{"task":"Speech Recognition","dataset":"AISHELL-1","metric_names":["Word Error Rate (WER)","Params(M)"],"rows":[{"id":7287,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"FireRedASR-AED","metrics":{"Params(M)":"1,100","Word Error Rate (WER)":"0.55"},"paper_url":"https://arxiv.org/abs/2501.14350v1","paper_title":"FireRedASR: Open-Source Industrial-Grade Mandarin Speech Recognition Models from Encoder-Decoder to LLM Integration","paper_date":"2025-01-24","code_links":[{"title":"fireredteam/fireredasr","url":"https://github.com/fireredteam/fireredasr"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":1,"source":"archive","tags":[]},{"id":7288,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Seed-ASR","metrics":{"Word Error Rate (WER)":"0.68"},"paper_url":"https://arxiv.org/abs/2407.04675v2","paper_title":"Seed-ASR: Understanding Diverse Speech and Contexts with LLM-based Speech Recognition","paper_date":"2024-07-05","code_links":[],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":1,"source":"archive","tags":[]},{"id":7289,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Qwen-Audio","metrics":{"Word Error Rate (WER)":"1.29"},"paper_url":"https://arxiv.org/abs/2311.07919v2","paper_title":"Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models","paper_date":"2023-11-14","code_links":[{"title":"alibaba-damo-academy/FunASR","url":"https://github.com/alibaba-damo-academy/FunASR"},{"title":"qwenlm/qwen-audio","url":"https://github.com/qwenlm/qwen-audio"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":1,"source":"archive","tags":[]},{"id":7290,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"MMSpeech With LM","metrics":{"Word Error Rate (WER)":"1.9"},"paper_url":"https://arxiv.org/abs/2212.00500v1","paper_title":"MMSpeech: Multi-modal Multi-task Encoder-Decoder Pre-training for Speech Recognition","paper_date":"2022-11-29","code_links":[{"title":"ofa-sys/ofa","url":"https://github.com/ofa-sys/ofa"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7291,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Paraformer-large","metrics":{"Params(M)":"220","Word Error Rate (WER)":"1.95"},"paper_url":"https://arxiv.org/abs/2305.11013v1","paper_title":"FunASR: A Fundamental End-to-End Speech Recognition Toolkit","paper_date":"2023-05-18","code_links":[{"title":"alibaba-damo-academy/FunASR","url":"https://github.com/alibaba-damo-academy/FunASR"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":1,"source":"archive","tags":[]},{"id":7292,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Zipformer+CR-CTC (no external language model)","metrics":{"Params(M)":"66.2","Word Error Rate (WER)":"4.02"},"paper_url":"https://arxiv.org/abs/2410.05101v4","paper_title":"CR-CTC: Consistency regularization on CTC for improved speech recognition","paper_date":"2024-10-07","code_links":[{"title":"k2-fsa/icefall","url":"https://github.com/k2-fsa/icefall"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7293,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Lightweight Transducer With LM","metrics":{"Params(M)":"45.3","Word Error Rate (WER)":"4.03"},"paper_url":"https://arxiv.org/abs/2409.13698v2","paper_title":"Lightweight Transducer Based on Frame-Level Criterion","paper_date":"2024-09-05","code_links":[{"title":"wangmengzhi/Lightweight-Transducer","url":"https://github.com/wangmengzhi/Lightweight-Transducer"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7294,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"SE-WSBO With LM","metrics":{"Params(M)":"46","Word Error Rate (WER)":"4.1"},"paper_url":"https://arxiv.org/abs/2207.11697v5","paper_title":"Improving Mandarin Speech Recogntion with Block-augmented Transformer","paper_date":"2022-07-24","code_links":[{"title":"LeonWlw/asr_blockformer","url":"https://github.com/LeonWlw/asr_blockformer"},{"title":"mininglamp-technology/asr-blockformer","url":"https://github.com/mininglamp-technology/asr-blockformer"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7295,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"CIF-HKD With LM","metrics":{"Params(M)":"47","Word Error Rate (WER)":"4.1"},"paper_url":"https://arxiv.org/abs/2301.13003v2","paper_title":"Knowledge Transfer from Pre-trained Language Models to Cif-based Speech Recognizers via Hierarchical Distillation","paper_date":"2023-01-30","code_links":[{"title":"MingLunHan/CIF-PyTorch","url":"https://github.com/MingLunHan/CIF-PyTorch"},{"title":"minglunhan/cif-hieradist","url":"https://github.com/minglunhan/cif-hieradist"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7296,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Lightweight Transducer","metrics":{"Params(M)":"45.3","Word Error Rate (WER)":"4.31"},"paper_url":"https://arxiv.org/abs/2409.13698v2","paper_title":"Lightweight Transducer Based on Frame-Level Criterion","paper_date":"2024-09-05","code_links":[{"title":"wangmengzhi/Lightweight-Transducer","url":"https://github.com/wangmengzhi/Lightweight-Transducer"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7297,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"UMA","metrics":{"Params(M)":"44.7","Word Error Rate (WER)":"4.7"},"paper_url":"https://arxiv.org/abs/2309.08150v2","paper_title":"Unimodal Aggregation for CTC-based Speech Recognition","paper_date":"2023-09-15","code_links":[{"title":"Audio-WestlakeU/UMA-ASR","url":"https://github.com/Audio-WestlakeU/UMA-ASR"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7298,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"U2","metrics":{"Params(M)":"47","Word Error Rate (WER)":"4.72"},"paper_url":"https://arxiv.org/abs/2012.05481v2","paper_title":"Unified Streaming and Non-streaming Two-pass End-to-end Model for Speech Recognition","paper_date":"2020-12-10","code_links":[{"title":"PaddlePaddle/PaddleSpeech","url":"https://github.com/PaddlePaddle/PaddleSpeech"},{"title":"TeaPoly/Conformer-Athena","url":"https://github.com/TeaPoly/Conformer-Athena"},{"title":"xianchao-wu/wenet-deep-sparse-conformer","url":"https://github.com/xianchao-wu/wenet-deep-sparse-conformer"},{"title":"joseewei/wenet","url":"https://github.com/joseewei/wenet"},{"title":"Vill-Lab/2023-TMM-Grad-SAS","url":"https://github.com/Vill-Lab/2023-TMM-Grad-SAS"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7299,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Paraformer","metrics":{"Params(M)":"46.3","Word Error Rate (WER)":"4.95"},"paper_url":"https://arxiv.org/abs/2305.11013v1","paper_title":"FunASR: A Fundamental End-to-End Speech Recognition Toolkit","paper_date":"2023-05-18","code_links":[{"title":"alibaba-damo-academy/FunASR","url":"https://github.com/alibaba-damo-academy/FunASR"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7300,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"BAT","metrics":{"Params(M)":"90","Word Error Rate (WER)":"4.97"},"paper_url":"https://arxiv.org/abs/2305.11571v1","paper_title":"BAT: Boundary aware transducer for memory-efficient and low-latency ASR","paper_date":"2023-05-19","code_links":[{"title":"alibaba-damo-academy/FunASR","url":"https://github.com/alibaba-damo-academy/FunASR"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7301,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"CTC-CRF 4gram-LM","metrics":{"Word Error Rate (WER)":"6.34"},"paper_url":"https://arxiv.org/abs/2005.13326v2","paper_title":"CAT: A CTC-CRF based ASR Toolkit Bridging the Hybrid and the End-to-end Approaches towards Data Efficiency and Low Latency","paper_date":"2020-05-27","code_links":[{"title":"thu-spmi/cat","url":"https://github.com/thu-spmi/cat"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7302,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"BRA-E","metrics":{"Params(M)":"8.5","Word Error Rate (WER)":"6.63"},"paper_url":"https://arxiv.org/abs/2303.13072v2","paper_title":"Beyond Universal Transformer: block reusing with adaptor in Transformer for automatic speech recognition","paper_date":"2023-03-23","code_links":[],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":7303,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"CTC/Att","metrics":{"Word Error Rate (WER)":"6.7"},"paper_url":"https://arxiv.org/abs/1909.06317v2","paper_title":"A Comparative Study on Transformer vs RNN in Speech Applications","paper_date":"2019-09-13","code_links":[{"title":"espnet/espnet","url":"https://github.com/espnet/espnet"},{"title":"MindSpore-scientific-2/code-11","url":"https://github.com/MindSpore-scientific-2/code-11/tree/main/ZoneoutRNN"}],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]},{"id":758735,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Latent","metrics":{"Word Error Rate (WER)":"8.4"},"paper_url":"https://paperswithcode.com/paper/latent-softmax-for-data-efficient-phoneme-based-multilingual-asr-across-tonal-and-non-tonal-languages","paper_title":"Latent Softmax for Data-Efficient Phoneme-Based Multilingual ASR Across Tonal and Non-Tonal Languages","paper_date":"2026-08-02","code_links":[],"metrics_order":null,"area":null,"uses_additional_data":null,"source":"auto","tags":[]},{"id":759093,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Retrieval-Augmented Self-Taught Reasonin","metrics":{"Word Error Rate (WER)":"17.96"},"paper_url":"https://paperswithcode.com/paper/retrieval-augmented-self-taught-reasoning-model-with-adaptive-chain-of-thought-for-asr-named-entity-correction","paper_title":"Retrieval-Augmented Self-Taught Reasoning Model with Adaptive Chain-of-Thought for ASR Named Entity Correction","paper_date":"2026-01-21","code_links":[],"metrics_order":null,"area":null,"uses_additional_data":null,"source":"auto","tags":[]},{"id":7304,"task":"Speech Recognition","parent_task":null,"dataset":"AISHELL-1","model_name":"Att","metrics":{"Word Error Rate (WER)":"18.7"},"paper_url":"http://arxiv.org/abs/1808.10088v2","paper_title":"End-to-end Speech Recognition with Adaptive Computation Steps","paper_date":"2018-08-30","code_links":[],"metrics_order":"[\"Word Error Rate (WER)\", \"Params(M)\"]","area":"Audio","uses_additional_data":0,"source":"archive","tags":[]}]}