pytorch/torch/csrc/autograd/python_variable_indexing.cpp
Edward Yang 4e21157e01 Revert "Revert D18171156: Merge Tensor and Variable." (#29299)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/29299

This reverts commit 9c43b16df9, but also
with the changes from D18348622.  Comments there:

thpp-compatibility is used by admarket/adreview/service:adreviewservice and
libtorch is too big for the service to deal with.

thpp-compatibility doesn't support autograd, so we hack around dispatching
variables by using AutoNonVariableTypeMode everywhere we call into ATen,
so we never attempt to call into Variable stubs.  If you get it wrong,
you'll get an error like:

```
what():  Could not run 'aten::empty' with arguments from the 'VariableTensorId' backend. 'aten::empty' is only available for these backends: [SparseCPUTensorId, CPUTensorId, MkldnnCPUTensorId]. (lookup_ at caffe2/aten/src/ATen/core/dispatch/DispatchTable.h:298)
```

Test Plan:
Imported from OSS

```
buck test //thpp-compatibility/...
buck build mode/opt-clang admarket/adreview/service:adreviewservice
```

adreviewservice canary: https://our.intern.facebook.com/intern/ads/canary/422290029716387895 (comparing against parent comment due to current breakage) ==> experiment store https://our.intern.facebook.com/intern/experiment_store/experiment/43990006/
adfinder canary: https://our.intern.facebook.com/intern/ads/canary/422268535840333934
adindexer canary: https://our.intern.facebook.com/intern/ads/canary/422268550559034675

adreview second canary:  https://our.intern.facebook.com/intern/ads/canary/422307863515591925

canary without thpp-compat fixups https://our.intern.facebook.com/intern/ads/canary/422308951649168772

Reviewed By: dreiss

Differential Revision: D18353504

Pulled By: ezyang

fbshipit-source-id: 65feaba39fa07bb66762810909aeb38868668a30
2019-11-08 09:11:20 -08:00

422 lines
14 KiB
C++

#include <torch/csrc/autograd/python_variable_indexing.h>
#include <torch/csrc/DynamicTypes.h>
#include <torch/csrc/Exceptions.h>
#include <torch/csrc/THP_export.h>
#include <torch/csrc/autograd/function.h>
#include <torch/csrc/autograd/python_variable.h>
#include <torch/csrc/autograd/utils/wrap_outputs.h>
#include <torch/csrc/autograd/variable.h>
#include <torch/csrc/utils/python_compat.h>
#include <torch/csrc/utils/python_numbers.h>
#include <torch/csrc/utils/tensor_new.h>
#include <torch/csrc/jit/tracer.h>
#include <torch/csrc/jit/ir.h>
#include <torch/csrc/utils/tensor_types.h>
#include <ATen/DeviceGuard.h>
#include <ATen/ExpandUtils.h>
#include <c10/core/TensorOptions.h>
#include <ATen/core/LegacyTypeDispatch.h>
#include <vector>
#include <tuple>
using namespace at;
using namespace torch::autograd::utils;
namespace torch { namespace autograd {
Py_ssize_t THPVariable_length(PyObject* self) {
HANDLE_TH_ERRORS
auto& self_ = reinterpret_cast<THPVariable*>(self)->cdata;
if (self_.dim() == 0) {
return 0;
}
return (Py_ssize_t)self_.size(0);
END_HANDLE_TH_ERRORS_RET(-1)
}
// We allow indexing by integers, slices, ellipsis, None, Variables,
// and tuples of those types. We also handle bools as if they were a
// Variable[ByteTensor].
static int64_t count_specified_dimensions(PyObject* index) {
// Count the number of indexed dimensions (everything but ellipsis and None)
int64_t count = 0;
auto size = PyTuple_GET_SIZE(index); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
for (Py_ssize_t i = 0; i < size; i++) {
PyObject* obj = PyTuple_GET_ITEM(index, i); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
if (THPVariable_Check(obj)) {
auto& var = reinterpret_cast<THPVariable*>(obj)->cdata;
if (var.scalar_type() == kByte || var.scalar_type() == kBool) {
count += var.dim();
} else {
count++;
}
} else if (obj != Py_None && obj != Py_Ellipsis && obj != Py_True && obj != Py_False) { // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
count++;
}
}
return count;
}
[[noreturn]]
static void invalid_index(PyObject* obj) {
throw IndexError(
"only integers, slices (`:`), ellipsis (`...`), None and long or byte "
"Variables are valid indices (got %s)", Py_TYPE(obj)->tp_name);
}
static Variable applySlice(const Variable& self, int64_t dim, PyObject* slice, bool ensure_view=false) {
Py_ssize_t start, stop, step;
auto length = self.size(dim);
if (!THPUtils_unpackSlice(slice, &start, &stop, &step)) {
throw python_error();
}
if (step == 0) {
throw ValueError("step cannot be zero");
}
if (step < 0) {
// TODO: implement negative step
throw ValueError("negative step not yet supported");
}
PySliceObject* sliceobj = (PySliceObject*) slice;
if (jit::tracer::isTracing() && THPVariable_Check(sliceobj->start)) {
auto& var = THPVariable_Unpack(sliceobj->start);
jit::tracer::ArgumentStash::stashValue(std::string("start"), 1, var, jit::IntType::get());
}
if (jit::tracer::isTracing() && THPVariable_Check(sliceobj->stop)) {
auto& var = THPVariable_Unpack(sliceobj->stop);
jit::tracer::ArgumentStash::stashValue(std::string("end"), 1, var, jit::IntType::get());
}
if (jit::tracer::isTracing() && THPVariable_Check(sliceobj->step)) {
auto& var = THPVariable_Unpack(sliceobj->step);
jit::tracer::ArgumentStash::stashValue(std::string("step"), 1, var, jit::IntType::get());
}
// Skip this optimization if we are tracing, as the trace may be polymorphic
// over the shape of the `self` tensor, and we still want to record
// the slice.
if (!ensure_view && start == 0 && stop == length && step == 1 && !jit::tracer::isTracing()) {
return self;
}
return self.slice(dim, start, stop, step);
}
static Variable applySelect(const Variable& self, int64_t dim, PyObject* index, int64_t real_dim=0) {
if (jit::tracer::isTracing() && THPVariable_Check(index)) {
auto& var = THPVariable_Unpack(index);
jit::tracer::ArgumentStash::stashValue(std::string("index"), 1, var, jit::IntType::get());
}
int64_t unpacked_index = THPUtils_unpackLong(index);
if (unpacked_index == 0 && dim == 0 && self.dim() == 0) {
throw IndexError(
"invalid index of a 0-dim tensor. "
"Use tensor.item() to convert a 0-dim tensor to a Python number");
}
int64_t size = self.size(dim);
if (unpacked_index < -size || unpacked_index >= size) {
throw IndexError("index %lld is out of bounds for dimension %lld with size %lld",
unpacked_index, real_dim, size);
}
// if the index is negative, do not normalize it because that would fix the index
// on the current tensor size in the tracer.
// aten::select also works on negative indices
return self.select(dim, unpacked_index);
}
static Variable sequenceToVariable(c10::TensorTypeId type_id, PyObject* seq) {
return torch::utils::indexing_tensor_from_data(type_id, kLong, c10::nullopt, seq);
}
static Variable valueToTensor(c10::TensorOptions options, PyObject* value) {
if (THPVariable_Check(value)) {
return reinterpret_cast<THPVariable*>(value)->cdata;
}
at::AutoNonVariableTypeMode guard;
if (THPUtils_checkLong(value) || PyBool_Check(value)) {
return at::scalar_tensor(Scalar(THPUtils_unpackLong(value)), options);
}
if (PyFloat_Check(value)) {
return at::scalar_tensor(Scalar(THPUtils_unpackDouble(value)), options);
}
throw TypeError(
"can't assign a %s to a %s",
Py_TYPE(value)->tp_name,
torch::utils::type_to_string(getNonVariableDeprecatedTypeProperties(options.backend(), typeMetaToScalarType(options.dtype()))).c_str());
}
static Variable boolToIndexingTensor(const Variable& self, bool value) {
// booleans add a dimension of size 1. true indexes this dimension as if 0:, false as empty.
if (value) {
return at::zeros({1}, self.options().dtype(kLong));
} else {
return at::empty({0}, self.options().dtype(kLong));
}
}
static Variable applySlicing(const Variable& self, PyObject* index, variable_list& outIndices) {
int64_t size = PyTuple_GET_SIZE(index); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
int64_t dim = 0;
int64_t specified_dims = count_specified_dimensions(index);
auto handle_var = [&](const Variable& var) {
// TODO: check scalarType
outIndices.resize(dim + 1);
outIndices[dim] = var;
dim++;
};
if (specified_dims > self.dim()) {
throw IndexError("too many indices for tensor of dimension %d", (int)self.dim());
}
Variable result = self;
for (int64_t i = 0; i < size; i++) {
PyObject* obj = PyTuple_GET_ITEM(index, i); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
if (THPUtils_checkLong(obj)) {
result = applySelect(result, dim, obj, i);
} else if (PySlice_Check(obj)) {
result = applySlice(result, dim, obj);
dim++;
} else if (obj == Py_Ellipsis) {
dim += self.dim() - specified_dims;
} else if (obj == Py_None) {
result = result.unsqueeze(dim);
dim++;
} else if (PyBool_Check(obj)) {
result = result.unsqueeze(dim);
handle_var(boolToIndexingTensor(result, obj == Py_True)); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
} else if (THPVariable_Check(obj)) {
auto& var = THPVariable_Unpack(obj);
auto scalar_type = var.scalar_type();
if (var.dim() == 0 && at::isIntegralType(scalar_type, /*includeBool=*/true)) {
if (scalar_type != at::kByte && scalar_type != at::kBool) {
result = applySelect(result, dim, obj, i);
} else {
result = result.unsqueeze(dim);
if(scalar_type == at::kBool) {
handle_var(boolToIndexingTensor(result, var.item<bool>() != 0));
} else {
handle_var(boolToIndexingTensor(result, var.item<uint8_t>() != 0));
}
}
} else {
handle_var(var);
}
} else if (PySequence_Check(obj)) {
// TODO: Naughty naughty get out of jail free
// (Fixing this means I have to fix the call chain though :/)
handle_var(sequenceToVariable(legacyExtractTypeId(self), obj));
} else {
auto index = THPObjectPtr(PyNumber_Index(obj));
if (!index) {
PyErr_Clear();
invalid_index(obj);
}
result = applySelect(result, dim, index, i);
}
}
return result;
}
static std::vector<Tensor> typeConvertIndices(const Variable& self, const variable_list& indices) {
std::vector<Tensor> converted_inds(indices.size());
for (size_t i = 0; i < indices.size(); ++i) {
const auto &ind = indices[i];
if (ind.defined()) {
converted_inds[i] = ind.to(ind.options().device(self.device()));
} else {
converted_inds[i] = indices[i];
}
}
return converted_inds;
}
static Variable dispatch_index(const Variable& self, const variable_list& indices) {
AutoNoGIL no_gil;
std::vector<Tensor> converted_indices = typeConvertIndices(self, indices);
OptionalDeviceGuard device_guard(device_of(self));
return self.index(converted_indices);
}
static Variable dispatch_index_put_(Variable& self, const variable_list& indices, const Variable& value) {
AutoNoGIL no_gil;
std::vector<Tensor> converted_indices = typeConvertIndices(self, indices);
OptionalDeviceGuard device_guard(device_of(self));
return self.index_put_(converted_indices, value);
}
static bool treatSequenceAsTuple(PyObject* index) {
if (PyTuple_Check(index)) {
return true;
}
if (!PySequence_Check(index)) {
return false;
}
// This uses a heuristics from NumPy for determining whether to treat
// non-tuple sequences as if they were a tuple. From the NumPy code comments:
//
// "At this point, we're left with a non-tuple, non-array, sequence:
// typically, a list. We use some somewhat-arbitrary heuristics from here
// onwards to decided whether to treat that list as a single index, or a
// list of indices. Backwards compatibility only takes effect for short
// sequences - otherwise we treat it like any other scalar."
auto n = PySequence_Size(index);
if (n < 0) {
// Negative size indicates a Python error in the PySequence_Size call.
PyErr_Clear();
return false;
}
if (n >= 32) {
return false;
}
for (Py_ssize_t i = 0; i < n; i++) {
auto obj = THPObjectPtr{PySequence_GetItem(index, i)};
if (!obj.get()) {
PyErr_Clear();
return false;
}
if (THPVariable_Check(obj.get()) || PySequence_Check(obj.get()) || PySlice_Check(obj.get())) {
return true;
}
if (obj.get() == Py_Ellipsis || obj.get() == Py_None) {
return true;
}
}
return false;
}
static THPObjectPtr wrapTuple(PyObject* index) {
THPObjectPtr res;
if (treatSequenceAsTuple(index)) {
res = PySequence_Tuple(index);
} else {
res = PyTuple_Pack(1, index); // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
}
if (!res) throw python_error();
return res;
}
PyObject* THPVariable_getitem(PyObject* self, PyObject* index) {
HANDLE_TH_ERRORS
auto& self_ = reinterpret_cast<THPVariable*>(self)->cdata;
OptionalDeviceGuard device_guard(device_of(self_));
// handle simple types: integers, slices, ellipsis
if (index == Py_None) {
return wrap(self_.unsqueeze(0));
} else if (index == Py_Ellipsis) {
return wrap(at::alias(self_));
} else if (THPUtils_checkLong(index)) {
return wrap(applySelect(self_, 0, index));
} else if (PySlice_Check(index)) {
return wrap(applySlice(self_, 0, index, true));
}
// wrap index in a tuple if it's not already one
THPObjectPtr holder = wrapTuple(index);
variable_list variableIndices;
Variable sliced = applySlicing(self_, holder.get(), variableIndices);
if (variableIndices.empty()) {
if (sliced.is_same(self_)) {
// ensure we return a shallow copy for things like x[...]
sliced = at::alias(sliced);
}
return wrap(sliced);
}
// indexing by tensors ("advanced" indexing)
return wrap(dispatch_index(sliced, variableIndices));
Py_RETURN_NONE;
END_HANDLE_TH_ERRORS
}
// To match numpy semantics:
// As a special case for backwards compatibility,
// strip away unit dimensions from the left of 'src'
static IntArrayRef slicePrefix1sSize(IntArrayRef sizes) {
size_t first_non1_src = sizes.size();
for (size_t i = 0; i < sizes.size(); ++i) {
if (sizes[i] != 1) {
first_non1_src = i;
break;
}
}
return sizes.slice(first_non1_src);
}
static void copy_to(Variable dst, const Variable& src) {
Tensor b_src;
IntArrayRef sliced_src_sizes = slicePrefix1sSize(src.sizes());
std::tie(b_src) = expand_inplace(dst, src.view(sliced_src_sizes), "setitem");
dst.copy_(b_src);
}
int THPVariable_setitem(PyObject* self, PyObject* index, PyObject* py_value) {
HANDLE_TH_ERRORS
if (py_value == nullptr) {
throw TypeError("Tensor does not support deleting items");
}
auto& self_ = reinterpret_cast<THPVariable*>(self)->cdata;
OptionalDeviceGuard device_guard(device_of(self_));
Variable value;
// TODO: This qint special case looks very suspicious...
if (isQIntType(self_.scalar_type())) {
value = valueToTensor(device(kCPU).dtype(kFloat), py_value);
} else {
value = valueToTensor(self_.options(), py_value);
}
// handle simple types: integers, slices, ellipsis, bool
if (index == Py_False) { // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
// do nothing for false (technically we should check the size, but we don't have
// real 0-sized shapes.
return 0;
} else if (index == Py_Ellipsis) {
copy_to(self_, value);
return 0;
} else if (index == Py_None || index == Py_True) { // NOLINT(cppcoreguidelines-pro-type-cstyle-cast)
copy_to(self_.unsqueeze(0), value);
return 0;
} else if (THPUtils_checkLong(index)) {
copy_to(applySelect(self_, 0, index), value);
return 0;
} else if (PySlice_Check(index)) {
copy_to(applySlice(self_, 0, index), value);
return 0;
}
// wrap index in a tuple if it's not already one
THPObjectPtr holder = wrapTuple(index);
variable_list variableIndices;
Variable sliced = applySlicing(self_, holder.get(), variableIndices);
if (variableIndices.empty()) {
copy_to(sliced, value);
return 0;
}
IntArrayRef slicedValueSizes = slicePrefix1sSize(value.sizes());
torch::autograd::Variable valuesSliced;
if (!value.sizes().equals(slicedValueSizes)) {
valuesSliced = value.view(slicedValueSizes);
} else {
valuesSliced = value;
}
dispatch_index_put_(sliced, variableIndices, valuesSliced);
return 0;
END_HANDLE_TH_ERRORS_RET(-1)
}
}} // namespace torch::autograd