Open bagustris opened 9 months ago
Current SER evaluation only split into five cross-validation using KFold. It is necessary to split based on session, use each session for test, and report average score (UA, WA) for benchmarking.
Solved in commit cd069a5.
Current SER evaluation only split into five cross-validation using KFold. It is necessary to split based on session, use each session for test, and report average score (UA, WA) for benchmarking.