Releases: thieu1995/IntelELM
Releases · thieu1995/IntelELM
v1.1.1
- Update seed value in all 4 classes to ensure reproducibility of your results
- Add mode, n_workers, and termination parameter in model.fit() of MhaElmRegressor and MhaElmClassifier classes
- These parameters are derived from Mealpy library
- With mode parameter, you can speed your training model
- With n_workers, you can set the number of threads or CPUs to speed up the training process
- With termination, you can set early stopping strategy for your model.
- Update docs, examples, and tests.
v1.1.0
v1.0.3
v1.0.2
v1.0.1
- Add "evaluate" function to all Estimators (ElmRegressor, ElmClassifier, MhaElmRegressor, MhaElmClassifier)
- Add new module "scaler"
- Our scaler can be utilized with multiple methods.
- Add "save_loss_train" and "save_metrics" functions to all Estimators
- Add "save_model" and "load_model" functions to all Estimators
- Add "save_y_predicted" function to all Estimators
- Update all examples and documents
v1.0.0
- Add supported information for each classes.
- Restructure intelelm module to based_elm module and model subpackage that includes mha_elm and standard_elm modules.
- Add traditional/standard ELM models (ElmRegressor and ElmClassifier classes) to standard_elm module.
- Add examples and tests for traditional models
- Add score and scores functions to all classes.
- Fix bug calculate metrics and objective in ELM-based models.
- Add examples with real-world datasets and examples with GridsearchCV to tune hyper-parameters of ELM-based models.
- Add documents