KaihuaTang / Scene-Graph-Benchmark.pytorch

A new codebase for popular Scene Graph Generation methods (2020). Visualization & Scene Graph Extraction on custom images/datasets are provided. It's also a PyTorch implementation of paper “Unbiased Scene Graph Generation from Biased Training CVPR 2020”
MIT License
1.06k stars 229 forks source link

Large difference in the accuracy of detection model #191

Closed qqqqhhqq closed 1 year ago

qqqqhhqq commented 1 year ago

❓ Questions and Help

According to the author's guidance, only the detection mAP of the training model is 0.25, but I use the weight of the detection model given by the author to train the precls detection model with the mAP of 0.99. Is this normal my training mode image

When I was the author's weight training precls image