Add KeOps MMD detector #548
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
Additional note: Need to check suitable error is raised when passed to |
| batch_size_permutations | ||
| KeOps computes the n_permutations of the MMD^2 statistics in chunks of batch_size_permutations. | ||
| Only relevant for 'keops' backend. |
There was a problem hiding this comment.
Maybe we should open an issue for this?
| elif backend == 'pytorch' and has_pytorch: | ||
| pop_kwargs += ['batch_size_permutations'] | ||
| detector = MMDDriftTorch | ||
| else: |
There was a problem hiding this comment.
I mistakingly opened a new issue #576 (comment) for this as thought these "if statements" were already present. Since they are actually added here (for the new pop_kwargs bit), it would make more sense to fix here.
The has_tensorflow and has_pytorch are unnecessary as BackendValidator should have already raised an error if backend='tensorflow' and has_tensorflow=False, or the PyTorch equivalent.
| @@ -89,7 +89,7 @@ def __init__( | |||
| # initialize kernel | |||
| sigma = torch.from_numpy(sigma).to(self.device) if isinstance(sigma, # type: ignore[assignment] | |||
| np.ndarray) else None | |||
| self.kernel = kernel(sigma) if kernel == GaussianRBF else kernel | |||
| self.kernel = kernel(sigma).to(self.device) if kernel == GaussianRBF else kernel | |||
There was a problem hiding this comment.
Not sure we actually fixed this?? I have opened an issue (#586) and will fix it tomorrow.
| return self.log_sigma.exp() | ||
|
|
||
| def forward(self, x: LazyTensor, y: LazyTensor, infer_sigma: bool = False) -> LazyTensor: | ||
|
|
There was a problem hiding this comment.
I agree with opening an issue since this applies to all kernels (even if the issue is just to review docstring conventions with the conclusion being keep-as-is!).
Codecov Report
@@ Coverage Diff @@
## master #548 +/- ##
=========================================
Coverage ? 83.51%
=========================================
Files ? 207
Lines ? 13777
Branches ? 0
=========================================
Hits ? 11506
Misses ? 2271
Partials ? 0 |
Add MMD detector using the KeOps (PyTorch) backend to further accelerate drift detection and scale up to larger datasets. This PR needs to be made compatible with the optional dependency management (incl. #538 and related).
This PR includes:
sigma_meanvs.sigma_medianand make foolproof.infer_sigmacheck.torchandtensorflow._mmd2-> check if results match that of the PyTorch implementation.sigma_meanfor both "usual" (non-batch) and batch setting (unusual and should probably use the first batch entry since it corresponds to the original(x, y)).Once this PR is merged, it will be followed up by a similar implementation for the Learned (Deep) Kernel detector.