mirror of
https://github.com/pytorch/pytorch.git
synced 2025-10-21 13:44:15 +08:00
ebc2365fd3
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/26339 Serializes per-channel tensor in both torch.serialization and jit. Since we didn't bind Quantizer properly yet, I chose to save a tuple representing quantizer settings. To avoid recursive tensor serialization calls, I'm using tuple instead of tensor to store scales and zero points. driazati - please check the serialization logic. Is there a good test that compares that JIT serialization and python serialization are equivalent? (I haven't tested it yet) Test Plan: Imported from OSS Differential Revision: D17443222 Pulled By: dzhulgakov fbshipit-source-id: a34758de1ffd2ec1cdc5355f5baf95284a4ccf4b
45 lines
1.2 KiB
C++
45 lines
1.2 KiB
C++
#include <torch/csrc/utils/tensor_qschemes.h>
|
|
|
|
#include <torch/csrc/DynamicTypes.h>
|
|
#include <torch/csrc/Exceptions.h>
|
|
#include <torch/csrc/QScheme.h>
|
|
#include <c10/core/QScheme.h>
|
|
|
|
#include <torch/csrc/python_headers.h>
|
|
#include <torch/csrc/utils/object_ptr.h>
|
|
|
|
|
|
namespace torch {
|
|
namespace utils {
|
|
|
|
static PyObject* thp_qscheme_array[at::COMPILE_TIME_NUM_QSCHEMES];
|
|
|
|
void initializeQSchemes() {
|
|
auto torch_module = THPObjectPtr(PyImport_ImportModule("torch"));
|
|
if (!torch_module) {
|
|
throw python_error();
|
|
}
|
|
|
|
for (int i = 0; i < at::COMPILE_TIME_NUM_QSCHEMES; ++i) {
|
|
auto qscheme = static_cast<at::QScheme>(i);
|
|
PyObject* qscheme_obj = THPQScheme_New(qscheme, toString(qscheme));
|
|
thp_qscheme_array[static_cast<int>(qscheme)] = qscheme_obj;
|
|
Py_INCREF(qscheme_obj);
|
|
if (PyModule_AddObject(
|
|
torch_module, toString(qscheme).c_str(), qscheme_obj) != 0) {
|
|
throw python_error();
|
|
}
|
|
}
|
|
}
|
|
|
|
PyObject* getTHPQScheme(at::QScheme qscheme) {
|
|
auto qscheme_ = thp_qscheme_array[static_cast<int>(qscheme)];
|
|
if (!qscheme_) {
|
|
throw std::invalid_argument("unsupported QScheme");
|
|
}
|
|
return qscheme_;
|
|
}
|
|
|
|
} // namespace utils
|
|
} // namespace torch
|