Skip to content

[Eager, Performance Optimization] Optimize clone interface #46190

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

Merged
merged 2 commits into from
Sep 21, 2022
Merged

[Eager, Performance Optimization] Optimize clone interface #46190

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ad4b628
[Eager] polish clone interface
veyron95 Sep 19, 2022
295d91c
rm clone in python, add clone in eager_method.cc
veyron95 Sep 19, 2022
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
22 changes: 22 additions & 0 deletions paddle/fluid/pybind/eager_method.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -443,6 +443,24 @@ static PyObject* tensor_method_copy_(TensorObject* self,
EAGER_CATCH_AND_THROW_RETURN_NULL
}

static PyObject* tensor_method_clone(TensorObject* self,
PyObject* args,
PyObject* kwargs) {
EAGER_TRY

PADDLE_ENFORCE_EQ(
self->tensor.initialized(),
true,
paddle::platform::errors::InvalidArgument(
"We can only support initialized tensor in clone, however we got "
"uninitialized tensor %s, please check your code.",
self->tensor.name()));

auto out = assign_ad_func(self->tensor);
return ToPyObject(out);
EAGER_CATCH_AND_THROW_RETURN_NULL
}

static PyObject* tensor_retain_grads(TensorObject* self,
PyObject* args,
PyObject* kwargs) {
Expand Down Expand Up @@ -1841,6 +1859,10 @@ PyMethodDef variable_methods[] = {
(PyCFunction)(void (*)(void))tensor_method_copy_,
METH_VARARGS | METH_KEYWORDS,
NULL},
{"clone",
(PyCFunction)(void (*)(void))tensor_method_clone,
METH_VARARGS | METH_KEYWORDS,
NULL},
{"reconstruct_from_",
(PyCFunction)(void (*)(void))tensor_method_reconstruct_from_,
METH_VARARGS | METH_KEYWORDS,
Expand Down
12 changes: 0 additions & 12 deletions python/paddle/fluid/dygraph/varbase_patch_methods.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -815,17 +815,6 @@ def _set_grad_ivar(self, value):
raise TypeError(
"_set_grad_ivar is only supported for Parameter Tensor")

@framework.dygraph_only
def clone(self):
if in_dygraph_mode():
return _C_ops.assign(self)

if _in_legacy_dygraph():
output = core.VarBase()
else:
output = core.eager.Tensor()
return _legacy_C_ops.assign(self, output)

@framework.dygraph_only
def value(self):
return self
Expand Down Expand Up @@ -1009,7 +998,6 @@ def to_sparse_coo(self, sparse_dim):
if framework._in_eager_mode_:
setattr(core.eager.Tensor, "_grad_ivar", _grad_ivar)
setattr(core.eager.Tensor, "_set_grad_ivar", _set_grad_ivar)
setattr(core.eager.Tensor, "clone", clone)
setattr(core.eager.Tensor, "value", value)
setattr(core.eager.Tensor, "cpu", cpu)
setattr(core.eager.Tensor, "cuda", cuda)
Expand Down