mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-07 00:21:07 +01:00
d701357d92
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/63612 This makes Tensor inherit from a new class TensorBase, that provides a subset of Tensor that doesn't directly depend on native_functions.yaml. Code that only includes TensorBase.h with thus not need to be rebuilt every time someone changes an operator signature. Making `Tensor` inherit from this class means that `const TensorBase&` parameters will be callable with an ordinary `Tensor`. I've also made `Tensor` constructible and assignable from `TensorBase` to minimize friction in code mixing the two types. To help enforce that `Tensor.h` and `Functions.h` aren't accidentally included, I've added an error into `Operators.h` if `TORCH_ASSERT_NO_OPERATORS` is defined. We can either set this in the build system for certain folders, or just define it at the top of any file. I've also included an example of manually special-casing the commonly used `contiguous` operator. The inline function's slow path defers to `TensorBase::__dispatch_contiguous` which is defined in `Tensor.cpp`. I've made it so `OptionalTensorRef` is constructible from `TensorBase`, so I can materialize a `Tensor` for use in dispatch without actually increasing its refcount. Test Plan: Imported from OSS Reviewed By: gchanan Differential Revision: D30728580 Pulled By: ezyang fbshipit-source-id: 2cbc8eee08043382ee6904ea8e743b1286921c03
45 lines
1.2 KiB
C++
45 lines
1.2 KiB
C++
#include <torch/csrc/autograd/cpp_hook.h>
|
|
#include <torch/csrc/autograd/variable.h>
|
|
#include <torch/csrc/autograd/custom_function.h>
|
|
|
|
namespace {
|
|
using torch::autograd::Variable;
|
|
void check_single_result (const at::TensorBase &value, const at::TensorBase &result, std::string hook_name) {
|
|
if (!value.defined()) {
|
|
throw std::runtime_error("can't replace a empty gradient with a non-empty value");
|
|
}
|
|
torch::autograd::check_variable_result(value, result, hook_name);
|
|
}
|
|
}
|
|
|
|
namespace torch { namespace autograd {
|
|
|
|
// NOLINTNEXTLINE(modernize-pass-by-value)
|
|
CppFunctionPreHook::CppFunctionPreHook(const std::shared_ptr<hooks_list> &hooks, int value_idx)
|
|
: hooks_(hooks)
|
|
, value_idx_(value_idx)
|
|
{}
|
|
|
|
variable_list CppFunctionPreHook::operator()(const variable_list& values) {
|
|
auto value = values[value_idx_];
|
|
for (unsigned i = 0; i < hooks_->size(); ++i) {
|
|
auto &hook = (*hooks_)[i];
|
|
if (!hook) {
|
|
// hook was removed
|
|
continue;
|
|
}
|
|
auto res = hook(value);
|
|
if (!res.defined()) {
|
|
// Don't change gradient
|
|
continue;
|
|
}
|
|
check_single_result(value, res, c10::to_string(i));
|
|
value = std::move(res);
|
|
}
|
|
variable_list results(values);
|
|
results[value_idx_] = value;
|
|
return results;
|
|
}
|
|
|
|
}} // namespace torch::autograd
|