Automatic Speech Recognition on Fearless Steps Challenge (FSC) Phase-2 (dev)
22.7WERWavLM
Evaluation results
| Method | Links | |
|---|---|---|
| WavLMArchitecture Components=Enc./Dec.2026.04 | 22.7 | |
| Chen et al.2026.04 | 26.2 | |
| Gorin et al.2026.04 | 28.6 | |
| Whisper baseEvaluation Protocol=fine-tuned, Model Variant=base, Number of Parameters=74 million2026.04 | 29.3 | |
| HuBERTArchitecture Components=Enc./Dec.2026.04 | 31.8 | |
| Whisper mediumEvaluation Protocol=zero-shot, Model Variant=medium, Number of Parameters=769 million2026.04 | 53.1 |