Removed no_grad from solver #19
Closed
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
mhavasi
force-pushed
the
marton/no_no_grad
branch
from
b61b6d0 to
68f0f96
Compare
bkarrer
reviewed
Dec 17, 2024
| step_size=step_size if method != "dopri5" else None, | ||
| time_grid=time_grid, | ||
| method=method, | ||
| enable_grad=True, |
There was a problem hiding this comment.
Check grads are not computed without this?
| @@ -105,6 +127,7 @@ def dummy_log_p(x: Tensor) -> Tensor: | |||
| log_p0=dummy_log_p, | |||
| step_size=step_size, | |||
| exact_divergence=True, | |||
| enable_grad=True, | |||
There was a problem hiding this comment.
Check grads not computed without this?
| @@ -174,16 +175,15 @@ def dynamics_func(t, states): | |||
| y_init = (x_1, torch.zeros(x_1.shape[0], device=x_1.device)) | |||
| ode_opts = {"step_size": step_size} if step_size is not None else {} | |||
|
|
|||
| with torch.no_grad(): | |||
There was a problem hiding this comment.
Was this no_grad unnecessary previously?
There was a problem hiding this comment.
This was unnecessary yes.
|
Do the docs for the affected methods get updated with the |
|
Moving this PR to internal |
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Outdated suggestions cannot be applied.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.
It was suggested that no part of the core library, such as the sampler and likelihood computation, should be wrapped in
no_grad. This is because if the code is expected to be integrated into other people's projects, it should not enforceno_gradand instead let the user decide whether to track the computation graph. While users can add their ownno_grad, it is impossible for them to remove ano_gradthat has already been applied.This PR removes
no_gradfrom the library.I tested it with the example notebooks and I also added a unit test to make sure we can differentiate through the ode solver and the likelihood computation.