[timeseries] Fix prediction failing if context contains < 3 observations #4070
Conversation
This file contains hidden or 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
Contributor
|
2 similar comments
Contributor
|
Contributor
|
shchur
force-pushed
the
fix-prediction-failing-if-series-have-one-observation
branch
from
f309903 to
a3d0bc4
Compare
Contributor
|
canerturkmen
approved these changes
Apr 10, 2024
Contributor
canerturkmen
left a comment
canerturkmen
left a comment
There was a problem hiding this comment.
LGTM! Thanks! Just one question.
| def _deferred_init_params_aux(self, dataset: TimeSeriesDataFrame) -> None: | ||
| """Update GluonTS specific parameters with information available only at training time.""" | ||
| self.freq = dataset.freq or self.freq | ||
| if not self.freq: |
Contributor
There was a problem hiding this comment.
1/ Why remove self.freq = dataset.freq or self.freq? We never depend on deferred initialization for freq in any other model? Is freq always set for all models by the trainer?
2/ Why remove this assertion? maybe assert self.freq?
Collaborator
Author
There was a problem hiding this comment.
- No other model currently depends on deferred initialization of
freq(both local & MLForecast models assume thatfreqis provided at creation time); trainer always setsfreqwhen creating models (https://github.com/autogluon/autogluon/blob/master/timeseries/src/autogluon/timeseries/trainer/abstract_trainer.py#L576) - Assuming you are referring to the assertion inside
SimpleGluonTSDataset.__init__: after pandas 2.2 upgrade, we will anyway need to remove this part completely and always feed a dummy frequencyWto GluonTS. So the assertion would be removed anyway after [WIP] Upgrade pandas to 2.2 #4074 is merged.
shchur
deleted the
fix-prediction-failing-if-series-have-one-observation
branch
LennartPurucker
pushed a commit
to LennartPurucker/autogluon
that referenced
this pull request
Jun 1, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Description of changes:
data.freqof the data passed tomodel.predict(data)when generating the forecast. It may happen that some time series indatapassed topredictcontain just a single observation. In this case,freqinference is impossible and the model fails.self.freqattribute of the model instead ofdata.freqwhenever frequency is requireddata.fill_missing_values()works even nofreqis available (as this is too strong of a requirement for this method)By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.