testadmin
/
Parakeet
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update README

This commit is contained in:
liuyibing01 2020-03-10 04:55:12 +00:00
parent 8c6a17b6bd
commit db6a96543f
2 changed files with 15 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
README.md
View File

@ -9,7 +9,7 @@ Parakeet aims to provide a flexible, efficient and state-of-the-art text-to-spee
In particular, it features the latest [WaveFlow] (https://arxiv.org/abs/1912.01219) model proposed by Baidu Research. In particular, it features the latest [WaveFlow] (https://arxiv.org/abs/1912.01219) model proposed by Baidu Research.
- WaveFlow can synthesize 22.05 kHz high-fidelity speech around 40x faster than real-time on a Nvidia V100 GPU without engineered inference kernels, which is faster than [WaveGlow] (https://github.com/NVIDIA/waveglow) and serveral orders of magnitude faster than WaveNet. - WaveFlow can synthesize 22.05 kHz high-fidelity speech around 40x faster than real-time on a Nvidia V100 GPU without engineered inference kernels, which is faster than [WaveGlow] (https://github.com/NVIDIA/waveglow) and serveral orders of magnitude faster than WaveNet.
- WaveFlow is a small-footprint flow-based model for raw audio. It has only 5.9M parameters, which is 15x smalller than WaveGlow (87.9M) and comparable to WaveNet (4.6M). - WaveFlow is a small-footprint flow-based model for raw audio. It has only 5.9M parameters, which is 15x smalller than WaveGlow (87.9M).
- WaveFlow is directly trained with maximum likelihood without probability density distillation and auxiliary losses as used in Parallel WaveNet and ClariNet, which simplifies the training pipeline and reduces the cost of development. - WaveFlow is directly trained with maximum likelihood without probability density distillation and auxiliary losses as used in Parallel WaveNet and ClariNet, which simplifies the training pipeline and reduces the cost of development.
## Overview ## Overview
@ -100,26 +100,26 @@ Parakeet also releases some well-trained parameters for the example models, whic
</th> </th>
<th> <th>
<audio id="audio" controls="" preload="none"> <audio id="audio" controls="" preload="none">
<source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_0.wav"> <source src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_0.wav">
</audio> <br> </audio> <br>
<audio id="audio" controls="" preload="none"> <audio id="audio" controls="" preload="none">
<source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_1.wav"> <source src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_1.wav">
</audio><br> </audio><br>
<audio id="audio" controls="" preload="none"> <audio id="audio" controls="" preload="none">
<source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_2.wav"> <source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_2.wav">
</audio><br> </audio><br>
<audio id="audio" controls="" preload="none"> <audio id="audio" controls="" preload="none">
<source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_3.wav"> <source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_3.wav">
</audio><br> </audio><br>
<audio id="audio" controls="" preload="none"> <audio id="audio" controls="" preload="none">
<source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_4.wav"> <source id="mp3" src="https://paddlespeech.bj.bcebos.com/Parakeet/clarinet_ljspeech_samples_1.0/step_500000_sentence_4.wav">
</audio> </audio>
</th> </th>
</tr> </tr>
</tbody> </tbody>
</table> </table>
**Note:** The input samples are drawn from validation dataset that are not visible in training. **Note:** The input samples are drawn from validation dataset that are not visible in training.
- TTS models - TTS models

2
examples/waveflow/README.md
View File

@ -3,7 +3,7 @@
PaddlePaddle dynamic graph implementation of [WaveFlow: A Compact Flow-based Model for Raw Audio](https://arxiv.org/abs/1912.01219). PaddlePaddle dynamic graph implementation of [WaveFlow: A Compact Flow-based Model for Raw Audio](https://arxiv.org/abs/1912.01219).
- WaveFlow can synthesize 22.05 kHz high-fidelity speech around 40x faster than real-time on a Nvidia V100 GPU without engineered inference kernels, which is faster than [WaveGlow] (https://github.com/NVIDIA/waveglow) and serveral orders of magnitude faster than WaveNet. - WaveFlow can synthesize 22.05 kHz high-fidelity speech around 40x faster than real-time on a Nvidia V100 GPU without engineered inference kernels, which is faster than [WaveGlow] (https://github.com/NVIDIA/waveglow) and serveral orders of magnitude faster than WaveNet.
- WaveFlow is a small-footprint flow-based model for raw audio. It has only 5.9M parameters, which is 15x smalller than WaveGlow (87.9M) and comparable to WaveNet (4.6M). - WaveFlow is a small-footprint flow-based model for raw audio. It has only 5.9M parameters, which is 15x smalller than WaveGlow (87.9M).
- WaveFlow is directly trained with maximum likelihood without probability density distillation and auxiliary losses as used in Parallel WaveNet and ClariNet, which simplifies the training pipeline and reduces the cost of development. - WaveFlow is directly trained with maximum likelihood without probability density distillation and auxiliary losses as used in Parallel WaveNet and ClariNet, which simplifies the training pipeline and reduces the cost of development.
## Project Structure ## Project Structure