Transformer Architecture for NetsDB
This work provides an incremental improvement for deploying transformer models in database systems, potentially benefiting users in scalable and efficient deep learning inference.
The authors tackled the problem of implementing the transformer encoder for efficient model serving in NetsDB, a relational database system for deploying large-scale deep learning models, and demonstrated its efficacy by comparing it with existing frameworks like PyTorch and TensorFlow across metrics such as inference time and model size.
Transformers models have become the backbone of the current state-of-the-art models in language, vision, and multimodal domains. These models, at their core, utilize multi-head self-attention to selectively aggregate context, generating dynamic contextual embeddings and modeling long-range dependencies for a clear contextual understanding. Lixi et al. \cite{zhou2022serving} proposed a method to use relational databases for deploying large-scale deep learning models and created an open-source implementation called NetsDB for the same. We build upon the previous work of these authors by creating an end-to-end implementation of the Encoder part of the transformer for model serving in NetsDB. Specifically, we construct a two-block encoder that includes Multi-Head Attention and its accompanying self-attention mechanism, Layer-Norm, Dropout, FeedForward Layers, and the necessary residual connections. We load out weights from our model for distributed processing, deployment, and efficient inferencing. To prove the efficacy of our implementation, we conduct a comprehensive performance analysis by comparing it with existing implementations in PyTorch, Tensorflow, Flax, and MxNet across key metrics such as inference time and model size.