Automatic Speech Recognition on LibriSpeech ASR (test)
2WERQwen2-Audio-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen2-Audio-7BArchitecture=Qwen2, Parameters=7B2025.02 | 2 | 0.7 | — | — | |
| whisper-large-v3Model Size=Large, Version=v32025.02 | 2.2 | 0.3 | — | — | |
| hubert-large-ls960-ftModel Size=Large, Finetuning=ls9602025.02 | 2.2 | 0.1 | — | — | |
| whisper-large-v3-turboModel Size=Large, Version=v3-turbo2025.02 | 3.3 | 0.1 | — | — | |
| wav2vec2-large-xlsr-53-englishModel Size=Large, Architecture=wav2vec2-xlsr2025.02 | 6.5 | 0 | — | — | |
| Ground TruthRIR Source=Ground Truth2026.03 | — | — | 8 | 0 | |
| Proposed text-conditioned RIR generationRIR Source=Ours2026.03 | — | — | 12 | 0 |