Assigned
Status Update
No update yet.
Comments
sh...@google.com
Screenshots are attached
sh...@google.com
The AI Platform (Unified) engineering team is aware of this issue and are investigating implementations. There is no ETA at this time for a release, but all further updates and communications should occur here.
mi...@google.com
Vertex AI (previously know as AI Platform Unified) is currently not supporting image segmentation, you can see the message when you select this option, I have attached a screenshot. As specified in the message, you can opt to create a custom-trained model.
Currently this functionality is in preview, please see link[1] for more information about the launch stages of GCP products , I suggest you to monitor the release notes to see when this feature will be released [2].
[1]https://cloud.google.com/products#product-launch-stages
[2]https://cloud.google.com/vertex-ai/docs/resources/release-notes?hl=es-419
Currently this functionality is in preview, please see link[1] for more information about the launch stages of GCP products , I suggest you to monitor the release notes to see when this feature will be released [2].
[1]
[2]
ma...@google.com
Hello, kindly, do you have an ETA when image segmentation might be available in Vertex AI Datasets?
Description
We are unable to save the image label in our dataset.
== Summary==
We would like to annotate/ label our images for model training. We tried as per the steps bellow in the GCP Console > AI Platform (Unified) > Dashboard
== What you expected to happen:==
1) Once the annotation is done on the image portion, it should be saved for further use.
2) There should be a 'SAVE' button to same the annotation _OR_ auto saving feature
3) If we go to the Dashboard or other pages and come back again to the the dataset the previously annotated labels should be found
== Steps to reproduce:==
1) Console > AI Platform (Unified) > Dashboard
2) Select Region from drop-down
3) Create Dataset
4) Dataset name
5) Select a data type and objective: select option "Image segmentation" from the 'IMAGE' tab
6) Hit "Create" button
7) Add images to your dataset: Select Option 'Polygon'
8) Select an import method: Upload images from your computer
9) Hit "SELECT FILES" button
10) Select your image
11) Select a Cloud Storage path: BROWSE (set your GCS bucket)
12) Hit "CONTINUE" button
Import in progress........
13) ADD NEW LABEL
14) Draw polygon on your image > annotate with apprppriate label from the Popup
15) Repeat 14 as many times as you want
16) There is no 'SAVE' button to save your work
17) Click Datasets from the left pane
18) Select the dataset from the datasets list to come back again
19) You can see the name of the Labels with '0' annotations and click the image itself there is no annotations done previously
So There is no 'SAVE' button to save and annotations were not saved automatically either.