AWS helps seismic AI model training jump from 6 months to 5 days

Maya Chen

AI Business Correspondent3d agoAWS ML Blog✓Verified across 1 source

The Brief

TGS used Amazon SageMaker HyperPod to achieve near-linear scaling for their Vision Transformer seismic foundation model, reducing training time by 97% while expanding context windows for larger data analysis. The breakthrough enables faster oil and gas exploration modeling.

✓Verified across 1 independent source

Sources

01https://aws.amazon.com/blogs/machine-learning/scaling-seismic-foundation-models-on-aws-distributed-training-with-amazon-sagemaker-hyperpod-and-expanding-context-windows/

AWS helps seismic AI model training jump from 6 months to 5 days

Google Releases MedGemma 1.5, an Open Medical AI Model for CT Scans, MRIs, and Clinical Records

Apple Research Finds Optimal Mix of Real and Synthetic Training Data

Apple Releases ProText Benchmark to Measure AI Misgendering in Long-Form Text