[2401.11851] BETA: Binarized Energy-Efficient Transformer Accelerator at the Edge

Pi Day is finally here – and so is Giving Day for arXiv! Donate today to directly support arXiv initiatives and help keep open science open. Infinite decimals. Infinite new ideas to discover. Infinite reasons to give. Pi Day is Giving Day: arXiv depends on donations to operate and keep science open for all. Give back to arXiv on 3.14.24! Help | Advanced Search arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. arXiv Operational Status Get status notifications via email or slack

View PDF HTML (experimental) Abstract:Existing binary Transformers are promising in edge deployment due to their compact model size, low computational complexity, and considerable inference accuracy. However, deploying binary Transformers faces challenges on prior processors due to inefficient execution of quantized matrix multiplication (QMM) and the energy consumption overhead caused by multi-precision activations. To tackle the challenges above, we first develop a computation flow abstraction method for binary Transformers to improve QMM execution efficiency by optimizing the computation…

saved by

• Pavel

related reading

• [2011.14203] EdgeBERT: Sentence-Level Energy Optimizations for Latency-Aware Multi-Task NLP Inferencearxiv.org
• [2201.06618] VAQF: Fully Automatic Software-Hardware Co-Design Framework for Low-Bit Vision Transformerarxiv.org
• FTRANS: Energy-Efficient Acceleration of Transformers using FPGAarxiv.org
• [2304.07493] OliVe: Accelerating Large Language Models via Hardware-friendly Outlier-Victim Pair Quantizationarxiv.org
• AttentionLego: An Open-Source Building Block For Spatially-Scalable Large Language Model Accelerator With Processing-In-Memory Technologyarxiv.org
• [2312.15159] Understanding the Potential of FPGA-Based Spatial Acceleration for Large Language Model Inferencearxiv.org
• [2209.10797] DFX: A Low-latency Multi-FPGA Appliance for Accelerating Transformer-based Text Generationarxiv.org
• [2403.00579] NeuPIMs: NPU-PIM Heterogeneous Acceleration for Batched LLM Inferencingarxiv.org
• Hardware Accelerator for Multi-Head Attention and Position-Wise Feed-Forward in the Transformerieeexplore.ieee.org
• [2403.06664] Smart-Infinity: Fast Large Language Model Training using Near-Storage Processing on a Real Systemarxiv.org
• A Mathematical Framework for Transformer Circuitstransformer-circuits.pub
• [2207.00032] DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scalearxiv.org