v0.6.1

HumanSignal/label-studiov0.6.1May 18, 2020by niklub

AI Summary

This is a post-release 0.6.0 patch release focused on bug fixes and minor improvements. Key improvements include new command-line options for active learning sampling and serving local files, fixes for deployment issues on Heroku and Google Cloud, and compatibility with Anaconda environments.

Key Highlights

Added active learning mode with `--sampling=prediction-score-min` for task sampling
Enabled serving locally stored images and audios via `--allow-serving-local-files` flag
Fixed deployment issues on Heroku and Google Cloud (one-click deploy)
Fixed Docker Compose startup issues
Added HTML classification example

New Features

Active learning mode with prediction-score-min sampling
Local file serving for images and audio
HTML classification example
Anaconda environment compatibility
CONLL export support for consecutive spans

Full Release Notes

Post release 0.6.0 fixes & improvements:

- Remove individual task from UI https://github.com/heartexlabs/label-studio/issues/267
- More options for task sampling - enable _active learning_ mode using `--sampling=prediction-score-min`
- Enabling serving locally stored images and audios by `--allow-serving-local-files`
- Repaired broken default localhost https://github.com/heartexlabs/label-studio/issues/287
- Fixed one-click deploy on Heroku & Google Cloud https://github.com/heartexlabs/label-studio/issues/257
- Fixed docker compose start https://github.com/heartexlabs/label-studio/issues/291
- Fixed examples for ML backend runs https://github.com/heartexlabs/label-studio/issues/294
- Fixed audio wave rendering https://github.com/heartexlabs/label-studio/issues/195
- Added HTML classification example https://github.com/heartexlabs/label-studio/issues/289
- Compatibility with Anaconda environments
- CONLL export now works fine with consecutive spans https://github.com/heartexlabs/label-studio-converter/pull/4