nsynth_wavenet
nsynth_wavenet copied to clipboard
bfs18
parallel wavenet based on nsynth
Hi, Great work. Do you have samples that generated by model with only KL divergence as loss function?
I've been trying to reproduce the parallel wavenet results, however I'm running into some issues with training the teacher model. I have trained it on the LJ Speech dataset with...
I am talking about systhesising not training. This is what I did, I downloaded pretrained model : https://drive.google.com/file/d/17bt6laBcTcjZCsLUzdTjtaH2hQQyki8E/view created a directory `PWN_LOGDIR` and renmed checkpoints to fix [this](https://github.com/bfs18/nsynth_wavenet/issues/31) and then...
Does it represent the minimum length of each speech? Why do you want to send each segment of speech into training? Is it the reason for memory resources?If my audio...
Hi, @bfs18 I have trained a teacher model using default configure of MOL loss where I have set up 'num_layers=10' because of memory limit. The performance of 114k step model...
Hello. When you train for 400 000 or 200 000 iteration. Which batch size are you using? is it 28? On 1 GPU (nvidia Geforce 1080 with 11178MiB of memory)...
Hi when reloading ns_wn-gauss model for teacher continue training, I got ``` INFO:tensorflow:Restoring parameters from ns_wn-gauss/model.ckpt-200000 2018-08-09 15:04:30.388764: W tensorflow/core/framework/op_kernel.cc:1202] OP_REQUIRES failed at save_restore_v2_ops.cc:184 : Not found: Key beta1_power not...
Hi, There is a very strange error when running `test_reader.py`. The original `test_reader.py` runs ok. However, once I remove line https://github.com/bfs18/nsynth_wavenet/blob/master/tests/test_reader.py#L26, I got a `Segmentation fault` error, for example: 2018-09-20...
Is it possible to integrate this wavenet implementation with r9y9's Tacotron 2 as r9y9's wavenet_vocoder.
