v0.16.3

Major Changes:

• pySIPFENN can now handle ONNX models with vector outputs (e.g., multi-property/multi-taks/MTL models)! This was done by adjustment of the low-level makePredictions function, which now automatically populates predictedProperties list then used by result-outputting functions. This change does maintain backwards-compatibility, so no workflow based around single-property models needs to be adjusted. In a future release, the naming scheme will be possible to be automatically overridden from the level of models JSON / YAML definitions.

Minor Changes:

• README and documentation updates; listing new information and pointing to the up-to-date scientific articles.
• Small bugfix by @ricardonpa in #22
• Official support for Intel MacOS has been dropped due to GitHub deprecating the test runners.

Full Changelog: v0.16.2...v0.16.3