Uformer

Uformer copied to clipboard

您好，我今天查看Uformer模型的参数量，发现只有400多万个参数，占用17MB的memory，这比我预想的要低很多，请问就是这么低吗？

BIOVPEPPER

loss问题

2

1. calloss_magmse全带幅度谱loss最后是除以batch和频点维度，而分段幅度谱loss(calloss_magmse_subband)除的是batch和帧数，因为output_mag.shape[2]应该是T维度； 2. 另外请问为什么不在T维度求平均呢？一般使用F.l1_loss的reduction直接用Mean就会在batch和F和T求平均，这样是对效果有啥影响吗？

shenbuguanni

关于复现网络出现的问题

5

作者您好，昨天的问题不知道怎么就不见了。我邮箱收到了您的回复，训练loss是正常的。我现在依然在查找问题的根源所在。如果您能把test 模型的源码公开，这对我解决问题会有很大的帮助。有一点需要咨询一下您，计算loss时为什么是返回三个值？期望您的回复。 昨天的问题是：测试模型时输出为静音。

Cagemoney

Training

1

Is there any script to run and use for training? Can you update the training script？

yygg678

your work's torch version

1

Thanks your creative work. But I encountered a problem in the process of reproduction, and I suspect it was a version problem.The questions are as follows： File "D:\DLSPEECHENHANCEMENT\Uformer\Uformer-main\uformer\conv2d_cplx.py", line 37,...

Cagemoney

Hi, I am confused whether the Uformer works in real time. Best Regards!

zhangxingtao

Input error when running uformer.py

10

The paper is one of the best out there, congrats! I am trying to run `uformer.py` but I get the following error: `RuntimeError: Given normalized_shape=[12], expected input with shape [*,...

jvel07

Question about torch version

1

Hi felix, Thanks for sharing this project, great work! I encountered some issues when running your code directly by: python uformer.py I have tried torch version from 1.8.1 to 1.10....

yjw123456