TODO

Before ESSV

create tiny dataset in benchmark_data.py
score for tongue tip and tongue middle
implement scores in score.py
/babi/ /babu/ /baba/ datasets
score for tongue height
segment model
create website / polish readme with table of model comparisons and link to ESSV paper
add page with score description

After ESSV

create datasets small and normal in benchmark_data.py
merge paule_benchmark to this one

Later

new task speaking rate; onset offset comparison
KEC: /eigentlich/ vgl. publication by Tomaschek