ljuvela/resgan

Core code for my ICASSP 2018 paper

resgan's Issues

GlottDNN

Any ETA on when GlottDNN will be available? Very interested in your code but want to have the full end-to-end experience.

F0 model

The code we used for the F0 model is available at http://tonywangx.github.io/index.html#DAR
We replaced the inputs with MFCCs and only used frame-tier feedbacks. For support on this, @TonyWangX knows more.

We're planning to do more experiments on various F0 models. These will be in TensorFlow and added to this repo.

TensorFlow implementation with WGAN

Everything is so much nicer in TensorFlow. No CPU/GPU memory swapping for FFTs, for example.

Stay tuned.

It is my first time using keras.

./01_run_train.sh
Using TensorFlow backend.
Traceback (most recent call last):
File "train.py", line 327, in
context_len=args.rnn_context_len)
File "train.py", line 68, in train_pls_model
pls_model = time_glot_model(timesteps=timesteps)
File "/home/zyq/lianxi/ResGAN-master/models.py", line 100, in time_glot_model
x_t = GRU(50, activation='relu', kernel_initializer='glorot_normal', return_sequences=False, unroll=False)(x_t)
File "/home/zyq/.local/lib/python3.5/site-packages/keras/layers/recurrent.py", line 500, in call
return super(RNN, self).call(inputs, **kwargs)
File "/home/zyq/.local/lib/python3.5/site-packages/keras/engine/base_layer.py", line 460, in call
output = self.call(inputs, **kwargs)
File "/home/zyq/.local/lib/python3.5/site-packages/keras/layers/recurrent.py", line 1589, in call
initial_state=initial_state)
File "/home/zyq/.local/lib/python3.5/site-packages/keras/layers/recurrent.py", line 609, in call
input_length=timesteps)
File "/home/zyq/.local/lib/python3.5/site-packages/keras/backend/tensorflow_backend.py", line 2957, in rnn
maximum_iterations=input_length)
TypeError: while_loop() got an unexpected keyword argument 'maximum_iterations'

TFRecord data providers

Coming soon...