Currently, when performing in-browser training or a raw calculation, only the labeled frame at the current time is considered, however, we want to be able to consider data across all frames (eg. if working with a tiled image). Additionally, this leads to the training machine breaking when data exists in other frames. The ability for a user to make this distinction and obtain desired results should be implemented.
Currently, when performing in-browser training or a raw calculation, only the labeled frame at the current time is considered, however, we want to be able to consider data across all frames (eg. if working with a tiled image). Additionally, this leads to the training machine breaking when data exists in other frames. The ability for a user to make this distinction and obtain desired results should be implemented.