COBRA co-designs a true 1-bit matrix multiplication engine and a shifted polarized softmax attention substitute, achieving 3,894 GOPS and 448.7 GOPS/W on an edge FPGA with a small GLUE accuracy drop.
BERT: Pre-training of deep bidirectional transformers for language understanding,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
COBRA: Algorithm-Architecture Co-optimized Binary Transformer Accelerator for Edge Inference
COBRA co-designs a true 1-bit matrix multiplication engine and a shifted polarized softmax attention substitute, achieving 3,894 GOPS and 448.7 GOPS/W on an edge FPGA with a small GLUE accuracy drop.