mlsanityData to 14 Oct 2026

COBRA: Algorithm-Architecture Co-optimized Binary Transformer Accelerator for Edge Inference

arXiv 2504.16269 ↗First posted 22 Apr 2025cs.AR cs.LGcited by 4

Not on the map. Why?

Abstract (arXiv)

Transformer-based models have demonstrated superior performance in various fields, including natural language processing and computer vision. However, their enormous model size and high demands in computation, memory, and communication limit their deployment to edge platforms for local, secure inference. Binary transformers offer a compact, low-complexity solution for edge deployment with reduced bandwidth needs and acceptable accuracy. However, existing binary transformers perform inefficiently on current hardware due to the lack of binary specific optimizations. To address this, we introduce COBRA, an algorithm-architecture co-optimized binary Transformer accelerator for edge computing. COBRA features a real 1-bit binary multiplication unit, enabling matrix operations with -1, 0, and +1 values, surpassing ternary methods. With further hardware-friendly optimizations in the attention block, COBRA achieves up to 3,894.7 GOPS throughput and 448.7 GOPS/Watt energy efficiency on edge FPGAs, delivering a 311x energy efficiency improvement over GPUs and a 3.5x throughput improvement over the state-of-the-art binary accelerator, with only negligible inference accuracy degradation.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 19 of its 26 references that are in the map. First posted on arXiv 22 Apr 2025.

Builds on

19 of 26 references.

References · 19

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  4. 2022FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenesscited by 3,673
  5. 2022ProgPrompt: Generating Situated Robot Task Plans using Large Language Modelscited by 611
  6. 2019Q8BERT: Quantized 8Bit BERTcited by 300
  7. 2020BinaryBERT: Pushing the Limit of BERT Quantizationcited by 178
  8. 2022A Fast Post-Training Pruning Framework for Transformerscited by 162
  9. 2020TernaryBERT: Distillation-aware Ultra-low Bit BERTcited by 148
  10. 2022BiBERT: Accurate Fully Binarized BERTcited by 73
  11. 2022BiT: Robustly Binarized Multi-distilled Transformercited by 52
  12. 2020FracBNN: Accurate and FPGA-Efficient Binary Neural Networks with Fractional Activationscited by 32
7 more
  1. 2022VAQF: Fully Automatic Software-Hardware Co-Design Framework for Low-Bit Vision Transformercited by 27
  2. 2021Hardware Acceleration of Fully Quantized BERT for Efficient Natural Language Processingcited by 22
  3. 2022A Two-Stage Efficient 3-D CNN Framework for EEG Based Emotion Recognitioncited by 7
  4. 2024BETA: Binarized Energy-Efficient Transformer Accelerator at the Edgecited by 6
  5. 2018XNORBIN: A 95 TOp/s/W Hardware Accelerator for Binary Convolutional Neural Networkscited by 5
  6. 2023Support for Stock Trend Prediction Using Transformers and Sentiment Analysiscited by 4
  7. 2024Co-Designing Binarized Transformer and Hardware Accelerator for Efficient End-to-End Edge Deploymentcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 2.

4 papers.

  1. 2025TeLLMe v2: An Efficient End-to-End Ternary LLM Prefill and Decode Accelerator with Table-Lookup Matmul on Edge FPGAscited by 2
  2. 2025PD-Swap: Prefill-Decode Logic Swapping for End-to-End LLM Inference on Edge FPGAs via Dynamic Partial Reconfigurationcited by 1
  3. 2026FASQ: Flexible Accelerated Subspace Quantization for Calibration-Free LLM Compressioncited by 1
  4. 2026Recent Developments in Transformer Inference Deployment on FPGA Platforms: A Surveycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025EdgeFlex-Transformer: Transformer Inference for Edge Devices
  2. 2021A Competitive Edge: Can FPGAs Beat GPUs at DCNN Inference Acceleration in Resource-Limited Edge Computing Applications?
  3. 2025IntAttention: A Fully Integer Attention Pipeline for Efficient Edge Inference
  4. 2026Lightweight Transformer Architectures for Edge Devices in Real-Time Applications
  5. 2025Atleus: Accelerating Transformers on the Edge Enabled by 3D Heterogeneous Manycore Architectures
  6. 2026Transformer Accelerator (TFA): A Macro-Op INT8 Hardware Chip for Transformer Inference and Machine Translation
  7. 2025TeLLMe: An Energy-Efficient Ternary LLM Accelerator for Prefilling and Decoding on Edge FPGAs
  8. 2025CoFormer: Collaborating with Heterogeneous Edge Devices for Scalable Transformer Inference