List of known issues for ftc-ml

This is a list of known issues for ftc-ml. Additional issues will be added once verified, and will be struck once resolved.

Issue 1 - Videos may not load properly on first load. After uploading a video, and the extraction process has completed, clicking on the description of the video will open the Label Video tool and load the video. On first load, the video may fail to load - this will be indicated in the upper-right next to the progress indicator, the text “Loading Failed” may be displayed (and the “Failed” will likely wrap around under the progress indicator bar).
Workaround for Issue 1 - Reload the page, but only after the progress bar has completed (to give the video additional time to be processed internally). Video loading should ONLY fail on first load. If a video fails to load on successive attempts, please add a topic to the Bug Report category.

Issue 2 - Clicking outside a popup will close the popup. This is a UI issue we’re aware of, and in situations where this happens it should be safe to wait up to 5 minutes to determine if the job behind the scenes has completed and updated the UI. For large tasks, the wait could be longer. If there is an error during processing, you will never see the error because the popup is not visible, and it might be necessary to try again.
Workaround for Issue 2 - Wait up to 5 minutes, and possibly try again.

Issue 3 - Can’t update bounding boxes when bounding box limit is reached. This is a UI issue where you aren’t allowed to change the bounding box size/position by dragging them once you reach the limit of 10 bounding boxes.
Workaround for Issue 3 - You can, however, modify the bounding box values in the Region Labels area (as marked in the Manual, pages 11 and 12) manually. Just be careful about not specifying points that are not physically in the bounds of the image (like selecting point 2000,2000 in a 1280x720 frame).