Jasonx0123 comments

Results 7 comments of


                                            Jasonx0123

Training errow

Thanks for your reply， I didn't change the version of pytorch. I add this code into train_amp.py and the problem was solved. parse.add_argument('--local_rank', defautle=-1, type=int, help='node rank for distributed training')

AssertionError: Invalid type

> Try to change '_C.LOSS.MULTI_HEAD_LAMBDA = None' to be '_C.LOSS.MULTI_HEAD_LAMBDA = False', works for me, good luck :-) thanks 发现是torch版本没对

复现之后，mAP达不到那么高

> ‍博主大神，一直关注你的微博和视频学到现在，我复现了你的SSD—pytorch，训练了VOC2012数据集，但是怎么达不到你的模型效果，什么参数都没改，结果却不理想。Epoch=150,loss=1.4+0.25左右，哪些参数需要调整吗？真诚希望得到你的指导！多谢！‍ 请问解决了嘛？亲我的也是达不到作者给的结果

在 make -j8出现了错误

![f8074dfb974b1f2e6953550f894f778](https://user-images.githubusercontent.com/46207106/139251069-4fd32116-6088-45af-ba5b-440a3fc69025.png)

您好，请问利用模型测试时，可以不用成对的阴影和mask图像，只用带有阴影的图像测试吗？

我在利用elvatulate.m验证的时候，发现SSIM指标为0，请问是哪里可能出问题呢？

ValueError: too many values to unpack

Hi,@devsergioraar ,I got the same problem. Did you solve this problem?

When will you have the code for paper "From Shadow Segmentation to Shadow Removal" open to public?

Is the author's code open source?