functorch.jvp support for autograd.Function #90077
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
WIP, to be filled in [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/90077
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 FailuresAs of commit 018bcc1: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Dec 2, 2022
Contributor
Author
|
Not ready for review yet, getting some CI signal. |
zou3519
added
the
release notes: torch.func
This was referenced Dec 5, 2022
WIP, to be filled in [ghstack-poisoned]
WIP, to be filled in [ghstack-poisoned]
WIP, to be filled in [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
zou3519
added a commit
that referenced
this pull request
Dec 6, 2022
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db ghstack-source-id: 026f85e Pull Request resolved: #90077
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
albanD
approved these changes
Dec 13, 2022
torch/_functorch/pyfunctorch.py
Outdated
| def lower(self): | ||
| prev_fwd_grad_mode = self.prev_fwd_grad_mode() | ||
| if not self.prev_fwd_grad_mode: | ||
| return contextlib.nested(_set_fwd_grad_enabled(False), super().lower()) |
Collaborator
There was a problem hiding this comment.
isn't contextlib.nested deprecated since python 2.7?
Contributor
Author
There was a problem hiding this comment.
You're right, it's not in the Python 3 docs. I'll find an alternative
This PR adds functorch.jvp support for autograd.Function. It does so by adding a jvp rule for custom_function_call. For a regular PyTorch operation (like at::sin), the VariableType kernel: - re-dispatches to at::sin - calls the jvp rule for at::sin The jvp rule for custom_function_call does just that. It constructs a new autograd.Function (because the above logic already exists). Inside the forward, it re-dispatches to custom_function_call. In the jvp rule, it just calls whatever the jvp rule is supposed to be. Since this logic is really close to the custom_function_call_grad, I just put them together. Test Plan: - added jvp rules to the autograd.Function in autograd_function_db [ghstack-poisoned]
Contributor
|
This pull request has been merged in 4809e83. |
This was referenced Dec 15, 2022
Closed
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:
This PR adds functorch.jvp support for autograd.Function. It does so by
adding a jvp rule for custom_function_call.
For a regular PyTorch operation (like at::sin), the VariableType kernel:
The jvp rule for custom_function_call does just that. It constructs a
new autograd.Function (because the above logic already exists). Inside
the forward, it re-dispatches to custom_function_call. In the jvp rule,
it just calls whatever the jvp rule is supposed to be.
Since this logic is really close to the custom_function_call_grad, I
just put them together.
Test Plan: