Skip to content

Speech synthesis platform based on tensorflow and sonnet

License

Notifications You must be signed in to change notification settings

npuichigo/voicenet

Repository files navigation

Voicenet: Speech Synthesis Platform

Overview

Voicenet is an open source speech synthesis framework based on tensorflow and sonnet. This flexible architecture lets you validate your new neural network based acoustic model quickly in your experiments, and owing to the deployment capability of tensorflow, we think it's easy to deploy new algorithms and experiments online for serving.

Installation Instructions

We have simplify the dependencies of voicenet, so you will need to install tensorflow>=v1.7.0 and progress only.

pip install -r requirements.txt

Getting Started

  1. Clone this repository with git clone https://github.com/npuichigo/voicenet.git
  2. Go to voicenet/tools and run the script compile_tools.sh to compile third_party tools
  3. Go to voicenet/egs/slt_arctic/ and run the script run.sh
  4. For your own dataset, just make a new directory under voicenet/egs/, copy voicenet/egs/local and voicenet/egs/run.sh to the new workspace

WARNING:You should change the training parameters for your own dataset. For the purpose of demostration, batch_size is set to one in voicenet/egs/slt_arctic/.

IMPORTANT:We remove the dependency of sonnet in the latest version of voicenet. The main reason is that we want to keep track of tensorflow's rapid updates. In addition, starting in tensorflow 1.2, dataset iterator is added for reading data into tensorflow. In using tensorflow's dataset api, iterators of dataset_train and dataset_valid can be merged into one iterator, which can be switched between different datasets conveniently, so variable reuse is no longer needed.

model = LSTM(...)
iterator = tf.contrib.data.Iterator.from_structure(
    dataset_train.batched_dataset.output_types,
    dataset_train.batched_dataset.output_shapes)
input, _ = iterator.get_next()
output = model(input)