Pipeline text specification and CV cache ideas - Githubissues

ContinuumIO / elm

Phase I & part of Phase II of NASA SBIR - Parallel Machine Learning on Satellite Data

http://ensemble-learning-models.readthedocs.io

44 stars 23 forks source link

Pipeline text specification and CV cache ideas #214

Closed PeterDSteinberg closed 7 years ago

PeterDSteinberg commented 7 years ago

I made this branch regarding experimentation I'm doing with Elm-Earthio-NLDAS repo - I'm using the soil moisture ML model there to break/prototype elm and related tools. This PR has some ideas for the Pipeline/estimator YAML specification and for CV cache ideas related to xarray_filters. We can open separate branches to get the work done further - this is just to show some ideas.

YAML -spec

See spec_mixins.py with some ideas for serialization
- An example of .spec with a Pipeline - shown here in wiki

A couple fixes in here in addition to the new stuff above

elm.pipeline.steps was wrapping some classes as they were found by way of imports into other sklearn classes.
removes the installation of attrs

PeterDSteinberg commented 7 years ago

See this notebook showing current status of EaSearchCV