Major Features
New tuner support
- Support network morphism
Training Service Improvements
- Migrate Kubeflow training service's dependency from kubectl CLI to Kubernetes API client
- Pytorch-operator support for Kubeflow training service.
- Improvement on local code files uploading to OpenPAI HDFS.
WebUI improvements
- Enable modify concurrency number during experiment.
- Add feedback link to NNI github 'create issue' page
- Enable customize top 10 trials regarding to metric numbers (largest or smallest)
- Enable download logs for dispatcher & nnimanager
- Enable automatic scaling of axes for metric number
New example
- FashionMnist, work together with network morphism tuner.
- Distributed MNIST example written in PyTorch.
Others
- Show version information both in nnictl and WebUI. You can run nnictl -v to show your current installed NNI version.
- Bug Fix
• fix the bug that WebUI doesn't show latest trial job status, which is caused by OpenPAI token expiration.
• Update annotation to support displaying real choice in searchspace