New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

训练结果离预期有差距 #10

Open

laocan opened this issue Jul 6, 2022 · 1 comment

laocan commented Jul 6, 2022

您好，用 AGRNet 在 celebahqmask 上跑了几十个epoch后，遇到些问题，请教下：

pred_edge 的效果跟 label 比起来，相差太多（参考图片）
测试图片时，一般虚化背景会比较好，如果背景干扰一多，对解析的结果影响蛮大
小目标（比如人眼等）结果有点糟糕

请问，是训练问题吗，还是模型本身的限制？有什么优化办法吗？
谢谢！

Owner

tegusi commented Jul 7, 2022

可以设置不同的阈值减小edge宽度
尝试单独先matting背景
裁剪图片局部进行分割

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment