Share your thoughts, 1 month free Claude Pro on usSee more

Automatic Speech Recognition on AISHELL-1 (test)

0.57CER

FireRedASR2-AED

Updated 8d ago

Evaluation Results

Method
FireRedASR2-AED 2026.03	0.57	-	-	-	-	-	-	-	-
Ours 2026.04	0.59	-	-	-	-	-	-	-	-
FireRedASR2-LLM 2026.03	0.64	-	-	-	-	-	-	-	-
FireRedASR-LLM 2026.04	0.73	-	-	-	-	-	-	-	-
Step-Audio2-mini 2026.04	0.81	-	-	-	-	-	-	-	-
Qwen3-Omni-Inst 2026.04	0.92	-	-	-	-	-	-	-	-
Qwen2.5-Omni-7B 2026.03	1.12	-	-	-	-	-	-	-	-
Qwen3-ASR 2026.03	1.48	-	-	-	-	-	-	-	-
Qwen3-ASR-1.7B 2026.04	1.51	-	-	-	-	-	-	-	-
Doubao-ASR 2026.03	1.52	-	-	-	-	-	-	-	-
Fun-ASR 2026.03	1.64	-	-	-	-	-	-	-	-
Fun-ASR-nano 2026.04	1.81	-	-	-	-	-	-	-	-
Baichuan-Omni-1.5 2026.03	2.37	-	-	-	-	-	-	-	-
GLM-ASR-nano 2026.04	2.41	-	-	-	-	-	-	-	-
GLM-4-Voice 2026.03	2.46	-	-	-	-	-	-	-	-
O_cp 2023.09	3.1	-	-	-	-	-	-	-	-
SE-WSBO 2022.07	4.05	-	-	-	-	-	-	-	-
UOT-BERT-CTC 2025.09	4.06	-	-	-	-	-	-	-	-
Neural Transducer 2023.01	4.1	-	-	-	-	-	-	-	-
CIF + HKD 2023.01	4.1	-	-	-	-	-	-	-	-
O_nb 2023.09	4.1	-	-	-	-	-	-	-	-
UOT-BERT-CTC 2025.09	4.13	-	-	-	-	-	-	-	-
UOT-BERT-CTC 2025.09	4.13	-	-	-	-	-	-	-	-
Neural Transducer+LFMMI 2022.07	4.18	-	-	-	-	-	-	-	-
OT-BERT-CTC 2025.09	4.19	-	-	-	-	-	-	-	-
CIF + HKD 2023.01	4.2	-	-	-	-	-	-	-	-
UOT-BERT-CTC 2025.09	4.21	-	-	-	-	-	-	-	-
Base-WSBO 2022.07	4.22	-	-	-	-	-	-	-	-
K2 2022.07	4.26	-	-	-	-	-	-	-	-
Base-WSBO+Softmax 2022.07	4.28	-	-	-	-	-	-	-	-
Zipformer-L 2023.10	4.28	-	-	-	-	-	-	-	-
SE-WSBO 2022.07	4.29	-	-	-	-	-	-	-	-
Icefall 2023.01	4.3	-	-	-	-	-	-	-	-
WeNet 2022.07	4.36	-	-	-	-	-	-	-	-
Branchformer 2023.01	4.4	-	-	-	-	-	-	-	-
WeNet 2023.01	4.4	-	-	-	-	-	-	-	-
CIF + LRD 2023.01	4.4	-	-	-	-	-	-	-	-
Zipformer-M 2023.10	4.4	-	-	-	-	-	-	-	-
Branchformer 2023.10	4.43	-	-	-	-	-	-	-	-
Base-WSBO 2022.07	4.48	-	-	-	-	-	-	-	-
CIF + LRD 2023.01	4.5	-	-	-	-	-	-	-	-
E-Branchformer in ESPnet 2023.10	4.5	-	-	-	-	-	-	-	-
Base-WSBO+Softmax 2022.07	4.54	-	-	-	-	-	-	-	-
CIF + ACD 2023.01	4.6	-	-	-	-	-	-	-	-
Hybrid (Conformer) 2023.09	4.6	-	-	-	-	-	-	-	-
WeNet 2022.07	4.61	-	-	-	-	-	-	-	-
Conformer in WeNet 2023.10	4.61	-	-	-	-	-	-	-	-
Zipformer-S 2023.10	4.67	-	-	-	-	-	-	-	-
LAS 2024.09	4.67	-	-	-	-	-	-	-	-
NAR-BERT-ASR 2025.09	4.68	-	-	-	-	-	-	-	-
Espnet 2022.07	4.7	-	-	-	-	-	-	-	-
ESPnet Conformer 2023.01	4.7	-	-	-	-	-	-	-	-
CIF + ACD 2023.01	4.7	-	-	-	-	-	-	-	-
UMA + self-condition 2023.09	4.7	-	-	-	-	-	-	-	-
Lightweight Transducer 2024.09	4.76	-	-	-	-	-	-	-	-
CIF 2023.01	4.8	-	-	-	-	-	-	-	-
UMA 2023.09	4.8	-	-	-	-	-	-	-	-
Conformer+CTC/AED 2025.09	4.82	-	-	-	-	-	-	-	-
Espnet 2022.07	4.9	-	-	-	-	-	-	-	-
ESPnet Conformer 2023.01	4.9	-	-	-	-	-	-	-	-
CIF 2023.01	4.9	-	-	-	-	-	-	-	-
Self-conditioned CTC 2023.09	4.9	-	-	-	-	-	-	-	-
Conformer in ESPnet 2023.10	4.9	-	-	-	-	-	-	-	-
Proposed 2026.01	4.9	-	-	-	-	-	-	-	-
H2T-LoRA 2023.09	5	-	-	-	-	-	-	-	-
WeNet-U2 2026.01	5	-	-	-	-	-	-	-	-
Transducer 2024.09	5.07	-	-	-	-	-	-	-	-
Proposed 2026.01	5.1	-	-	-	-	-	-	-	-
Paraformer 2023.09	5.2	-	-	-	-	-	-	-	-
BESTOW 2026.01	5.3	-	-	-	-	-	-	-	-
U2 2020.12	5.33	-	-	-	-	-	-	-	-
SPEECH-OMNI-LITE 2026.03	5.46	-	-	-	-	-	-	-	-
Hybrid (Conformer) 2023.09	5.6	-	-	-	-	-	-	-	-
Conformer+CTC 2025.09	5.76	-	-	-	-	-	-	-	-
SPEECH-OMNI-LITE 2026.03	5.77	-	-	-	-	-	-	-	-
BTI 2026.01	5.9	-	-	-	-	-	-	-	-
CTC 2023.09	6.1	-	-	-	-	-	-	-	-
Baseline 2023.09	6.3	-	-	-	-	-	-	-	-
Baseline-non-stream 2026.01	6.5	-	-	-	-	-	-	-	-
MMA 2020.12	6.6	-	-	-	-	-	-	-	-
LASO-large 2023.09	6.6	-	-	-	-	-	-	-	-
Baseline-stream 2026.01	6.9	-	-	-	-	-	-	-	-
RNN-T 2023.05	6.96	4,633	100	230	-	-	-	-	-
BAT 2023.05	7.35	4,523	10	140	-	-	-	-	-
SCAMA 2020.12	7.39	-	-	-	-	-	-	-	-
RNN-T + FE 2023.05	7.44	4,517	-20	120	-	-	-	-	-
BAT 2023.05	7.63	4,483	-10	120	-	-	-	-	-
RNN-T + FE 2023.05	7.67	4,416	-110	30	-	-	-	-	-
Sync-Transformer 2020.12	8.91	-	-	-	-	-	-	-	-
SPEECH-OMNI-LITE 2026.03	9.01	-	-	-	-	-	-	-	-
FireRedASR-AED 2025.01	55	-	-	-	-	-	-	-	-
Seed-ASR 2025.01	68	-	-	-	-	-	-	-	-
FireRedASR-LLM 2025.01	76	-	-	-	-	-	-	-	-
Qwen-Audio 2025.01	130	-	-	-	-	-	-	-	-
Paraformer-Large 2025.01	168	-	-	-	-	-	-	-	-
SenseVoice-L 2025.01	209	-	-	-	-	-	-	-	-
Whisper-Large-v3 2025.01	514	-	-	-	-	-	-	-	-
Wav2Vec 2.0 2022.11	-	-	-	-	4.7	-	-	-	-
HuBERT 2022.11	-	-	-	-	4.3	-	-	-	-
MMSpeech 2022.11	-	-	-	-	2.1	-	-	-	-

Showing 100 of 159 rows