Skip to content

Conversation

@justinchuby
Copy link
Collaborator

Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Meta Open Source bot. label Jul 8, 2025
Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchuby@users.noreply.github.com>
@titaiwangms
Copy link
Collaborator

I will have another commit to move it to done.

Copy link
Contributor

@jbschlosser jbschlosser left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Awesome, thanks!


#### `onnxscript>=0.3.1` is required for the `dynamo=True` option ([#157017](https://github.com/pytorch/pytorch/pull/157017))

You must upgrade `onnxscript` to version 0.3.1 or higher for it to be compatible with PyTorch 2.8.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we specify the reason?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it's fine? We have a new api in this version that pytorch uses

@titaiwangms
Copy link
Collaborator

cc @justinchuby @xadupre for final review.

@justinchuby
Copy link
Collaborator Author

@jbschlosser could you grant write access to me and @titaiwangms ?

@justinchuby justinchuby requested a review from jbschlosser July 8, 2025 22:23
@jbschlosser
Copy link
Contributor

@jbschlosser could you grant write access to me and @titaiwangms ?

sure thing, one sec

@jbschlosser
Copy link
Contributor

@titaiwangms @justinchuby I just added you to the repo; you should have received an invite email

@justinchuby justinchuby merged commit 107439d into meta-pytorch:main Jul 9, 2025
1 check passed
@justinchuby justinchuby deleted the justinchu/onnx branch July 9, 2025 15:36
@justinchuby
Copy link
Collaborator Author

Looks like the file was directly merged to done. @jbschlosser let us know if there are any changes you would like us to make


The default will be `dynamo=True` starting from PyTorch 2.9. You are encouraged to migrate to use the `dynamo=True` option in `torch.onnx.export`. This flag makes `torch.export.export` the default export path, replacing `TorchScript`, as `TorchScript` is nearing end-of-life"

To maintain the old behavior, set `dynamo=True`. You are encouraged to also experiment with the `fallback=True` option that will make the exporter fall back to the `dynamo=False` path if there are errors.
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

dynamo=False?


#### Support for symbolic operators in the `dynamo=True` export path ([#148905](https://github.com/pytorch/pytorch/pull/148905), [#149678](https://github.com/pytorch/pytorch/pull/149678), [#150038](https://github.com/pytorch/pytorch/pull/150038))

Two operators `torch.onnx.ops.symbolic` and `torch.onnx.ops.symbolic_multi_out` are defined to allow you to create symbolic ONNX operators directly in your PyTorch models. You can use them in a forward function:
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You can use them in a function:

jhavukainen pushed a commit to jhavukainen/torch-release-notes that referenced this pull request Jul 9, 2025
* dataloader release notes updated

* address comments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Meta Open Source bot.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants