[Feature] Torch Module Parameters Gradient Estimator #31
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
tests/test_gradients.py
Outdated
Comment on lines
386
to
389
| async def test_torch_param_backward_estimator(): | ||
| torch_module = torch.nn.Linear(4, 1) | ||
| torch_op = TorchOp(torch_module) | ||
| estimator = TorchParamBackwardEstimator(torch_module) |
There was a problem hiding this comment.
This looks pretty good! Following our conversation this morning, do you mind testing for output dim >1? It's worth checking that the shapes are correct in that case too, since we apply a few squeezes in the code.
|
I extended the tests. I noticed that at the end of the graph we set output_grads to [0] And then we try to expand the dimensions in TorchOp But |
sidnarayanan
requested changes
Sep 17, 2024
jamesbraza
reviewed
Sep 17, 2024
Co-authored-by: James Braza <jamesbraza@gmail.com>
jamesbraza
reviewed
Sep 17, 2024
Co-authored-by: James Braza <jamesbraza@gmail.com>
jamesbraza
reviewed
Sep 18, 2024
sidnarayanan
approved these changes
Sep 18, 2024
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Outdated suggestions cannot be applied.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.
Adds a
TorchParamBackwardEstimatorclass, which includes abackwardmethod. This method can be used to monkey-patch theTorchOpclass during gradient computation. The patch allows the gradients to be computed with respect to the internal parameters of theTorchOp'sunderlying torch.nn.Module.