0.2.1

• Add supports for starting hugging face models other than classification tasks.
• Skip the model loading for the random sampling strategy.
• Add Docker support (need CD scripts).

0.2.0

• new server API with support for both local start and parameter start
• fixed some issues

0.1.7

• support more AL strategies
• fixed some issues
• support large dataset requests

0.1.6

• new framework structure
• removed deep learning dependiencies
• update documents