Intelligent Engine Release Notes¶
This page lists the Release Notes for Intelligent Engine, so that you can learn its evolution path and feature changes.
2024-05-30¶
v0.5.0¶
Features¶
- Added Support for adding
Tensorboard
analysis dashboard when creating tasks withbaizectl
. - Added Support for binding
Job
to custom environments created inEnvironment Management
. - Added Optimizations for custom environment configuration updates and improvements to the
Python
version selector inEnvironment Management
. - Added Support for viewing resource monitoring dashboards in the details of
Inference Service
. - Added Support for binding
Inference Service
to custom environments created inEnvironment Management
.
Fixes¶
- Fix the issue where
Python
version prompts permission problems in certain cases within environment management. - Fix the issue where the inference service does not support stopping during exceptions.
2024-04-30¶
v0.4.0¶
Features¶
- Added
Notebook
now supports local SSH access, compatible with various development tools such asPycharm
,VS Code
, etc. - Added Upgrade
Notebook
image to support the built-inCLI
toolbaizectl
, for command-line task submission and management. - Added
Notebook
adds affinity scheduling strategy configuration. - Added Distributed training tasks can now configure
SHM size
through the UI. - Added One-click restart function for training tasks.
- Added Model training tasks support custom cluster scheduler specification.
- Added Training task analysis tool
Tensorboard
support, can be launched with one click inNotebook
and training tasks. - Added When editing queue quotas, hints are provided for the shared resource configuration of the current workspace.
- Added Upgrade and adapt Kueue version
v0.6.2
.
Fixes¶
- Fixed Occasional sync anomaly issue with
Notebook
CRD
. - Fixed The query interface for
Notebook
affinity configuration parameters did not return.
2024-04-01¶
v0.3.0¶
Features¶
- Added the Notebooks module, supporting development tools like
Jupyter Notebook
. - Added the Job Center module, supporting the training of jobs with various mainstream development frameworks such as
Pytorch
,Tensorflow
, andPaddle
. - Added the Model Inference module, supporting rapid deployment of
Model Serving
, compatible with any model algorithm and large language models. - Added the Data Management module, supporting the integration of mainstream data sources such as
S3
,NFS
,HTTP
, andGit
, with support for automatic data preheating.