp:219

pulsatrix::Tensor
N-dimensional tensor. Owns its data buffer exclusively; a DeviceBackend* is injected (not owned) – th...
Definition tensor.hpp:29

device_backend.hpp
Abstract interface isolating vendor-specific memory/compute operations from Tensor/ComputationGraph.

pulsatrix
Definition acquisition_functions.hpp:16

tensor.hpp
N-dimensional tensor – owns a buffer via DeviceBackend*, RAII (Rule of Five).