Day30 - Baseline在FLIR上的表现

以yolov5n/s/m和yolov8n/s作为baseline,在FLIR上训练200epochs

• Yolov5n

  Parameters: 2.655M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.604	0.445	0.487	0.293
  person	0.832	0.605	0.74	0.415
  bike	0.517	0.435	0.411	0.247
  car	0.855	0.742	0.835	0.588
  motor	0.757	0.566	0.634	0.339
  bus	0.771	0.475	0.611	0.425
  truck	0.241	0.326	0.152	0.102
  light	0.764	0.373	0.484	0.22
  other vehicle	0.0947	0.0416	0.031	0.01

• Yolov5s

  Parameters: 9.153M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.637	0.495	0.553	0.34
  person	0.866	0.65	0.788	0.467
  bike	0.605	0.518	0.497	0.312
  car	0.88	0.777	0.867	0.628
  motor	0.793	0.618	0.755	0.379
  bus	0.772	0.453	0.639	0.461
  truck	0.307	0.435	0.27	0.183
  light	0.797	0.459	0.573	0.272
  other vehicle	0.0798	0.0476	0.0369	0.0169

• Yolov5m

  Parameters: 25.111M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.676	0.54	0.579	0.371
  person	0.871	0.677	0.808	0.498
  bike	0.597	0.565	0.55	0.359
  car	0.887	0.798	0.881	0.651
  motor	0.771	0.655	0.731	0.418
  bus	0.812	0.553	0.691	0.509
  truck	0.407	0.457	0.283	0.191
  light	0.804	0.503	0.607	0.303
  other vehicle	0.261	0.111	0.0816	0.0423

• Yolov8n

  Parameters: 3.157M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.6	0.433	0.48	0.288
  person	0.835	0.614	0.743	0.418
  bike	0.498	0.412	0.404	0.246
  car	0.859	0.747	0.838	0.595
  motor	0.796	0.569	0.61	0.314
  bus	0.78	0.455	0.605	0.413
  truck	0.195	0.239	0.134	0.0827
  light	0.723	0.381	0.473	0.221
  other vehicle	0.111	0.0476	0.0367	0.0171

• Yolov8s

  Parameters: 11.167M

  Class	Precision	Recall	mAP50	mAP50-95
  all	0.626	0.522	0.55	0.347
  person	0.834	0.681	0.79	0.474
  bike	0.53	0.559	0.51	0.313
  car	0.841	0.808	0.87	0.633
  motor	0.82	0.618	0.70	0.399
  bus	0.768	0.535	0.66	0.477
  truck	0.278	0.413	0.23	0.162
  light	0.765	0.468	0.57	0.272
  other vehicle	0.173	0.0952	0.087	0.0469