Fixing bugs

Fixing bugs

v0.5.0

News

• Feature/add module
• Feature/module plugins
• Code Refactoring

v0.4.0

News

• Better support to streamlit
• Add interpret-ml on Wrapper and Trainer modules
• Add SkLearn Unsupervised Class in Trainer module
• Support to new feature selection algorithms
• Many bug fixes

v0.3.4

Enhancements

• Wrapper has 2 functions( train_interpret, local_interpret) to interpret model results.
• TrainerSklearn class has a clause on train to train interpret if classification model.
• Example_notebook has a train_interpret example
• Add new class TrainerSklearnUnsupervised

and minor bug fixes.

v0.3.3

News:

• Wrapper has 2 functions( train_interpret, local_interpret) to interpret model results.
• TrainerSklearn class has a clause on train to train interpret if classification model.
• Example_notebook has a train_interpret example
• Add new class TrainerSklearnUnsupervised

v0.3.2

Bug Fixes

• Regression Analysis plot bug fix on pull request #22

v0.3.1

• Remove dependency with conda, and Anaconda Python.
• Builds venv virtual environments.
• Close bug #21

v0.3

News

Improvements

• api folder with a Flask + WSGI API
• api/myrequests.py module with an example code on how to make requests to the API
• Dockerfile that builds a docker image of the API
• two new CLI commands, hermione build and hermione run to build images and run containers
• Implemented version is the default now

0.2

Improvements

• predict.py script to make batch predictions
• Dockerfile
• Wrapper has a get_preprocessing method to access its own preprocessing instance

Bug fixes

• Wrapper save_model method now saves the wrapper instance, not the inner dictionary
• train.pysaves the wrapper artifact in the output folder to make it easier for building a docker image
• hermione has a new predict command to call predict.py from terminal

V0.1.5

fix #10