add varbasecopy func to fix the ParamBase type bug in layers.to API #32789
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for your contribution! |
MingMingShangTian
changed the title
chenwhql
reviewed
May 8, 2021
| static void VarBaseCopy(const imperative::VarBase &src, | ||
| imperative::VarBase &dst, const P &dst_device, | ||
| const bool blocking) { | ||
| if (dst.SharedVar()->IsEmpty()) { |
There was a problem hiding this comment.
add else branch and throw error to avoid bug
paddle/fluid/pybind/imperative.cc
Outdated
| dst.SetDataType(src.DataType()); | ||
| dst.SetType(src.Type()); | ||
| dst.SetOverridedStopGradient(src.OverridedStopGradient()); | ||
| if (!src.SharedVar()->IsEmpty()) { |
There was a problem hiding this comment.
same above, add else branch and throw error to avoid bug, can use PADDLE_THROW
| @@ -1639,6 +1672,10 @@ void BindImperative(py::module *m_ptr) { | |||
| self.nrings_ = nrings; | |||
| }); | |||
|
|
|||
| m.def("varbase_copy", &VarBaseCopy<platform::Place>); | |||
| m.def("varbase_copy", &VarBaseCopy<platform::CPUPlace>); | |||
| m.def("varbase_copy", &VarBaseCopy<platform::CUDAPlace>); | |||
| new_t = t._copy_to(device, blocking) | ||
| if dtype is not None and dtype != t.dtype: | ||
| new_t = new_t.cast(dtype=dtype) | ||
| if isinstance(t, framework.ParamBase): |
There was a problem hiding this comment.
the ParamBase._copy_to is still error, whether override the ParamBase._copy_to method to fix this bug?
zhwesky2010
reviewed
May 8, 2021
| new_t = new_t.cast(dtype=dtype) | ||
| if isinstance(t, framework.ParamBase): | ||
| state = copy.deepcopy(t.__dict__) | ||
| new_param = framework.ParamBase(t.shape, dtype, **state) |
There was a problem hiding this comment.
能否给ParamBase添加一个_copy_to的方法覆盖原来的？
chenwhql
reviewed
May 11, 2021
| auto *dst_tensor = dst.MutableVar()->GetMutable<framework::LoDTensor>(); | ||
| dst_tensor->set_lod(src_tensor.lod()); | ||
| framework::TensorCopy(src_tensor, dst_device, dst_tensor); | ||
| if (blocking) { |
There was a problem hiding this comment.
if (!blocking), do we need IncreaseVarbaseReferenceCountUntilCopyComplete?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
Bug fixes
PR changes
APIs
Describe
Fix the bug that the layers.to API will not keep the parameters type. It will change the type from paddle.fluid.framework.ParamBase to paddle.Tensor and lose the attribution of raw type.