Trackable Version Control
Enable version control of raw data, annotations, and metadata, without having to manually record the changes of the dataset. Manage data versions like Git and easily trace the iteration process of data, sync your data iteration with AI product iteration, and use data to optimize products more conveniently
Start for Free
version banner
Clear Version Control
Similar to Git, you can record changes on data through Commit
Version information can only be recorded with different folders or Excel documents
Every time a version is created, the previous version must be fully copied, which is laborious and time-consuming, resulting in a large amount of redundant data
Commit edited data in a single click and automatically record data versions
Comment on version information and tags
Version control only indexes changes of data to avoid storage waste caused by data duplication
Switch Versions and View Data Swiftly
Consistent version control of the raw data, annotations, and metadata. Different data versions of the dataset can be switched and accessed through different Commit, and the complete data information of each version can be viewed online
Separating storage and management of raw data, annotations, and metadata makes it extremely complicated to locate and compare data of different versions
Flexible version switching and quick view of version data details make the comparison of different versions an easy process
Version Iteration Visualization
By visualizing data versions, the data iteration process is clearly visible
Limited ways to record data iteration make it hard to view the data iteration clearly
The management cost of iterations on previous versions is high, so it is difficult to use data versions to promote the continuous model optimization
The data iteration process is clearly visible by presenting versions in list and tree view
Data iteration continues based on the previous versions, syncing with the model iteration
how it works
How It Works
how it works
When your data changes, you can commit a version, leaving messages and assigning tags to it. You can also select a version, see if the data meets your needs, and start a new iteration based on the version via developer tools
Create a Dataset
Create a Dataset Version
Commit at any time during any stage of data processing to create a version
View Previous Versions
View the version's data list through Manage Data on the details page or via developer tools
Manage Version Tags
Flexibly edit the version tags to quickly locate previous versions
Read Data Versions
Access different data versions via developer tools
Start Building AI Now