Automatic Speech Recognition on LRS2
3.47WER (%)Qwen2.5-Omni
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-OmniPreprocessing=Raw video2026.03 | 3.47 | |
| HumanOmni-SpeakerPreprocessing=Raw video2026.03 | 3.47 | |
| OLAPreprocessing=Raw video2026.03 | 5.5 |
Loading the SOTA2 catalog…