Clover-2: Accurate Inference for Regressive Lightweight Speculative Decoding
This work addresses efficiency issues in LLM inference for text generation tasks, representing an incremental improvement over prior methods.
The paper tackles the inefficiency of large language models in text generation by introducing Clover-2, an improved RNN-based draft model for speculative decoding that achieves accuracy comparable to attention decoder layers while maintaining low computational overhead, as demonstrated in experiments with Vicuna 7B and LLaMA3-Instruct 8B models.
Large Language Models (LLMs) frequently suffer from inefficiencies, largely attributable to the discord between the requirements of auto-regressive decoding and the architecture of contemporary GPUs. Recently, regressive lightweight speculative decoding has garnered attention for its notable efficiency improvements in text generation tasks. This approach utilizes a lightweight regressive draft model, like a Recurrent Neural Network (RNN) or a single transformer decoder layer, leveraging sequential information to iteratively predict potential tokens. Specifically, RNN draft models are computationally economical but tend to deliver lower accuracy, while attention decoder layer models exhibit the opposite traits. This paper presents Clover-2, an advanced iteration of Clover, an RNN-based draft model designed to achieve comparable accuracy to that of attention decoder layer models while maintaining minimal computational overhead. Clover-2 enhances the model architecture and incorporates knowledge distillation to increase Clover's accuracy and improve overall efficiency. We conducted experiments using the open-source Vicuna 7B and LLaMA3-Instruct 8B models. The results demonstrate that Clover-2 surpasses existing methods across various model architectures, showcasing its efficacy and robustness.