Speech Emotion Recognition on MSP-Podcast (WER/CCC)
0WERground truth
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ground truthtranscription_source=ground truth text2026.01 | 0 | 0.613 | |
| ASR Error-Robust Modality-Gated Fusiontranscription_source=ASR with error correction2026.01 | 12.85 | 0.618 | |
| best transcriptiontranscription_source=best automated transcription2026.01 | 15.02 | 0.616 |