Error in Creating a Labelling Project in Azure

This is for an Azure Computer Vision project. I created a dataset from a container in a storage account. But when I tried to create a labelling project, 'Add Azure ML Data Labeling project to dataset failed' this error message was displayed. How to fix this?

Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.
romungi-MSFT 45,961 Reputation points Microsoft Employee

2024-01-11T06:37:24.3766667+00:00

@Malith Srineth Amarawickrama Are you referring to this documentation or steps that are failing in this case? For creating custom models, there is a pre-req to request access for training custom models. Is your subscription approved to use the same? If yes, then creating a labeling project in Azure ML requires an Azure ML workspace and the subsequent steps should create a new workspace and use the labeling project from azure ML portal. I think the error above indicates that there could some issues in your storage account being accessed by Azure ML workspace. Is the storage account used while dataset creation from vision studio open to all networks?
romungi-MSFT 45,961 Reputation points Microsoft Employee

2024-01-17T10:08:58.04+00:00

@Malith Srineth Amarawickrama Did you get a chance to review the above comments? Thanks!!
Vibhor Jain 30 Reputation points

2024-01-19T10:48:28.0633333+00:00

@romungi-MSFT I was giving assignment to gain credentials and certifications. There I had to create a custom image classification project given their credentials. Here is the link for that computer vision assessment. Today when I started to try it out with my own account and subscription, I got the same error. I have requested access now but I don't think that would be possible in the assessment. Yes I have checked that my storage account is accessible from all networks.
Abhijat 10 Reputation points

2024-01-21T11:52:06.0066667+00:00

If there is a solution, please let me know. I faced the same error too.
Mailendiran S 5 Reputation points

2024-01-23T14:53:17.0733333+00:00

Any solution for this? I'm also using it in assessment and getting 'Add Azure ML Data Labeling project to dataset failed' error.
Khushal Patel 0 Reputation points

2024-01-26T04:32:05.4933333+00:00

facing the same issue
Vivek Koppu 0 Reputation points

2024-01-27T17:54:33.8533333+00:00

I am also getting same error. Is there any solution

1 answer

Vibhor Jain 30 Reputation points

2024-01-21T12:45:58.3766667+00:00

I was able to do it with Azure Machine Learning Labeling. After having created a workshop in computer vision portal. Go to the resource and workshop created. With project created add and classify dataset. Export the labels in Coco file, download it, add it to same blob container. At the vision portal use import drop-down and import from blob container. Your dataset is ready to be used in a model. After having tried for 2-3 times I finally got a trained model.
Please sign in to rate this answer.

5 people found this answer helpful.
Mailendiran S 5 Reputation points

2024-01-27T05:20:10.5566667+00:00

It worked, thanks. If someone is having doubts , reach me on linkedin
https://www.linkedin.com/in/mailendiran/

Manas Mohanty (QUANTUM LEAP CONSULTING PRIVAT) 0 Reputation points Microsoft Vendor

2024-01-29T14:13:41.04+00:00

In the same context, You can also create a labelling project from the workspace you are creating from AI vision and select the labelling project if it is already fully labelled. You can select a vendor to do the labelling on your behalf or can do the manual labelling on the image dataset and export coco json file if needed as mentioned by Vibhor. On UI it is there even with ML studio permission and trying with public storage . Have Reached and Requested the concerned product team for a bug fix from UI side already.

Praveen Choudhary 0 Reputation points

2024-01-29T16:31:35.91+00:00

I have tried this method of downloading the labeled COCO file and uploaded it but in the model training the status as "Failed (DatasetCorrupt)" P.S.- Never mind it still accepted it and passed the assessment with 100%

Vivek Koppu 0 Reputation points

2024-01-30T18:02:59.11+00:00

Thank you very much. The solution helped me to solve a problem.
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

Error in Creating a Labelling Project in Azure

1 answer

Your answer