Import Training Data from Google Cloud Storage into Datature Nexus
Connecting Google Cloud Storage to Your ML Pipeline
Teams running their data infrastructure on Google Cloud need a direct path from GCS buckets to their training platform. Manual file transfers break when datasets grow beyond a few thousand images, and downloading locally just to re-upload defeats the purpose of cloud storage.
What This Tutorial Covers
Datature walks through connecting a Google Cloud Storage bucket to Nexus. The process includes:
- Setting up a GCS connection in your Datature workspace
- Providing the service account credentials for bucket access
- Choosing which folders or paths to import
- Confirming assets load correctly into your project
Setup takes under two minutes. After the initial connection, pulling new data from the same bucket requires no additional configuration.
Who This Is For
This approach works well for teams that collect training images through Google Cloud pipelines, store labeled datasets in GCS, or use BigQuery alongside their vision AI workflows. Whether the data is for object detection, classification, or segmentation tasks, the import handles standard image and video formats without conversion.
If your data sits in a different cloud provider, Datature also supports AWS S3 and Azure Blob Storage connections.
