Release notes: 2020-09-04
- 2 minutes to read
- Logging fix for
microsoft-bonsai-api. Version v0.1.2 fixes a bug where
enable_logging argument in
BonsaiClientConfig was not printing debug
logs. The new client is available on PyPi.
- Updated Reinforcement Learning algorithm support. The training engine
now uses a "pre-training" phase to make better decisions about which RL
algorithm to use.
- Algorithm choice preserved across training sessions. If you pause and
resume training on a pre-trained brain, it preserves the algorithm selected
- Algorithm information available in UI. Algorithm selection by the
training engine now appears in the UI as a notification message.