Parse and register schema declarations lazily #9801
Conversation
zdevito
requested review from
apaszke,
colesbury,
ezyang,
gchanan and
soumith
as code owners
torch/csrc/jit/script/compiler.cpp
Outdated
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
torch/csrc/jit/operator.cpp
Outdated
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
@pytorchbot retest this please |
|
@pytorchbot retest this please |
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
@zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
This should prevent slow startup times, and will not report as many errors during static initialization time which are hard to debug
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
facebook-github-bot
left a comment
facebook-github-bot
left a comment
There was a problem hiding this comment.
zdevito has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
@pytorchbot retest this please |
Summary: zdevito explained that the attributed versions of `Operator`s are no longer necessary. This PR does two things: 1. Removes all code associated with attributed operators, 2. Adds a second kind of state to `Operator` where it is constructed with an `Operation` directly instead of an `OperationCreator`. This will be useful to test custom operators which don't require a node (you can just retrieve it directly). Now rebased on top of #9801 zdevito Pull Request resolved: #10080 Differential Revision: D9113668 Pulled By: goldsborough fbshipit-source-id: 1276a191c7cf89da1c38488769f2105ce2664750
Summary: This should prevent slow startup times, and will not report as many errors during static initialization time which are hard to debug ezyang apaszke Pull Request resolved: pytorch#9801 Reviewed By: goldsborough Differential Revision: D8986603 Pulled By: zdevito fbshipit-source-id: 440d43ab5e8cffe0b15118cb5fda36391ed06dbc
Summary: zdevito explained that the attributed versions of `Operator`s are no longer necessary. This PR does two things: 1. Removes all code associated with attributed operators, 2. Adds a second kind of state to `Operator` where it is constructed with an `Operation` directly instead of an `OperationCreator`. This will be useful to test custom operators which don't require a node (you can just retrieve it directly). Now rebased on top of pytorch#9801 zdevito Pull Request resolved: pytorch#10080 Differential Revision: D9113668 Pulled By: goldsborough fbshipit-source-id: 1276a191c7cf89da1c38488769f2105ce2664750
3 participants
This should prevent slow startup times, and will not report as many
errors during static initialization time which are hard to debug
@ezyang @apaszke