## [![Benchmarking ML Training Throughput on Tigris Object Storage](/blog/assets/images/ml-bench-entitlement-4734acfe3fd514f32dc9f23e8822f2f0.webp)](/blog/training-object-storage/.md)[Benchmarking ML Training Throughput on Tigris Object Storage](/blog/training-object-storage/.md)

Training ML models directly from object storage is possible with the right access patterns and client tooling. We benchmarked Tigris against AWS S3 and introduced the Tigris Acceleration Gateway (TAG), a local caching service that significantly reduces epoch duration and worker requirements for mu…

Ovais Tariq · March 17, 2026 ·

<!-- -->

14 min read