A single weather transformer with inference-time timestep conditioning matches specialist models across 1 to 24 hour steps and improves rollout stability under mixed-timestep training.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Timestep-Conditioned Transformers for Global Weather Forecasting
A single weather transformer with inference-time timestep conditioning matches specialist models across 1 to 24 hour steps and improves rollout stability under mixed-timestep training.