16337 d7f56683c7 perf: GPU NMS + 日志优化 + 数组预分配 ...

- GPU NMS: torchvision.ops.nms 替代 CPU NMS, 50-80% 提升
- 日志优化: 每10帧输出一次性能日志, 减少90%日志开销
- 数组预分配: 预分配8400框缓冲区, 避免重复创建
- 预过滤: 置信度>0.3的框先过滤, 减少NMS计算量

性能对比:
- 优化前: 40-50ms
- 优化后: 17-22ms (60% 提升)

2026-02-02 16:37:24 +08:00

..

__init__.cpython-310.pyc

feat(project): move edge_inference_service contents to root and update paths

2026-01-29 18:43:19 +08:00

config_sync.cpython-310.pyc

fix: YOLO TensorRT 输出解析修复

2026-02-02 15:02:58 +08:00

postprocessor.cpython-310.pyc

perf: GPU NMS + 日志优化 + 数组预分配

2026-02-02 16:37:24 +08:00

preprocessor.cpython-310.pyc

perf: batch=1 优化减少延迟

2026-02-02 15:25:13 +08:00

result_reporter.cpython-310.pyc

fix: YOLO TensorRT 输出解析修复

2026-02-02 15:02:58 +08:00

tensorrt_engine.cpython-310.pyc

perf: batch=1 优化减少延迟

2026-02-02 15:25:13 +08:00

video_stream.cpython-310.pyc

feat(project): move edge_inference_service contents to root and update paths

2026-01-29 18:43:19 +08:00