-
Hello Evangelos,
Thanks for sharing your code!
I get a question here:
In dataset.py, is there any different between your implementation of ‘_log_spectgram’ and the ‘librosa.feature.melspectrogra…
-
1. What should I do to make my custom layers available to be loaded using `load_model()`?
I used my custom layers in [this repo](https://github.com/keunwoochoi/kapre/blob/master/kapre/TimeFrequency…
-
When i training tacotron2 model with my own data, some of mel-spetrogram has timestep > 2000 so I always get this error:
> tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatib…
-
At this point, I feel like we definitely update the current STFT implementation with more native STFT operations for speed/stability. Out of three Keras backends -- tf, theano, cntk -- only tf has its…
-
hi，I have doubt about the preprocessing_mel function. I use the following preprocessing method. The generated audio file is muted.
def melspectrogram(wav, hparams):
D = _stft(preemphasis(wav, hpa…
-
Why Augmentation file has been removed?
Why is not applied power 2.0 to Magnitude Layer before of filter bank Mel?
-
Hi,
May I know how you get the MFCC features? Like the parameters you used. Thank you very much!
-
In looking at the code, it seems like it might be possible to skip the synthesizing step. (I'm using the standard datasets but I would like to use an existing wav file and NOT one that was synthesize…
-
**Describe your question**
I would like to write and train a model in Nemo which, given a ground truth labelling of where I think speech or silence is, returns a list of start and end indices in th…
-
Hi @keunwoochoi !
I'm enjoying your Kapre library, as I'm currently using it in my music-based deep learning project...
In my deep network, I'm using a similar architecture as the ones you have …