Skip to content

[Runtime][Builtin] Handle mismatched type on argument #0 when calling Builtin Runtime Operators #18837

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

Open
locnd182644 wants to merge 6 commits into
base: main
Choose a base branch
from
+32 −4
Open

[Runtime][Builtin] Handle mismatched type on argument #0 when calling Builtin Runtime Operators #18837

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
6d6d330
[Runtime][Builtin] Handle mismatched type on argument #0 when calling…
locnd182644 Feb 27, 2026
5a29e6a
Merge branch 'apache:main' into Runtime/Builtin
locnd182644 Feb 27, 2026
e1cb235
Using std::make_unique and nullptr following Review
locnd182644 Feb 27, 2026
d446cf3
Updated
locnd182644 Feb 27, 2026
857f182
rerun
locnd182644 Feb 28, 2026
7c3e1f3
rerun again
locnd182644 Feb 28, 2026
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
36 changes: 32 additions & 4 deletions src/runtime/vm/builtin.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -526,11 +526,39 @@ TVM_FFI_STATIC_INIT_BLOCK() {
TVM_FFI_STATIC_INIT_BLOCK() {
namespace refl = tvm::ffi::reflection;
refl::GlobalDef()
.def_method("vm.builtin.shape_of", [](Tensor data) -> ffi::Shape { return data.Shape(); })
.def_method("vm.builtin.shape_of",
[](ffi::Any any) -> ffi::Shape {
if (auto opt_tensor = any.try_cast<Tensor>()) {
return opt_tensor.value().Shape();
} else if (auto opt_dltensor = any.try_cast<DLTensor*>()) {
DLTensor* ptr = opt_dltensor.value();
return ffi::Shape(ptr->shape, ptr->shape + ptr->ndim);
} else {
TVM_FFI_THROW(TypeError)
<< "vm.builtin.shape_of expects a Tensor or DLTensor*, but get "
<< any.GetTypeKey();
}
})
.def("vm.builtin.copy", [](ffi::Any a) -> ffi::Any { return a; })
.def(
"vm.builtin.reshape",
[](Tensor data, ffi::Shape new_shape) { return data.CreateView(new_shape, data->dtype); })
.def("vm.builtin.reshape",
[](ffi::Any any, ffi::Shape new_shape) {
if (auto opt_tensor = any.try_cast<Tensor>()) {
Tensor data = opt_tensor.value();
return data.CreateView(new_shape, data->dtype);
} else if (auto opt_dltensor = any.try_cast<DLTensor*>()) {
DLTensor* ptr = opt_dltensor.value();
auto tmp = std::make_unique<DLManagedTensor>();
tmp->dl_tensor = *ptr;
tmp->manager_ctx = nullptr;
tmp->deleter = nullptr;
Tensor data = Tensor::FromDLPack(tmp.release());
return data.CreateView(new_shape, data->dtype);
} else {
TVM_FFI_THROW(TypeError)
<< "vm.builtin.reshape expects a Tensor or DLTensor*, but get "
<< any.GetTypeKey();
}
})
.def("vm.builtin.null_value", []() -> std::nullptr_t { return nullptr; })
.def_packed("vm.builtin.to_device", [](ffi::PackedArgs args, ffi::Any* rv) {
Tensor data = args[0].cast<Tensor>();
Expand Down