seorim0
commented
2 years ago Hi! Thank you for your interest in our project.
Yes, similar issues can be found here. (#8)
The minMaxNorm function was designed to set the input value between -1 and 1. You may abandon abs() depending on your purpose.
And actually, in this project, since inputs = torch.clamp_(inputs, -1, 1) is used in dataloader.py, you do not need to use the function.
I hope this answer was helpful to you!
lixinlun
Hello,同学你好！
In the last few days, I was trying to change the dataloader from loading .npy huge file to load wav files from folders directly on your DCCRN project. (https://github.com/seorim0/DCCRN-with-various-loss-functions)
Fortunately, I saw a same requset from the Issues (https://github.com/seorim0/DCCRN-with-various-loss-functions/issues/4), which helped me a lot. Now, I find a function in NUNet-TLS --> tools.py --> minMaxNorm(wav, eps=1e-8), However, is the function result value correct?
tools.py line72-76:
def minMaxNorm(wav, eps=1e-8):max = np.max(abs(wav))min = np.min(abs(wav))wav = (wav - min) / (max - min + eps)return wavthe wav data should have negtive values, so... should we change the code to like this?
def minMaxNorm(wav, eps=1e-8):max = np.max(wav)min = np.min(wav)wav = (wav - min) / (max - min + eps)return wavI mean, should we abandon the abs()? thank you!!!