List your company

ResearchFormerly WizWand

Transformer Inference

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

12-layer Transformer 1024 tokens (inference)

342.74Speedup

24

May 28, 2026

Spartan-7 XC7S15 FPGA

Resource-aware Mixed-precision Quantization

3.77RMSE

3

Apr 23, 2026