This patch updates metadata in `setup.py` and comments in other files to
clarify that the current license is AGPLv3, as is specified in the
current `LICENSE` file.
This patch also removes the "this agreement applies to this version and
all previous versions" text from the README and code comments, because
it is redundant; the previous code was already released under a _more
liberal_ license, so giving the option of AGPLv3 is not useful (and
perhaps confusing).
# 0.1.1
# - Added support for Python 3.7
# - Added full support for Windows, including multiprocessing parallelization support in cleanlab/pruning.py
# - Improved PEP-8 adherence in core cleanlab/ code.
# 0.1.0 - Release of confident learning paper based on this version of cleanlab: https://arxiv.org/pdf/1911.00068.pdf
# - Documentation increase
# - Add examples to find label errors in mnist, cifar, imagenet
# - re-organized examples and added readme.
0.0.8 - Multiple major changes
- Finding label errors is no fully parallelized.
- prune_count_method parameter has been removed.
- estimate_confident_joint_from_probabilities now automatically calibrates confident joint to be a true joint estimate.
- Confident joint algorithm changed! When an example is found confidently as 2+ labels, choose class with max probability.
Curtis G. Northcutt