ps_roi_align.cpp

#include "ps_roi_align.h"

#include <torch/autograd.h>
#include <torch/types.h>

#if defined(WITH_CUDA) || defined(WITH_HIP)
#include <ATen/autocast_mode.h>
#endif

namespace vision {
namespace ops {

std::tuple<at::Tensor, at::Tensor> ps_roi_align(
    const at::Tensor& input,
    const at::Tensor& rois,
    double spatial_scale,
    int64_t pooled_height,
    int64_t pooled_width,
    int64_t sampling_ratio) {
  static auto op = c10::Dispatcher::singleton()
                       .findSchemaOrThrow("torchvision::ps_roi_align", "")
                       .typed<decltype(ps_roi_align)>();
  return op.call(
      input, rois, spatial_scale, pooled_height, pooled_width, sampling_ratio);
}

#if defined(WITH_CUDA) || defined(WITH_HIP)
std::tuple<at::Tensor, at::Tensor> ps_roi_align_autocast(
    const at::Tensor& input,
    const at::Tensor& rois,
    double spatial_scale,
    int64_t pooled_height,
    int64_t pooled_width,
    int64_t sampling_ratio) {
  c10::impl::ExcludeDispatchKeyGuard no_autocast(c10::DispatchKey::Autocast);
  auto result = ps_roi_align(
      at::autocast::cached_cast(at::kFloat, input),
      at::autocast::cached_cast(at::kFloat, rois),
      spatial_scale,
      pooled_height,
      pooled_width,
      sampling_ratio);

  return std::make_tuple(
      std::get<0>(result).to(input.scalar_type()),
      std::get<1>(result).to(input.scalar_type()));
}

TORCH_LIBRARY_IMPL(torchvision, Autocast, m) {
  m.impl("ps_roi_align", ps_roi_align_autocast);
}
#endif

at::Tensor _ps_roi_align_backward(
    const at::Tensor& grad,
    const at::Tensor& rois,
    const at::Tensor& channel_mapping,
    double spatial_scale,
    int64_t pooled_height,
    int64_t pooled_width,
    int64_t sampling_ratio,
    int64_t batch_size,
    int64_t channels,
    int64_t height,
    int64_t width) {
  static auto op =
      c10::Dispatcher::singleton()
          .findSchemaOrThrow("torchvision::_ps_roi_align_backward", "")
          .typed<decltype(_ps_roi_align_backward)>();
  return op.call(
      grad,
      rois,
      channel_mapping,
      spatial_scale,
      pooled_height,
      pooled_width,
      sampling_ratio,
      batch_size,
      channels,
      height,
      width);
}

TORCH_LIBRARY_FRAGMENT(torchvision, m) {
  m.def(
      "ps_roi_align(Tensor input, Tensor rois, float spatial_scale, int pooled_height, int pooled_width, int sampling_ratio) -> (Tensor, Tensor)");
  m.def(
      "_ps_roi_align_backward(Tensor grad, Tensor rois, Tensor channel_mapping, float spatial_scale, int pooled_height, int pooled_width, int sampling_ratio, int batch_size, int channels, int height, int width) -> Tensor");
}

namespace {

class PSROIAlignFunction
    : public torch::autograd::Function<PSROIAlignFunction> {
 public:
  static torch::autograd::variable_list forward(
      torch::autograd::AutogradContext* ctx,
      const torch::autograd::Variable& input,
      const torch::autograd::Variable& rois,
      double spatial_scale,
      int64_t pooled_height,
      int64_t pooled_width,
      int64_t sampling_ratio) {
    ctx->saved_data["spatial_scale"] = spatial_scale;
    ctx->saved_data["pooled_height"] = pooled_height;
    ctx->saved_data["pooled_width"] = pooled_width;
    ctx->saved_data["sampling_ratio"] = sampling_ratio;
    ctx->saved_data["input_shape"] = input.sizes();
    at::AutoNonVariableTypeMode g;
    auto result = ps_roi_align(
        input,
        rois,
        spatial_scale,
        pooled_height,
        pooled_width,
        sampling_ratio);

    auto output = std::get<0>(result);
    auto channel_mapping = std::get<1>(result);
    ctx->save_for_backward({rois, channel_mapping});
    ctx->mark_non_differentiable({channel_mapping});

    return {output, channel_mapping};
  }

  static torch::autograd::variable_list backward(
      torch::autograd::AutogradContext* ctx,
      const torch::autograd::variable_list& grad_output) {
    // Use data saved in forward
    auto saved = ctx->get_saved_variables();
    auto rois = saved[0];
    auto channel_mapping = saved[1];
    auto input_shape = ctx->saved_data["input_shape"].toIntList();
    auto grad_in = _ps_roi_align_backward(
        grad_output[0],
        rois,
        channel_mapping,
        ctx->saved_data["spatial_scale"].toDouble(),
        ctx->saved_data["pooled_height"].toInt(),
        ctx->saved_data["pooled_width"].toInt(),
        ctx->saved_data["sampling_ratio"].toInt(),
        input_shape[0],
        input_shape[1],
        input_shape[2],
        input_shape[3]);

    return {grad_in,
            torch::autograd::Variable(),
            torch::autograd::Variable(),
            torch::autograd::Variable(),
            torch::autograd::Variable(),
            torch::autograd::Variable()};
  }
};

// TODO: There should be an easier way to do this
class PSROIAlignBackwardFunction
    : public torch::autograd::Function<PSROIAlignBackwardFunction> {
 public:
  static torch::autograd::variable_list forward(
      torch::autograd::AutogradContext* ctx,
      const torch::autograd::Variable& grad,
      const torch::autograd::Variable& rois,
      const torch::autograd::Variable& channel_mapping,
      double spatial_scale,
      int64_t pooled_height,
      int64_t pooled_width,
      int64_t sampling_ratio,
      int64_t batch_size,
      int64_t channels,
      int64_t height,
      int64_t width) {
    at::AutoNonVariableTypeMode g;
    auto grad_in = _ps_roi_align_backward(
        grad,
        rois,
        channel_mapping,
        spatial_scale,
        pooled_height,
        pooled_width,
        sampling_ratio,
        batch_size,
        channels,
        height,
        width);

    return {grad_in};
  }

  static torch::autograd::variable_list backward(
      torch::autograd::AutogradContext* ctx,
      const torch::autograd::variable_list& grad_output) {
    TORCH_CHECK(0, "double backwards on ps_roi_align not supported");
  }
};

} // namespace

std::tuple<at::Tensor, at::Tensor> ps_roi_align_autograd(
    const at::Tensor& input,
    const at::Tensor& rois,
    double spatial_scale,
    int64_t pooled_height,
    int64_t pooled_width,
    int64_t sampling_ratio) {
  auto result = PSROIAlignFunction::apply(
      input, rois, spatial_scale, pooled_height, pooled_width, sampling_ratio);

  return std::make_tuple(result[0], result[1]);
}

at::Tensor ps_roi_align_backward_autograd(
    const at::Tensor& grad,
    const at::Tensor& rois,
    const at::Tensor& channel_mapping,
    double spatial_scale,
    int64_t pooled_height,
    int64_t pooled_width,
    int64_t sampling_ratio,
    int64_t batch_size,
    int64_t channels,
    int64_t height,
    int64_t width) {
  return PSROIAlignBackwardFunction::apply(
      grad,
      rois,
      channel_mapping,
      spatial_scale,
      pooled_height,
      pooled_width,
      sampling_ratio,
      batch_size,
      channels,
      height,
      width)[0];
}

TORCH_LIBRARY_IMPL(torchvision, Autograd, m) {
  m.impl("ps_roi_align", ps_roi_align_autograd);
  m.impl("_ps_roi_align_backward", ps_roi_align_backward_autograd);
}

} // namespace ops
} // namespace vision