Closed MaFuyan closed 3 years ago
Maybe @WasifurRahman @matalvepu could answer this better since I was mostly involved with modeling code and experiments.
At least for the last MOSEI we posted, the pre-processing applied are written in pull request: #10.
I really appreciate it that you reply so quickly!!! I will follow the process on MOSEI and have a try.
Closing due to inactivity, feel free to re-open if any problem persists
Hi. Thank you for your brilliant work.
Would you mind giving some instructions on how to get such formatted pickle files like your processed MOSI and MOSEI? Previous processed IEMOCAP pickle files (http://immortal.multicomp.cs.cmu.edu/raw_datasets/processed_data/iemocap/) don't contain keys "ID" or "SEGMENT" to identify which utterance the raw transcripts belong to.
Thanks in advance.