Skip to content

Add torch2.6 support for ms_deform_attn_cuda #94

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Apr 17, 2025
Merged

Add torch2.6 support for ms_deform_attn_cuda #94

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
49 changes: 31 additions & 18 deletions grounding_dino/groundingdino/models/GroundingDINO/csrc/MsDeformAttn/ms_deform_attn_cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,11 +15,24 @@
#include <ATen/cuda/CUDAContext.h>
#include <cuda.h>
#include <cuda_runtime.h>
#include <torch/extension.h>
#include <torch/version.h>

// Check PyTorch version and define appropriate macros
#if TORCH_VERSION_MAJOR >= 2 && TORCH_VERSION_MINOR >= 6
// PyTorch 2.x and above
#define GET_TENSOR_TYPE(x) x.scalar_type()
#define IS_CUDA_TENSOR(x) x.device().is_cuda()
#else
// PyTorch 1.x
#define GET_TENSOR_TYPE(x) x.type()
#define IS_CUDA_TENSOR(x) x.type().is_cuda()
#endif

namespace groundingdino {

at::Tensor ms_deform_attn_cuda_forward(
const at::Tensor &value,
const at::Tensor &value,
const at::Tensor &spatial_shapes,
const at::Tensor &level_start_index,
const at::Tensor &sampling_loc,
Expand All @@ -32,11 +45,11 @@ at::Tensor ms_deform_attn_cuda_forward(
AT_ASSERTM(sampling_loc.is_contiguous(), "sampling_loc tensor has to be contiguous");
AT_ASSERTM(attn_weight.is_contiguous(), "attn_weight tensor has to be contiguous");

AT_ASSERTM(value.type().is_cuda(), "value must be a CUDA tensor");
AT_ASSERTM(spatial_shapes.type().is_cuda(), "spatial_shapes must be a CUDA tensor");
AT_ASSERTM(level_start_index.type().is_cuda(), "level_start_index must be a CUDA tensor");
AT_ASSERTM(sampling_loc.type().is_cuda(), "sampling_loc must be a CUDA tensor");
AT_ASSERTM(attn_weight.type().is_cuda(), "attn_weight must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(value), "value must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(spatial_shapes), "spatial_shapes must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(level_start_index), "level_start_index must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(sampling_loc), "sampling_loc must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(attn_weight), "attn_weight must be a CUDA tensor");

const int batch = value.size(0);
const int spatial_size = value.size(1);
Expand All @@ -51,7 +64,7 @@ at::Tensor ms_deform_attn_cuda_forward(
const int im2col_step_ = std::min(batch, im2col_step);

AT_ASSERTM(batch % im2col_step_ == 0, "batch(%d) must divide im2col_step(%d)", batch, im2col_step_);

auto output = at::zeros({batch, num_query, num_heads, channels}, value.options());

const int batch_n = im2col_step_;
Expand All @@ -62,7 +75,7 @@ at::Tensor ms_deform_attn_cuda_forward(
for (int n = 0; n < batch/im2col_step_; ++n)
{
auto columns = output_n.select(0, n);
AT_DISPATCH_FLOATING_TYPES(value.type(), "ms_deform_attn_forward_cuda", ([&] {
AT_DISPATCH_FLOATING_TYPES(GET_TENSOR_TYPE(value), "ms_deform_attn_forward_cuda", ([&] {
ms_deformable_im2col_cuda(at::cuda::getCurrentCUDAStream(),
value.data<scalar_t>() + n * im2col_step_ * per_value_size,
spatial_shapes.data<int64_t>(),
Expand All @@ -82,7 +95,7 @@ at::Tensor ms_deform_attn_cuda_forward(


std::vector<at::Tensor> ms_deform_attn_cuda_backward(
const at::Tensor &value,
const at::Tensor &value,
const at::Tensor &spatial_shapes,
const at::Tensor &level_start_index,
const at::Tensor &sampling_loc,
Expand All @@ -98,12 +111,12 @@ std::vector<at::Tensor> ms_deform_attn_cuda_backward(
AT_ASSERTM(attn_weight.is_contiguous(), "attn_weight tensor has to be contiguous");
AT_ASSERTM(grad_output.is_contiguous(), "grad_output tensor has to be contiguous");

AT_ASSERTM(value.type().is_cuda(), "value must be a CUDA tensor");
AT_ASSERTM(spatial_shapes.type().is_cuda(), "spatial_shapes must be a CUDA tensor");
AT_ASSERTM(level_start_index.type().is_cuda(), "level_start_index must be a CUDA tensor");
AT_ASSERTM(sampling_loc.type().is_cuda(), "sampling_loc must be a CUDA tensor");
AT_ASSERTM(attn_weight.type().is_cuda(), "attn_weight must be a CUDA tensor");
AT_ASSERTM(grad_output.type().is_cuda(), "grad_output must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(value), "value must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(spatial_shapes), "spatial_shapes must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(level_start_index), "level_start_index must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(sampling_loc), "sampling_loc must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(attn_weight), "attn_weight must be a CUDA tensor");
AT_ASSERTM(IS_CUDA_TENSOR(grad_output), "grad_output must be a CUDA tensor");

const int batch = value.size(0);
const int spatial_size = value.size(1);
Expand All @@ -128,11 +141,11 @@ std::vector<at::Tensor> ms_deform_attn_cuda_backward(
auto per_sample_loc_size = num_query * num_heads * num_levels * num_point * 2;
auto per_attn_weight_size = num_query * num_heads * num_levels * num_point;
auto grad_output_n = grad_output.view({batch/im2col_step_, batch_n, num_query, num_heads, channels});

for (int n = 0; n < batch/im2col_step_; ++n)
{
auto grad_output_g = grad_output_n.select(0, n);
AT_DISPATCH_FLOATING_TYPES(value.type(), "ms_deform_attn_backward_cuda", ([&] {
AT_DISPATCH_FLOATING_TYPES(GET_TENSOR_TYPE(value), "ms_deform_attn_backward_cuda", ([&] {
ms_deformable_col2im_cuda(at::cuda::getCurrentCUDAStream(),
grad_output_g.data<scalar_t>(),
value.data<scalar_t>() + n * im2col_step_ * per_value_size,
Expand All @@ -153,4 +166,4 @@ std::vector<at::Tensor> ms_deform_attn_cuda_backward(
};
}

} // namespace groundingdino
} // namespace groundingdino