Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

ORBIT-2: Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling

Conference ·
 [1];  [1];  [2];  [1];  [1];  [1];  [1];  [1];  [3];  [1];  [1];  [4];  [1];  [1];  [1];  [1];  [1]
  1. ORNL
  2. Fujitsu Research America
  3. AMD
  4. AIST, Japan

Sparse observations and coarse-resolution climate models limit effective regional decision-making, underscoring the need for robust downscaling. However, existing AI methods struggle with generalization across variables and geographies and are constrained by the quadratic complexity of Vision Transformer (ViT) self-attention. We introduce ORBIT-2, a scalable foundation model for global, hyper-resolution climate downscaling. ORBIT-2 incorporates two key innovations: (1) Residual Slim ViT (Reslim), a lightweight architecture with residual learning and Bayesian regularization for efficient, robust prediction; and (2) TILES, a tile-wise sequence scaling algorithm that reduces self-attention complexity from quadratic to linear, enabling long-sequence processing and massive parallelism. ORBIT-2 scales to 10 billion parameters across 65,536 GPUs, achieving up to 4.1 ExaFLOPS sustained throughput and 74–98% strong scaling efficiency. It supports downscaling to 0.9 km global resolution and processes sequences up to 4.2 billion tokens. On 7 km resolution benchmarks, ORBIT-2 achieves high accuracy with R2 scores in range of 0.98–0.99 against observation data.

Research Organization:
Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
Sponsoring Organization:
USDOE
DOE Contract Number:
AC05-00OR22725
OSTI ID:
3007902
Resource Relation:
SC '25: The International Conference for High Performance Computing, Networking, Storage, and Analysis - St. Louis, Missouri, United States of America - 11/16/2025-11/21/2025
Country of Publication:
United States
Language:
English

Similar Records

ORBIT-2 Weather and Climate Downscaling Software Repository
Software · Fri Oct 17 20:00:00 EDT 2025 · OSTI ID:code-167711

ORBIT-2 Dataset for Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling
Dataset · Fri Oct 10 00:00:00 EDT 2025 · OSTI ID:2589526

ORBIT-2 Dataset for Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling
Dataset · Mon Sep 01 00:00:00 EDT 2025 · OSTI ID:3002026

Related Subjects