What changed
- Second dummy data with noisy labels (see
misc\synthetic_data.ipynb
). - Removed the support for the Anaconda environment, all the packages are specified in
requirements.txt
(see this tutorial). - Upgraded
pytorch-lightning
anddask
packages to the newer versions. - Fixed the issue with the data augmentation described in here.
- Added new experiment (using two data sources) in
workflow_two_datasets.ipynb
. - Cleaned typos