◀️Release Version
Available after project annotation is complete
JTheta.AI allows you to release multiple dataset versions to capture finalized annotations after edits. This helps with version tracking, model retraining, and audit trails.
📁 Step-by-Step: How to Release a Dataset Version
Complete Project Annotation Once all annotations are done, go to the admin Dashboard.
Click "Dataset" Open the dataset linked to your project.
Choose or Create a Version You can release multiple versions of the dataset for the same project over time.
✏️ Modify Before Release
To release a version, at least one image must be modified and submitted.
If no changes are made, you cannot release a version.
Unmodified or unsubmitted images will remain in Draft state.
Draft images can still be edited, removed, or updated before release.
✅ Submitting Images for Release
After editing, click Submit on each image to finalize it.
Once an image is submitted, it is locked for the current version.
Only submitted images will be included in the released version.
You cannot modify an image after it’s been submitted in that version.
⚠️ A dataset version can only be released after at least one image has been submitted.
🔒 Finalized Versions
Once a version is released, it becomes read-only.
You can view, export, or train models on it, but cannot make changes.
For further changes, edit draft images and release a new version.
🔄 Multiple Versions Supported
There is no limit to how many versions you can create. Each version captures the exact state of your dataset at the time of release.
🧩 Why Use Versions?
Maintain a clean version history
Roll back to previous dataset states
Compare model performance across versions
Ensure data consistency and compliance
Last updated