Edge AI Demo – Quantization: FP32 vs INT8

FP32 — Full Precision

Click "Start Demo" to begin

FP32 — 32-bit floats · ~27 MB · baseline
INT8 — 8-bit integers · ~7 MB · quantized

Latency —ms

Throughput —fps

Objects found —

Model size —MB

FP32 latency—

INT8 latency—

        Speedup: —
      

Edge AI Quantization Demo