removing dependency on setting the Keras backend #79
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
hstojic
commented
Aug 17, 2022
hstojic
commented
Aug 17, 2022
| @@ -103,7 +106,7 @@ def __init__( | |||
| self.distribution_class = prior.__class__ | |||
| self.encoder = encoder | |||
| self.compositor = ( | |||
| compositor if compositor is not None else tf.keras.layers.Concatenate(axis=-1) | |||
| compositor if compositor is not None else tf.keras.layers.Concatenate(axis=-1, dtype=default_float()) | |||
There was a problem hiding this comment.
this needed to be passed an explicit dtype as well
hstojic
commented
Aug 17, 2022
| (fit_adam, fit_adam, {}), | ||
| (fit_adam, keras_fit_adam, {}), | ||
| (fit_natgrad, fit_natgrad, {}), | ||
| (fit_natgrad, keras_fit_natgrad, dict(atol=1e-7)), |
There was a problem hiding this comment.
this test required lowering tolerance slightly...
hstojic
commented
Aug 17, 2022
|
|
||
|
|
||
| @pytest.mark.parametrize( | ||
| "svgp_fitter, keras_fitter, tol_kw", | ||
| [ | ||
| (fit_adam, _keras_fit_adam, {}), | ||
| (fit_natgrad, _keras_fit_natgrad, dict(atol=1e-8)), | ||
| (fit_natgrad, _keras_fit_natgrad, dict(atol=1e-6)), |
There was a problem hiding this comment.
this one as well... seems like there was an issue with this one already, so I assumed it might be ok?
but not sure where the small differences arise really, perhaps keras is still somewhere using float32, not sure how exactly to check that - any idea?
hstojic
commented
Aug 17, 2022
Comment on lines
-32
to
-33
| tf.keras.backend.set_floatx("float64") | ||
|
|
There was a problem hiding this comment.
didnt try to run this as it needed additional packages
hstojic
commented
Aug 17, 2022
vdutor
reviewed
Aug 31, 2022
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.
this PR addresses #76, fix was fairly straightforward, the rest was removing backend calls throughout the repo
there was a tiny bit discrepancy in some tests in
GPflux/tests/integration/test_svgp_equivalence.py, couldn't pinpoint where it was coming fromwhile this avoids setting the backend, one still needs to be careful that float64 data is given to the model (which is the case in all the tests at the moment), @vdutor how do you want to go about this? just warn the user in docstrings and some notebooks, or implement data check in
DeepGPclass (perhaps somewhere else as well?)?