Why Cloud Storage Matters for Vision AI Training
Training computer vision models starts with data, and most teams store their images and videos in cloud buckets. Moving that data manually between storage and your ML platform wastes time and introduces version control problems. A direct connection between AWS S3 and your training environment keeps everything in sync.
What This Tutorial Covers
Datature covers the full setup for connecting an Amazon S3 bucket to Nexus so your training data flows directly into the platform. The walkthrough includes:
- Creating an S3 connection inside your Datature workspace
- Configuring IAM permissions so Nexus can read your bucket
- Selecting specific folders or prefixes to import
- Verifying that assets appear correctly in your project
The entire process takes about two minutes. Once connected, new files added to the S3 bucket can be pulled into Nexus without repeating the setup.
When to Use This Approach
S3 integration is useful when your team already has image data stored in AWS, or when your data pipeline deposits files into S3 buckets automatically. Instead of downloading and re-uploading, Nexus reads from S3 directly. This works for object detection datasets, segmentation masks, video frames, and raw image collections of any size.
For a step-by-step written guide with screenshots, see How to Connect Amazon S3 Bucket for Asset Uploading on the Datature blog.
