We released 3,000 samples of the annotated spoken sentences and our annotation rules.
This is a tool for disfluency detection in the pipeline speech translation system.
pip install -r requirements.txt
python train.py -c configs/test.yaml
Any questions, please contact me with the email: dongqianqian2016@ia.ac.cn