A drunk asr project is building based on opensmile tools and SVM classification.
├── drunk_main_sicheng.sh - here's a main bash file that is responsible for the whole pipeline
│
│
├── data - this folder contains original non-processed sicheng speech data
│
│
├── local - this folder contains the source code for generating features
│ └── prepare_data.py - the main code for preparing data
│
│
├── model - this folder contains the models, such as SVMs.
│
│
|
├── tools - this folder contains the some speech processing tools
│ └── opensmile-2.3.0 - OPENsmile tools
| └── vadtools - VAD tools based on webrtcvad
│
│
└── workspace - this folder contains all files during model training and test
└── exp - this folder contains all model files
└── feature - this folder contains various feature types from data preparation
└── package_feature - this folder contains all package features from feature folder