Speech Recognition On Librispeech Test Other
평가 지표
Word Error Rate (WER)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | Word Error Rate (WER) |
|---|---|
| espresso-a-fast-end-to-end-neural-speech | 8.7 |
| jasper-an-end-to-end-convolutional-neural | 7.84 |
| end-to-end-asr-from-supervised-to-semi | 5.18 |
| contextnet-improving-convolutional-neural | 4.1 |
| samba-asr-state-of-the-art-speech-recognition | 2.48 |
| specaugment-a-simple-data-augmentation-method | 5.8 |
| jasper-an-end-to-end-convolutional-neural | 8.79 |
| fully-convolutional-speech-recognition | 10.47 |
| fadam-adam-is-a-natural-gradient-optimizer | 2.49 |
| iterative-pseudo-labeling-for-speech | 3.83 |
| conformer-convolution-augmented-transformer | 4.3 |
| state-of-the-art-speech-recognition-using | 5.80 |
| contextnet-improving-convolutional-neural | 4.5 |
| hubert-self-supervised-speech-representation | 2.9 |
| cr-ctc-consistency-regularization-on-ctc-for | 3.95 |
| semi-supervised-speech-recognition-via-local | 15.28 |
| mt4ssl-boosting-self-supervised-speech | 9.6 |
| conformer-convolution-augmented-transformer | 5.0 |
| graph-convolutions-enrich-the-self-attention | 4.94 |
| cr-ctc-consistency-regularization-on-ctc-for | 4.35 |
| squeezeformer-an-efficient-transformer-for | 5.97 |
| crf-based-single-stage-acoustic-modeling-with | 10.65 |
| speechstew-simply-mix-all-available-speech | 3.3 |
| improved-noisy-student-training-for-automatic | 3.4 |
| quartznet-deep-automatic-speech-recognition | 7.25 |
| self-training-and-pre-training-are | 3.1 |
| qwen-audio-advancing-universal-audio | 4.2 |
| transformer-based-acoustic-modeling-for | 4.85 |
| speechstew-simply-mix-all-available-speech | 4.0 |
| asapp-asr-multistream-cnn-and-self-attentive | 4.46 |
| rwth-asr-systems-for-librispeech-hybrid-vs | 5.0 |
| specaugment-a-simple-data-augmentation-method | 6.5 |
| improving-rnn-transducer-based-asr-with | 4.20 |
| wav2vec-2-0-a-framework-for-self-supervised | 3.0 |
| e-branchformer-branchformer-with-enhanced | 3.65 |
| wavlm-large-scale-self-supervised-pre | 3.2 |
| fast-simpler-and-more-accurate-hybrid-asr | 4.20 |
| semi-supervised-speech-recognition-via-local | 20.84 |
| librispeech-transducer-model-with-internal | 5.6 |
| a-comparative-study-on-transformer-vs-rnn-in | 5.7 |
| pushing-the-limits-of-semi-supervised | 2.6 |
| end-to-end-asr-from-supervised-to-semi | 4.11 |
| snips-voice-platform-an-embedded-spoken | 16.5 |
| zipformer-a-faster-and-better-encoder-for | 4.38 |
| conformer-convolution-augmented-transformer | 3.9 |
| w2v-bert-combining-contrastive-learning-and | 2.5 |
| neural-network-language-modeling-with-letter | 7.63 |
| deep-speech-2-end-to-end-speech-recognition | 13.25 |
| data2vec-a-general-framework-for-self-1 | 3.7 |
| 모델 50 | 12.5 |
| contextnet-improving-convolutional-neural | 5.5 |
| relaxed-attention-a-simple-method-to-boost | 6.85 |
| wav2vec-2-0-a-framework-for-self-supervised | 4.1 |