Multi-talker Automatic Speech Recognition on Libri2Mix Clean (eval)
3.6WERSOP-Llama-3B
Evaluation results
| Method | Links | |
|---|---|---|
| SOP-Llama-3BLLM Usage=With LLMs2026.03 | 3.6 | |
| SOP-Llama-1BLLM Usage=With LLMs2026.03 | 4 | |
| SOT-Llama-3BLLM Usage=With LLMs2026.03 | 4.1 | |
| ID-8LLM Usage=With LLMs2026.03 | 4.1 | |
| SOT-Llama-1BLLM Usage=With LLMs2026.03 | 4.6 | |
| w/o decoder: CTCLLM Usage=Without LLMs, Encoder=SSL2026.03 | 6.1 | |
| GEncSepLLM Usage=Without LLMs, Encoder=SSL2026.03 | 6.6 | |
| Training from ScratchLLM Usage=Without LLMs, Encoder=SSL2026.03 | 7 | |
| WavLM-CLNLLM Usage=Without LLMs, Encoder=SSL2026.03 | 7.6 | |
| C-HuBERT LARGELLM Usage=Without LLMs, Encoder=SSL2026.03 | 7.8 | |
| W2V-Sidecar-ft.LLM Usage=Without LLMs, Encoder=SSL2026.03 | 8.1 |