bind OrderedDict() and Dict constructor #26170
Closed
Conversation
This file contains hidden or 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
[ghstack-poisoned]
pytorchbot
added
oncall: jit
This was referenced Sep 13, 2019
eellison
pushed a commit
to eellison/pytorch
that referenced
this pull request
Sep 13, 2019
ghstack-source-id: 442e055 Pull Request resolved: pytorch#26170
driazati
reviewed
Sep 13, 2019
| case TypeKind::FloatType: | ||
| case TypeKind::StringType: | ||
| case TypeKind::TensorType: | ||
| case TypeKind::VarType: |
Contributor
There was a problem hiding this comment.
How can a VarType that is not one of the specified key types work as a dict key?
Contributor
Author
There was a problem hiding this comment.
Otherwise it fails in schema parsing, since the dict is first instantiated with vartype
| TypePtr value_type = static_cast<const DictType*>(output_type.get())->getValueType(); | ||
| return [=](Stack& stack) { | ||
| auto vals = c10::impl::GenericDict(key_type, value_type); | ||
| vals.reserve(stack.size() / 2); |
Contributor
There was a problem hiding this comment.
How is stack's size related to the size of vals
Contributor
Author
There was a problem hiding this comment.
for (size_t i = 0; i < num_inputs; i += 2) {
auto val = pop(stack);
auto key = pop(stack);
vals.insert_or_assign(std::move(key), std::move(val));
}
Contributor
There was a problem hiding this comment.
Shouldn't it be num_inputs * 2 then?
Contributor
Author
There was a problem hiding this comment.
no because each iteration the dict size increases by one
| }, | ||
| aliasAnalysisSpecialCase()), | ||
| Operator( | ||
| "aten::dict((tKey, tVal)[] inputs) -> Dict(tKey, tVal)", |
Contributor
There was a problem hiding this comment.
This should just enumerate the supported keys instead of using vartype
Contributor
Author
|
Replaced by #26373 b/c ghstack was poisioned and i had to resubmit |
facebook-github-bot
pushed a commit
that referenced
this pull request
Sep 19, 2019
Summary:
Makes c10::Dict Ordered and bins binds the OrderedDict() and dict() constructor into torchscript. For the case of the empty constructor dict() i typed it as [str, Tensor] because:
• we're almost dropping support for python 2, at which point all dicts are ordered
• then it's more conventional to write x : Dict[int, int] = {} which is already supported
• It is possible to construct an arbitrarily typed empty OrderedDict through
OrderedDict(torch.jit.annotate(List[Tuple[key, value]], [])
We could consider dropping the no inputs aten::dict constructor since then the types would be more explicit.
This replaces #26170 and #26372 b/c ghstack was poisioned and i had to resubmit
Pull Request resolved: #26465
Differential Revision: D17481604
Pulled By: eellison
fbshipit-source-id: d2d49795a518c3489881afac45d070e5262c5849
zdevito
pushed a commit
to zdevito/ATen
that referenced
this pull request
Sep 19, 2019
Summary:
Makes c10::Dict Ordered and bins binds the OrderedDict() and dict() constructor into torchscript. For the case of the empty constructor dict() i typed it as [str, Tensor] because:
• we're almost dropping support for python 2, at which point all dicts are ordered
• then it's more conventional to write x : Dict[int, int] = {} which is already supported
• It is possible to construct an arbitrarily typed empty OrderedDict through
OrderedDict(torch.jit.annotate(List[Tuple[key, value]], [])
We could consider dropping the no inputs aten::dict constructor since then the types would be more explicit.
This replaces pytorch/pytorch#26170 and pytorch/pytorch#26372 b/c ghstack was poisioned and i had to resubmit
Pull Request resolved: pytorch/pytorch#26465
Differential Revision: D17481604
Pulled By: eellison
fbshipit-source-id: d2d49795a518c3489881afac45d070e5262c5849
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.
Stack from ghstack:
Binds the OrderedDict() and dict() constructor into torchscript. For the case of the empty constructor
dict()i typed it as [str, Tensor] because:• we're almost dropping support for python 2, at which point all dicts are ordered
• then it's more conventional to write
x : Dict[int, int] = {}which is already supported• It is possible to construct an arbitrarily typed empty OrderedDict through
OrderedDict(torch.jit.annotate(List[Tuple[key, value]], [])We could consider dropping the no inputs
aten::dictconstructor since then the types would be more explicit.