轻量级模型Sonny实现中短期天气预报，仅需单块A40显卡即可训练。
Sonny: Breaking the Compute Wall in Medium-Range Weather Forecasting
- 分两阶段设计：先慢后快，分别捕捉大气动力与热力过程。
- 在WeatherBench2上性能媲美业务系统，热带地区预报优势明显。
- 训练仅需单块A40 GPU，5.5天完成，适合算力有限团队。
天气预报对防范极端气象灾害至关重要。近年来，基于深度学习的数据驱动方法表现出色，常可媲美业务数值系统。但多数模型依赖大规模训练和高算力架构，限制了资源有限的学术机构使用。本文提出Sonny，一种高效分层变换器，在保持中短期预报竞争力的同时，计算成本可控。其核心为两阶段StepsNet设计：先通过窄速路径建模大尺度大气动力，再由全宽快速路径整合热力相互作用。为避免中长期滚动预测不稳，训练中引入指数移动平均（EMA）。在WeatherBench2上，Sonny展现稳健预报能力，性能与业务基准相当，并显著优于FastNet，尤其在热带地区延长预报时效。实际训练仅需单张NVIDIA A40 GPU，约5.5天即可收敛。
原文摘要 · Abstract (English)
Weather forecasting is a fundamental problem for protecting lives and infrastructure from high-impact atmospheric events. Recently, data-driven weather forecasting methods based on deep learning have demonstrated strong performance, often reaching accuracy levels competitive with operational numerical systems. However, many existing models rely on large-scale training regimes and compute-intensive architectures, which raises the practical barrier for academic groups with limited compute resources. Here we introduce Sonny, an efficient hierarchical transformer that achieves competitive medium-range forecasting performance while remaining feasible within reasonable compute budgets. At the core of Sonny is a two-stage StepsNet design: a narrow slow path first models large-scale atmospheric dynamics, and a subsequent full-width fast path integrates thermodynamic interactions. To stabilize medium-range rollout without an additional fine-tuning stage, we apply exponential moving average (EMA) during training. On WeatherBench2, Sonny yields robust medium-range forecast skill, remains competitive with operational baselines, and demonstrates clear advantages over FastNet, particularly at extended tropical lead times. In practice, Sonny can be trained to convergence on a single NVIDIA A40 GPU in approximately 5.5 days.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。