Open guangpintao opened 10 months ago
{"@timestamp":"2023-10-20 19:09:39.697","@message":" Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.329"} {"@timestamp":"2023-10-20 19:09:39.697","@message":" Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.513"} {"@timestamp":"2023-10-20 19:09:39.698","@message":" Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.346"} {"@timestamp":"2023-10-20 19:09:39.698","@message":" Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.196"} {"@timestamp":"2023-10-20 19:09:39.699","@message":" Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.364"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.449"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.275"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.494"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.557"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.364"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.593"} {"@timestamp":"2023-10-20 19:09:39.700","@message":" Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.735"} {"@timestamp":"2023-10-20 19:09:39.701","@message":"bbox AP seen: 59.74352140181269"} {"@timestamp":"2023-10-20 19:09:39.701","@message":"bbox AP unseen: 27.574026608034263"} {"@timestamp":"2023-10-20 19:09:42.705","@message":"Training time 3 days, 7:24:13"}
Using the code in repo and using the provided pre-trained model to test consistently with performance in paper, but when retraining, the metrics are only APall=51.3 APsee=59.7 APunseen=27.6, and the setting of parameter R is also different from that in the paper (3 in the paper and 12 in the code). What should be noted when reproducing the performance of the paper?