ivis: dimensionality reduction in very large datasets using Siamese Networks

Transition ivis to TensorFlow 2.0

ivis: dimensionality reduction in very large datasets using Siamese Networks

Added support for supervised multi-label dimensionality reduction.

ivis: dimensionality reduction in very large datasets using Siamese Networks

A number of major additions:

• Support for both classification- and regression-type supervision
• Access to all Keras losses for supervised dimensionality reduction
• Bug fixes and performance improvements

ivis: dimensionality reduction in very large datasets using Siamese Networks

This release introduces a number of new features into ivis:

• Windows support
• Code changes to support ivis on Python2
• R package received a major facelift - with big thanks to JOSS reviewers
• Added cosine distance metric in triplet loss function
• Minor bug fixes and performance improvements

1.2.4

Appropriate contribution assignments.

1.2.3-joss

ivis release following feedback from JOSS review.

1.2.3

Support for sparse matrices in supervised mode
Bug fixes

1.2.2

Added callbacks and sanity checks during module imports

1.2.1

Bug fixes and cleanup

1.2.0

Supervised mode added to ivis. Additional features:

• Add classification_weight parameter to allow users to tune balance between classification vs. triplet loss.
• Add Ivis callbacks module for ivis-specific callbacks such as checkpointing during training. Ivis object code changed to deal with provided callbacks.
• Tensorboard callbacks
• Sparse matrix support in supervised mode