# Run anti-spoofing for codec

- https://github.com/hbwu-ntu/CodecResynBenchmark/tree/main
- https://docs.google.com/spreadsheets/d/1Una1I_R3_c5LyAzCuPv0avBM9eVvONkmzrw5z13XyWs/edit?usp=sharing

## Get data

/media/hbwu/12TB/PublicData/CodecData

## Training
- set config/AASIST-L.conf

"database_path": "YourPathToCodecData/CodecData/datalist.csv"

- set run.sh

stage=0
traincodecname="SpeechTokenizer"

## Testing
- set run.sh

stage=0
model_path="Your/Model/Path"