-
- Description:
peewit intends to facilitate programming, running and result examination of machine learning experiments. It abstains from setting ML data types, engines, or interfaces, what are to be provided by the user herself or by other libraries, for instance the python-weka-wrapper, https://mloss.org/software/view/548/.
The user is to decompose the experimental code into named nodes that correpsond to aspects or dimension of an experiment. This breakdown gives the machine a grip on the compounds, such that it can support the user in
- keeping track of the dimensions in code, stored (intermediate) results and plots
- simple parallelization
- storing and reintergration of intermediate results.
It also provides some generall houskeeping services like
- regime on the file-names and paths, or automatic storing
- recovering former experiment versions.
Be aware that the prototype is under development and many features are experimental. Even the underlying experimental model may be modified from time to time.
The core modules depend on python-3 and numpy only but for full functionality you further have to provide unison, ssh with agent, git, matplotlib, and graphviz as well as libsvm for the example project. Feel free to contact us for questions.
- Changes to previous version:
v-cube with side-cubes
Comments
No one has posted any comments yet. Perhaps you'd like to be the first?
Leave a comment
You must be logged in to post comments.