search

PaddlePaddle / PaddleSeg

Easy-to-use image segmentation library with awesome pre-trained model zoo, supporting wide-range of practical tasks in Semantic Segmentation, Interactive Segmentation, Panoptic Segmentation, Image Matting, 3D Segmentation, etc.
https://arxiv.org/abs/2101.06175
Apache License 2.0
8.71k stars 1.68k forks source link

PaddleSeg跑cityscapes数据集的问题 #764

Closed yy11223 closed 1 year ago

yy11223 commented 3 years ago

模型评估

!cd work/PaddleSeg && python pdseg/eval.py --use_gpu \ --cfg configs/deeplabv3p_mobilenetv3_large_cityscapes.yaml \ TEST.TEST_MODEL saved_model/deeplabv3p_mobilenetv3_large_cityscapes/best_model 运行结果： [EVAL]#image=500 acc=0.9529 IoU=0.7251 [EVAL]Category IoU: [0.9779 0.8250 0.9087 0.4881 0.5561 0.5690 0.6265 0.7335 0.9160 0.6307 0.9382 0.7701 0.5423 0.9379 0.6543 0.7905 0.6180 0.5689 0.7254] [EVAL]Category Acc: [0.9895 0.8980 0.9444 0.7817 0.7766 0.7893 0.8002 0.8743 0.9502 0.8188 0.9594 0.8514 0.7229 0.9655 0.8730 0.8536 0.8881 0.7644 0.8347] 问题：EVAL输出的的 acc和IoU是怎么计算的，是Category IoU和Category Acc的均值吗？我计算了IoU是均值，acc不是

wuyefeilin commented 3 years ago

acc ：预测正确的像素数/总像素数

yy11223 commented 3 years ago

那[EVAL]#image=500 acc=0.9529 IoU=0.7251里的 IoU也不是平均指标是吧？

wuyefeilin commented 3 years ago

那[EVAL]#image=500 acc=0.9529 IoU=0.7251里的 IoU也不是平均指标是吧？

mIoU是取的各类的iou平均值

yy11223 commented 3 years ago

程序输出是IoU，意思其实是mIoU是吧？

wuyefeilin commented 3 years ago

程序输出是IoU，意思其实是mIoU是吧？

是的， 我们在新版本动态图已经进行了相应的规范，欢迎使用我们的动态图版本

github-actions[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.