Skip to content
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

[Release/2.6] Fix torch.lerp RuntimeError when weight is CPU scalar while input & end are XPU tensor #1201

Merged
merged 3 commits into from
Dec 24, 2024
Merged

[Release/2.6] Fix torch.lerp RuntimeError when weight is CPU scalar while input & end are XPU tensor #1201

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
e40188e
fix lerp
chunhuanMeng Dec 23, 2024
e335123
Update apply_torch_pr.py
xytintel Dec 23, 2024
d29533e
Update SDPUtils.cpp
xytintel Dec 23, 2024
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
2 changes: 1 addition & 1 deletion .github/scripts/apply_torch_pr.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@
# Fallback to CPU for XPU FP64
"https://github.com/pytorch/pytorch/pull/126516",
# Modify the tolerance level in TIMM benchmark
"https://github.com/pytorch/pytorch/pull/129735",
"https://github.com/pytorch/pytorch/pull/143739",
# [XPU] Update XPU C Shim Header
"https://github.com/pytorch/pytorch/pull/141086",
]
Expand Down
2 changes: 2 additions & 0 deletions src/ATen/native/transformers/SDPUtils.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,8 @@

namespace sdp {

using c10::array_of;

bool check_all_tensors_on_device(sdp_params const& params, bool debug) {
// Check that all tensors are on the GPU device
// This should be handled by the stub dispatch, but whe call
Expand Down
14 changes: 14 additions & 0 deletions src/ATen/native/xpu/sycl/LerpKernels.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,15 +57,29 @@ struct LerpScalarFunctor {
opmath_t weight_val_;
};

void lerp_scalar_kernel(
at::TensorIteratorBase& iter,
const c10::Scalar& weight);

void lerp_tensor_kernel(at::TensorIteratorBase& iter) {
auto dtype = iter.common_dtype();
if (at::isComplexType(dtype)) {
AT_DISPATCH_COMPLEX_TYPES_AND(kComplexHalf, dtype, "lerp_xpu", [&] {
if (iter.is_cpu_scalar(3)) {
auto weight_val = iter.scalar_value<scalar_t>(3);
iter.remove_operand(3);
return lerp_scalar_kernel(iter, weight_val);
}
gpu_kernel(iter, LerpTensorComplexFunctor<scalar_t>());
});
} else {
AT_DISPATCH_FLOATING_TYPES_AND2(
at::ScalarType::Half, at::ScalarType::BFloat16, dtype, "lerp_xpu", [&] {
if (iter.is_cpu_scalar(3)) {
auto weight_val = iter.scalar_value<scalar_t>(3);
iter.remove_operand(3);
return lerp_scalar_kernel(iter, weight_val);
}
gpu_kernel(iter, LerpTensorFunctor<scalar_t>());
});
}
Expand Down
Loading