Goals Accomplished This week:
- The sampling frequency of all the audio was checked.
- The problem in the table which used to analyze SVM ranking was solved.
- All female data were applied to train the rank models.
- The emotions from IS18 dataset were extracted to classify the features of a part of audio.
- Data processing on the dataset of heartbeat preextracted features was done.
Goals to be Accomplished Next Week :
- Analyze the result of all female data which was training now.
Technical Difficulties:
- The training of rank model was very slow.
- The torch hdf5 used in soundnet to extract features was not match with the update mac OS.
Weekly Report