Tensorboards

The Tensorboards menu is used to visualize, monitor, and analyze machine learning experiments interactively, making it easier for users to access model training results, such as loss graphs, accuracy, parameter value distribution, and result space mapping (embedding) in one interface.

Create Tensorboards

Create TensorBoards is used to create a new TensorBoard instance linked to the log files of the model training experiment. On the TensorBoards page press the New TensorBoards button.

Create Tensorboards

The New TensorBoards window appears, fill in the Name and select PVC or Object Store. If you select Object Store, enter the Object link that will be used, while if you select PVC select the PVC and Mount Path to use. In Configurations, check if you are giving Pipelines access. Press the Create button to create tensorboards.

Create Tensorboards

Delete Tensorboards

On the TensorBoards page, select the TensorBoards you want to delete by pressing the Delete icon. A window will appear for you to successfully agree to delete TensorBoards, press the Delete button to continue the deletion process.