Transformer Layer Injection: A Novel Approach for Efficient Upscaling of Large Language Models
This addresses the challenge of scaling transformer-based models for AI researchers and practitioners, offering a data-efficient and cost-effective method, though it appears incremental as it builds upon existing techniques like Depth Up-Scaling.
The paper tackles the problem of efficiently upscaling large language models by proposing Transformer Layer Injection (TLI), which reduces computational costs and maintains performance, achieving better initialization, fewer training steps, and superior accuracy on tasks like KoBEST and KMCQA with models up to 8B parameters.
In this paper, we propose Transformer Layer Injection (TLI), a novel method for efficiently upscaling large language models (LLMs) while minimizing computational costs and maintaining model performance. Model scale is a key factor in enhancing the quality of machine learning models, and TLI addresses the challenge of scaling by reducing initial loss, minimizing fine-tuning requirements, and preserving model complexity. Our approach improves upon the conventional Depth Up-Scaling (DUS) technique by injecting new layers into every set of K layers, enabling hidden representations to pass through transformer blocks with minimal disruption. We compare TLI with existing approaches, including Mixture of Experts (MoE) and DUS, and validate its efficiency through experiments on small LLMs (LLama3 1B, 3B, and 8B). Results show that TLI achieves better initialization, requires fewer training steps, and delivers superior accuracy on tasks such as KoBEST and KMCQA, with models performing effectively even without additional training. TLI is demonstrated to be both data-efficient and cost-effective, significantly outperforming existing methods. Its scalability and simplicity make it a promising solution for upscaling transformer-based models, with potential applications in scaling models from 10B to 405B parameters.