Multi-talker Automatic Speech Recognition on Libri2Mix Noisy (dev)
8.4WERID-20.D
Evaluation Results
| Method | Links | |
|---|---|---|
| ID-20.DLLM usage=With LLMs2026.03 | 8.4 | |
| ID-30.DLLM usage=With LLMs2026.03 | 8.6 | |
| SOP-Llama-3BLLM Usage=With LLMs2026.03 | 10.5 | |
| SOP-Llama-3BLLM usage=With LLMs, Backbone=Llama-3B, Protocol=SOP2026.03 | 10.5 | |
| ID-8LLM Usage=With LLMs, Method description=Proposed Encoder-only Model with distillation2026.03 | 10.7 | |
| SOT-Llama-3BLLM Usage=With LLMs2026.03 | 11.2 | |
| SOT-Llama-3BLLM usage=With LLMs, Backbone=Llama-3B, Protocol=SOT2026.03 | 11.2 | |
| SOP-Llama-1BLLM Usage=With LLMs2026.03 | 11.8 | |
| SOP-Llama-1BLLM usage=With LLMs, Backbone=Llama-1B, Protocol=SOP2026.03 | 11.8 | |
| SOT-Llama-1BLLM Usage=With LLMs2026.03 | 12.4 | |
| SOT-Llama-1BLLM usage=With LLMs, Backbone=Llama-1B, Protocol=SOT2026.03 | 12.4 | |
| w/o decoder: CTCLLM Usage=Without LLMs, Encoder=SSL2026.03 | 16.8 | |
| GEncSep w/o decoderLLM usage=Without LLMs, SSL for speech encoder=true, Alignment protocol=serialized CTC2026.03 | 16.8 | |
| GEncSepLLM Usage=Without LLMs, Encoder=SSL2026.03 | 17.2 | |
| GEncSepLLM usage=Without LLMs, SSL for speech encoder=true2026.03 | 17.2 | |
| Training from ScratchLLM Usage=Without LLMs, Encoder=SSL2026.03 | 19.4 | |
| Training from ScratchLLM usage=Without LLMs, SSL for speech encoder=true2026.03 | 19.4 |