Skip to content

Conversation

@wjmaddox
Copy link
Contributor

Summary: refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Differential Revision: D29937300

@facebook-github-bot facebook-github-bot added CLA Signed Do not delete this pull request or issue due to inactivity. fb-exported labels Jul 27, 2021
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

wjmaddox pushed a commit to wjmaddox/botorch that referenced this pull request Jul 28, 2021
Summary:
Pull Request resolved: meta-pytorch#883

refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Differential Revision: D29937300

fbshipit-source-id: 81aa5c9b639e030a947c8b4946f4b81024c8e2b7
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

wjmaddox pushed a commit to wjmaddox/botorch that referenced this pull request Jul 28, 2021
Summary:
Pull Request resolved: meta-pytorch#883

refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Differential Revision: D29937300

fbshipit-source-id: 82ff9be65247fc019bc58d973474793d3d1ed4b3
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

@codecov
Copy link

codecov bot commented Jul 28, 2021

Codecov Report

Merging #883 (bb8a4d3) into master (a3230a4) will not change coverage.
The diff coverage is 100.00%.

Impacted file tree graph

@@            Coverage Diff            @@
##            master      #883   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files          108       108           
  Lines         8113      8102   -11     
=========================================
- Hits          8113      8102   -11     
Impacted Files Coverage Δ
botorch/models/higher_order_gp.py 100.00% <100.00%> (ø)
botorch/posteriors/higher_order.py 100.00% <100.00%> (ø)

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a3230a4...bb8a4d3. Read the comment docs.

@facebook-github-bot
Copy link
Contributor

@wjmaddox has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

wjmaddox added a commit to wjmaddox/botorch that referenced this pull request Jul 30, 2021
Summary:
refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Pull Request resolved: meta-pytorch#883

Differential Revision: D29937300

Pulled By: wjmaddox

fbshipit-source-id: def55ebf83e9afa50d515c9fb5ad2457af90ca6f
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

1 similar comment
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

wjmaddox added a commit to wjmaddox/botorch that referenced this pull request Aug 2, 2021
Summary:
refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Pull Request resolved: meta-pytorch#883

Test Plan: unit tests.

Differential Revision: D29937300

Pulled By: wjmaddox

fbshipit-source-id: 14863186d95d4e92f79f8b1f3700ed9446202a9c
Summary:
refactors existing HOGP posterior to use forwards calls and to use cat_rows on the data joint covariance.

Pull Request resolved: meta-pytorch#883

Test Plan: unit tests.

Reviewed By: Balandat

Differential Revision: D29937300

Pulled By: wjmaddox

fbshipit-source-id: a32b81de9c55d013747b5c4b538c8e7b09e70f98
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D29937300

@facebook-github-bot
Copy link
Contributor

@wjmaddox merged this pull request in 8cbed8d.

@saitcakmak
Copy link
Contributor

@wjmaddox, test_higher_order_gp.test_posterior keeps failing for me (even with different seeds). Mind taking a look?

Error
Traceback (most recent call last):
  File "/Users/saitcakmak/PycharmProjects/botorch/test/models/test_higher_order_gp.py", line 90, in test_posterior
    posterior = self.model.posterior(test_x)
  File "/Users/saitcakmak/PycharmProjects/botorch/botorch/models/higher_order_gp.py", line 455, in posterior
    data_joint_covar = data_train_covar.cat_rows(
  File "/Users/saitcakmak/.conda/envs/botorch/lib/python3.8/site-packages/gpytorch/lazy/lazy_tensor.py", line 819, in cat_rows
    new_inv_root = stable_pinverse(new_root).transpose(-2, -1)
  File "/Users/saitcakmak/.conda/envs/botorch/lib/python3.8/site-packages/gpytorch/utils/pinverse.py", line 15, in stable_pinverse
    return torch.triangular_solve(Q.transpose(-1, -2), R).solution
RuntimeError: triangular_solve: For batch 0: U(11,11) is zero, singular U.

@wjmaddox
Copy link
Contributor Author

wjmaddox commented Aug 3, 2021

Ugh, that should have been resolved in gpytorch #1714 (cornellius-gp/gpytorch#1714); try rebasing off gpytorch master.

@saitcakmak
Copy link
Contributor

Yeah, my bad. Looks like I was missing that locally. Thanks for clarification!

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

Labels

CLA Signed Do not delete this pull request or issue due to inactivity. fb-exported Merged

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants