Fix None gradient to enable maximum-likelihood training. This fixes issue #59.

[yiboyang]

Fix None gradient to enable maximum-likelihood training.

This PR fixes resolves issue #59. Previously, likelihood-based training fails because the gradient of the log probability is None even when calling ODESolver.compute_likelihood with enable_grad=True. This PR fixes the issue by leaving out .detach() statements when enable_grad=True. The default behavior (with enable_grad=False) is unaffected.

[meta-cla]

Hi @yiboyang!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!

[meta-cla]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

[timonpalm]

I can confirm that this fix worked. In the plot, you can see the likelihoods of a grid of points computed by the flow (A), their gradients using the current state of the repo (B) and the gradients with this fix.

It is apparent that (B) is clearly wrong.

Wish this PR to be merged soon. :)

[itaigat]

Thanks @yiboyang! I am happy to merge this; can you please remove the private folders from the notebook and solve the CI issues?

[yiboyang]

The formatting issue should be fixed now with 04e5165.