search

microsoft / vscode-tensorboard

Launch and view Tensorboards in VS Code
MIT License
4 stars 3 forks source link

Extension asks for directory even after setting `tensorBoard.logDirectory` #28

Open nikhilweee opened 3 weeks ago

nikhilweee commented 3 weeks ago

Environment data

Expected behaviour

When launching tensorboard, the extension should not ask for a directory when tensorBoard.logDirectory is set.

Actual behaviour

The extension asks for a directory even when tensorBoard.logDirectory is set.

Steps to reproduce:

[NOTE: Self-contained, minimal reproducing code samples are extremely helpful and will expedite addressing your issue]

  1. Open settings JSON and edit tensorBoard.logDirectory.

    {
"tensorBoard.logDirectory": "/mnt/train/"
}

  2. Open command palette and launch tensorboard Python: Launch Tensorboard

  3. The extension opens a quickpick menu and asks for a directory.

Logs

Output for Jupyter in the Output panel (View→Output, change the drop-down the upper-right of the Output panel to Jupyter)

``` N/A ```

nikhilweee commented 3 weeks ago

The reason I am using a release version (v2023.10.1002992421) and not a newer pre-release version is because I get an error when switching to a pre-release version (I tried v2024.3.1001511100).

Failed to start a TensorBoard session due to the following error: Extension 'ms-toolsai.tensorboard' CANNOT use API proposal: portsAttributes. Its package.json#enabledApiProposals-property declares: but NOT portsAttributes. The missing proposal MUST be added and you must start in extension development mode or use the following command line switch: --enable-proposed-api ms-toolsai.tensorboard

I can get the pre-release version to work without this issue but for that I need to run Preferences: Configure Runtime Arguments and edit argv.json to add a special exception for the extension.

"enable-proposed-api": [
  "ms-toolsai.tensorboard"
]