- Updates certain package versions in requirements.txt and setup.py to address package dependency compatibility issues.
- Changes
create_tfrecords
andcheck_tfrecords
toconvert
andinspect
respectively - Adds
convert_and_load
function - Changes flexible schema to use
dataclasses
- Adds automated testing for notebooks
- Minor fixes and usability improvements
- Adds note regarding DataFrame header specification in README.md.
- Adds more informative error message when input is missing expected column
- Adds feature to load dataset from TFRecords generated by TFRecorder.
- Adds feature to generate TFRecords from an image directory.
- None known
- Python 3.8 is not compatible due to upstream requirements.
- Added flexible schemas - TFRecorder now works with any input schema, but the 'Image CSV' schema is still default.