Where Are Labeling Results Stored?
The ModelArts console provides data visualization capabilities, which allows you to view detailed data and labeling information on the console. To learn more about the path for storing labeling results, see the following description.
Background
When creating a dataset in ModelArts, set both Input Dataset Path and Output Dataset Path to OBS.
- Input Dataset Path: OBS path where the raw data is stored.
- Output Dataset Path: Under this path, directories are generated based on the dataset version after data is labeled in ModelArts and datasets are published. The manifest files (containing data and labeling information) used in ModelArts are also stored in this path. For details about the files, see Directory Structure of Dataset Versions.
Procedure
- Log in to the ModelArts console and choose Data Management > Datasets.
- Select your desired dataset and click the triangle icon on the left of the dataset name to expand the dataset details. You can obtain the OBS path set for Output Dataset Path.
Before obtaining labeling results, ensure that at least one dataset version is available.
Figure 1 Dataset details
- Log in to the OBS console and locate the directory of the corresponding dataset version from the OBS path obtained in 2 to obtain the labeling result of the dataset.
Figure 2 Obtaining the labeling result
Data Management FAQs
- Are There Size Limits for Images to be Uploaded?
- What Do I Do If Images in a Dataset Cannot Be Displayed?
- How Do I Integrate Multiple Object Detection Datasets into One Dataset?
- What Do I Do If Importing a Dataset Failed?
- What Do I Do to Import Locally Labeled Data to ModelArts?
- Why Does Data Fail to Be Imported Using the Manifest File?
- Where Are Labeling Results Stored?
- How Do I Download Labeling Results to a Local PC?
- Why Cannot Team Members Receive Emails for a Team Labeling Task?
- How Data Is Distributed Between Team Members During Team Labeling?
- How Do I Merge Two Datasets?
- Does Auto Labeling Support Polygons?
- What Do the Options for Accepting a Team Labeling Task Mean?
- Why Are Images Displayed in Different Angles Under the Same Account?
- Do I Need to Train Data Again If New Data Is Added After Auto Labeling Is Complete?
- Why Does the System Display a Message Indicating My Label Fails to Save on ModelArts?
- Can One Label By Identified Among Multiple Labels?
- Why Are Newly Added Images Not Automatically Labeled After Data Amplification Is Enabled?
- How Do I Use Soft-NMS to Reduce Bounding Box Overlapping?
- How Do I Add Images to a Validation or Training Dataset?
- Can I Customize Labels for an Object Detection Dataset?
- What ModelArts Data Management Can Be Used for?
- Why Is My New Dataset Version Unavailable in Versions?
- How Do I View the Size of a Dataset?
- How Do I View Labeling Details of a New Dataset?
- How Do I Export Labeled Data?
- How Do I Split a Dataset?
- How Do I Delete a Dataset Image?
- Why Is There No Sample in the ModelArts Dataset Downloaded from AI Gallery and Then an OBS Bucket?
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbotmore