Image Classification on MNIST (Training and FPGA Accuracy)
98.48Training AccuracyHan et al.
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Han et al.Year=2020, Model=LIF, Neurons=784-1024-1024-10, FPGA=C7Z045 SoC, freq. [MHz]=100, Cells log.=12,690, BRAM used=40.5, T/img [ms]=12.42, E/img [mJ]=5.92, Weight (bits)=6 (4)2026.05 | 98.48 | 97.06 | |
| Carpegna et al.Year=2024, Model=LIF, Neurons=784-128-10, FPGA=XC7Z020, freq. [MHz]=100, Cells log.=7,612, DSP used=0, BRAM used=18, T/img [ms]=0.78, E/img [mJ]=0.14, Weight (bits)=42026.05 | 96.83 | 93.85 | |
| SRCYear=2025, Model=SRC, Neurons=784-100-10, FPGA=XC7A200, freq. [MHz]=100, DSP used=100, BRAM used=341, T/img [ms]=1.748, E/img [mJ]=2.24, Weight (bits)=4, Variant=200 - 202026.05 | 96.48 | 95.35 | |
| SRCYear=2025, Model=SRC, Neurons=784-100-10, FPGA=XC7A200, freq. [MHz]=100, Cells log.=93,347, DSP used=100, BRAM used=341, T/img [ms]=0.436, E/img [mJ]=0.55, Weight (bits)=4, Variant=44 - 42026.05 | 96.48 | 92.89 |