Forward unnecessarily bloated?

[mueller91]

I noticed that the following code in the fordward method does not depend on input x. Is there a reason this is not run once, and only once, during model init? Seems like unneccesary overhead to run it during each fwd pass, or am i missing something?

for i in range(len(self.mel)-1):
    fmin=self.mel[i]
    fmax=self.mel[i+1]
    hHigh=(2*fmax/self.sample_rate)*np.sinc(2*fmax*self.hsupp/self.sample_rate)
    hLow=(2*fmin/self.sample_rate)*np.sinc(2*fmin*self.hsupp/self.sample_rate)
    hideal=hHigh-hLow
    
    self.band_pass[i,:]=Tensor(np.hamming(self.kernel_size))*Tensor(hideal)

band_pass_filter=self.band_pass.to(self.device)

self.filters = (band_pass_filter).view(self.out_channels, 1, self.kernel_size)