HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization
This work addresses training stability issues in transformers, which are critical for large language models, but it is incremental as it builds on existing normalization methods.
The paper tackles the problem of unstable training in deep transformer networks by proposing HybridNorm, a hybrid normalization strategy that integrates Pre-Norm and Post-Norm, resulting in consistent performance improvements across multiple benchmarks for large-scale transformer models.
Transformers have become the de facto architecture for a wide range of machine learning tasks, particularly in large language models (LLMs). Despite their remarkable performance, challenges remain in training deep transformer networks, especially regarding the position of layer normalization. While Pre-Norm structures facilitate more stable training owing to their stronger identity path, they often lead to suboptimal performance compared to Post-Norm. In this paper, we propose $\textbf{HybridNorm}$, a simple yet effective hybrid normalization strategy that integrates the advantages of both Pre-Norm and Post-Norm. Specifically, HybridNorm employs QKV normalization within the attention mechanism and Post-Norm in the feed-forward network (FFN) of each transformer block. We provide both theoretical insights and empirical evidence demonstrating that HybridNorm improves gradient flow and model robustness. Extensive experiments on large-scale transformer models, including both dense and sparse variants, show that HybridNorm consistently outperforms both Pre-Norm and Post-Norm approaches across multiple benchmarks. These findings highlight the potential of HybridNorm as a more stable and effective technique for improving the training and performance of deep transformer models. Code is available at https://github.com/BryceZhuo/HybridNorm.