v1.8.5

• Bug Fixed:
  • Fixed a bug when the SPNAS algorithm cluster training fails.
  • Fixed bugs such as model copy failure in safe mode.

v1.8.4

• Bug Fixed:

  • Fixed bug that ASHA failed to update data.
  • Fixed bug that loss is not updated on HCCL+Apex.
  • Add dictionary metrics.
  • Update the security configuration document.
  • Not Allowed Horovod and TensorFlow in safe mode.
  • Python 3.9 or later is required in the security model.

v1.8.3

Bug Fixed:

• Fixed bug of updating report.
• Fixed bug of Horovod training in multi-nodes.
• Fixed docs error of vega-kill.

v1.8.2

• Bug Fixed:

  • Fixed bad document links.
  • The model to be evaluated supports multiple imputs.
  • Fixed bug when using Apex on the NPU.

v1.8.1

• Bug fix:

  • The shared directory was not set when the evaluation service was started.
  • Failed to load the pre-trained model.

v1.8.0

• Feature enhancement:

  • Security enhancement: Security protocols communication.
  • Provide evaluation service release package.
  • Update the auto-lane model and provide auto-lane inference sample code.

v1.7.1

• Bug fixes:

  • Maximum number of evaluation service attempts.
  • Use SafeLoader to load the YAML file.
  • Catch evaluation service input parameter exceptions.

v1.7.0

• Feature enhancement:

  • Releases Ascend MindStudio version.
  • Provides data parallel training capabilities for Horovod (GPU) and HCCL (NPU).
  • Fixed bug: The BOHB algorithm may not automatically stop after more than three rounds.

v1.6.1

• Bug Fixes:
  • Evaluation time error in log.
  • Updating error model description while updating record.

v1.6.0

• Feature enhancement:

  • Supports simple quota settings, for example, quota: flops < 11.2 and params in [34.0, 56.0].
  • Supports running Vega in a Python virtual environment.
  • Supported running environments: Python 3.8 and PyTorch 1.9.
  • Fixed some bugs with parallel training and distributed search.