SMIL-SPCRAS / EMOLIPS

EMOLIPS: TWO-LEVEL APPROACH FOR LIP-READING EMOTIONAL SPEECH
GNU General Public License v3.0
3 stars 0 forks source link

Data Sets for CREMA-D #1

Closed david-gimeno closed 11 months ago

david-gimeno commented 12 months ago

Hi,

My name is David and, first of all, congratulations for you good work. It is really interesting.

I am also working on lipreading and i would like to do a similar case study for my PhD thesis. However, although based on your code, I am not able to obtain the same tranining, validation, and test sets you are describing in your paper. For this reason, I wanted to ask you if it would be possible to obtain three CSVs indicating just the video IDs (e.g. "1076_MTI_SAD_XX") for each subset. With theses IDs is enough, since the rest of information i can retrieve it using my own scripts. It would so helpful!

Thanks in advance. Best regards from Spain,

David

david-gimeno commented 11 months ago

I solved it! I do not know why but if I run your training notebook, I am getting the correct dataset splits.

David.