Does preprocess.sh have an effect if using datasets other than phoneix? I have successfully generated the *-goundtruth-[dev,test,train].stm file. Is that alone enough to train with another dataset?
As indicated in Line 14, the simplifications pertain specifically to the phoenix14/14T datasets. Please adjust the preprocessing step according to the evaluation requirements of any other dataset.
Does preprocess.sh have an effect if using datasets other than phoneix? I have successfully generated the *-goundtruth-[dev,test,train].stm file. Is that alone enough to train with another dataset?
I look forward to the response. Thank You