search

mravanelli / SincNet

SincNet is a neural architecture for efficiently processing raw audio samples.
MIT License
1.11k stars 260 forks source link

Is this a bug? #90

Open JunzheJosephZhu opened 4 years ago

JunzheJosephZhu commented 4 years ago

self.bn.append(nn.BatchNorm1d(N_filt,int((current_input-self.cnn_len_filt[i]+1)/self.cnn_max_pool_len[i]),momentum=0.05))

I pulled out the module and it looks like: BatchNorm1d(80, eps=2583, momentum=0.05, affine=True, track_running_stats=True) when I set batchnorm = [True, True, True] and layernorm = [False, False, False] in config.

From the SincNet class. Aren't the first two arguments supposed to be in a list?

ZaUt-bio commented 11 months ago

I have the same issue. could you find a solution?