“Mengyuan Zhao 2016-05-09”版本间的差异
来自cslt Wiki
（以“* sino-8k model: # The training process of silence-noise-separated xEnt model is about 40%. * sino-16k embedded model: # Start mpe training. * feature: # modifying w...”为内容创建页面） |
（没有差异）
|
2016年5月9日 (一) 00:08的最后版本
- sino-8k model:
- The training process of silence-noise-separated xEnt model is about 40%.
- sino-16k embedded model:
- Start mpe training.
- feature:
- modifying wav-cnn recipe.
- testing binary feature(timit), still unfinished.
