List your company

Automatic Speech Recognition on Chinese Dialect Scenario Aggregate

12Hallucination RateQwen3-ASR-1.7B
1117.7524.531.25Apr 9, 2026

Evaluation results

MethodLinks
Qwen3-ASR-1.7BDecoding Mode=offline2026.0412
OursDecoding Mode=offline, Beam Size=3, IA-SFT=true2026.0412.2
Step-Audio2-miniDecoding Mode=offline2026.0419.4
Ours (w/o IA-SFT)Decoding Mode=offline, Beam Size=3, IA-SFT=false2026.0419.8
GLM-ASR-nanoDecoding Mode=offline2026.0420.1
Fun-ASR-nanoDecoding Mode=offline2026.0421.7
FireRedASR-LLMDecoding Mode=offline2026.0422.8
Qwen3-Omni-InstDecoding Mode=offline2026.0437