Long

I also want to know this question, do you know it now?

You are right. I use a encoder-decoder architecture segmentation model, like UNet. Should I how to obtain a cam? Looking forward to your reply.

I have the same problem as https://github.com/frgfm/torch-cam/discussions/150. The aim of our work to obtain a CAM of semantic segmentation model, like UNet. However, I fails to get it.

我也是这样的，请问你解决了吗？

Has the effect improved after you changed to a different optimizer?

你好，你有做过实验嘛？我用这种方法得到的结果不是很理想（目标是三任务，三输出）

我想问batch_size取其他值是否可以呢？比如4，8，16等。还有为什么要权重初始化呢（weight_init函数）？

谢谢你的回答，我尝试设置batch_size为8，但是得到的loss非常大，几万，而且结果是全黑的，这是为什么呢

博主看到啦，但是具体的代码实施是在哪个文件呢？

谢谢你的答复。我想问一下你有遇到过Clutter类别的F1指标为nan的情况吗，其他五类却都是正常的。