Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

55 review models architecture #75

Merged
merged 12 commits into from
Jan 6, 2025
Merged

Conversation

ben-jy
Copy link
Collaborator

@ben-jy ben-jy commented Dec 18, 2024

In this PR, model architecture was improved. Specifically, this was done :

  • BenchmarkModel wrapper was removed and replace by a BenchmarkModelConfig class, that do not act as a model, but only contain an onTime Model, which it can init dynamically with the given configuration (closes Improve benchmark objects instanciation #68).
  • Evaluation for benchmarking was moved to a specific BenchmarkEvaluator class, and make wrapper around libraries much simpler and flexible.
  • A wrapper around Pytorch model was implemented
  • The onTime Model kind of factory has been updated, so that it can handle model instance as well as model class when creating a specific model (closes Check model choice with instance type in model.py #46)
  • Clean benchmark code (closes Clean benchmark code #71)

…at it can perform fit and predict and perform multivariate prediction even when the model is univariate.
…model class as well as model class instance. Update data module and dataset creation with torch. Update docs accordingly.
…his allows to ensure that evaluation pipeline is common to all models. Still, model must ensure to be able to process batch of data for inference (predict method).
…ata (a list of time series for now). It has been implemented for all models except scikit-learn compatible ones. Update documentation accordingly, and update dependencies.
…g holder. Refactor and clean some of the code.
…s and make some modification in model config class to pass dynamic parameters
…fixes in some modules. Benchmark is now running succesfully and we get the expected results.
@ben-jy ben-jy requested a review from fredmontet December 18, 2024 17:38
@ben-jy ben-jy self-assigned this Dec 18, 2024
@ben-jy ben-jy linked an issue Dec 18, 2024 that may be closed by this pull request
7 tasks
Copy link
Collaborator

@fredmontet fredmontet left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Great work

@@ -1,6 +1,6 @@
from .detection import detectors, abstract_detector
from .generation import generators, abstract_generator
from .modelling import Model, models, abstract_model
from .modelling import Model, model_interface, models
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

change to abstract_model

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

updated in new commit

# Darts Models
self.model = DartsForecastingModel(self.model, **self.params)
# This take all the sklearn regressors and pipelines
elif issubclass(self.model, BaseEstimator):
elif is_subclass_or_instance_of_subclass(self.model, BaseEstimator):
if size_of_ts > 1:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Make an issue to group the SkForecaster classes here so that the if isn't imbricated.

Copy link
Collaborator Author

@ben-jy ben-jy Jan 6, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done (#78)

Copy link
Collaborator

@fredmontet fredmontet left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I now approved

@ben-jy ben-jy merged commit 8748f30 into develop Jan 6, 2025
@ben-jy ben-jy deleted the 55-review-models-architecture branch January 6, 2025 13:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
2 participants