junxnone / tio

Log

Other

10 stars 5 forks source link

OpenVINO - benchmark #678

Open junxnone opened 4 years ago

junxnone commented 4 years ago

Reference

[ ] inference-engine/tools/benchmark_tool

Brief

Calculate Latency & Throughput

Latency : 取所有 request的中值
Throughput：async mode 取 inference images 除以 总时间

        times = request_queue.times
        times.sort()
        latency_ms = median(times)
        fps = batch_size * 1000 / latency_ms
        if self.api_type == 'async':
            fps = batch_size * iteration / total_duration_sec

junxnone commented 4 years ago

17