AWS Builder Center is the official home for builders on AWS. Share and read what others are working on, follow people who inspire you, explore training and workshops, and find tools to support what you're building.
SageMaker Data Wrangler throws an error when trying to export images
Hi there, I am trying to transform a bunch of images using Data Wrangler. I have a 3 node workflow consisting of
- Loading images from S3
- Rotating the images
- Exporting to S3
When trying to export the images I am getting the following error:
Variable '$input' got invalid value 'IMAGE' at 'input.outputConfig.config.fileFormatType'; Value 'IMAGE' does not exist in 'FileFormatType' enum.
In my export settings I have not changed anything special. The only file type for exporting is "Image" and "image_col" is the only column I can select.
I am grateful for any help or advice.
- Topics
- Machine Learning & AI
- Language
- English
asked 2 years ago209 views
- Newest
- Most votes
- Most comments
I apologize for the inconvenience you're experiencing with Amazon SageMaker Data Wrangler when trying to export your transformed images. The error you're encountering suggests that there might be a mismatch between the expected file format types and the "Image" option you're selecting.
When working with image data in Data Wrangler, the process typically involves importing, transforming, and exporting. However, it's important to note that for image data, Data Wrangler exports the transformations you've applied to all the images, rather than the actual transformed image files themselves.
The error message you're seeing indicates that the system is not recognizing "IMAGE" as a valid file format type for export. This could be due to limitations in the current version of Data Wrangler you're using or a potential bug in the export process for image data.
To work around this issue, you might consider the following approaches:
-
Instead of using the direct export option, try creating a destination node in your Data Wrangler flow. This might provide more flexibility in specifying the output format.
-
Check if there's an option to export your transformations as a Python script. This would allow you to apply the transformations to your images programmatically outside of Data Wrangler.
-
Ensure you're using the latest version of SageMaker Studio and Data Wrangler, as there might be updates that address this issue.
-
If the problem persists, you may need to consider alternative methods for applying your image transformations, such as using SageMaker Processing jobs or custom Python scripts to handle the image rotation and export.
If none of these suggestions resolve the issue, I recommend reaching out to AWS Support for further assistance, as they can provide more specific guidance based on your exact setup and the current state of the Data Wrangler service.
Sources
Prepare ML Data with Amazon SageMaker Data Wrangler - Amazon SageMaker
Transform data - Amazon SageMaker AI
Export - Amazon SageMaker AI
answered 2 years ago
Relevant content
asked 2 years ago
asked 3 years ago
This is sadly not helpful. Apart from the option to export to S3, I don't see any other options. I have the same issues as above and therefore I am currently stuck with my DataFlow which I can't further use.