Skip to content

Latest commit

 

History

History
 
 

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 

README.md

DSTC7 Challange

Prepare Environment

conda install mkl-service

Prepare data

All the data can be found in the ofiicial repository here.

You need to first make symbolic link under data folder like this.

data/
├── word_embedding 
├── textdata
├── cache
└── Audio-Visual-Feature 

word_embedding/
└── glove.6B.300d.txt

Training

    python -m src.train --modelType [SimpleModel|....] 

For more information, use --help flag. A simple example can be found in train_simple.sh

Testing

    python -m src.predict [parameter_timestep] 

For more information, use --help flag.

Code Structure

.
├── data (all data should be placed here)
│   ├── Audio-Visual-Feature 
│   ├── cache
│   ├── textdata 
│   └── word_embedding 
├── output (all model output would be place here)
│   ├── generate (after predicting, model prediced result will be place here as a JSON file)
│   ├── log (after training, training losses and validation losses will be placed here)
│   ├── metrics (after predicting, automatic evaluation metrics would be placed here)
│   ├── parameter (after training, model parameters would be placed here)
│   └── visualization 
└── src
    ├── components (Layers that can be reuse)
    ├── dataset (handle the dataset)
    ├── log (log configuration file)
    ├── model (all models are here)
    │   ├── components
    │   └── net
    ├── statistics (plot things or compute statistics information)
    └── util